Commit
•
e9a9d92
1
Parent(s):
3167145
Model save
Browse files- README.md +66 -0
- all_results.json +9 -0
- config.json +30 -0
- generation_config.json +8 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- runs/Aug26_20-32-44_GPU-4090s7/events.out.tfevents.1724697710.GPU-4090s7.3310924.0 +3 -0
- special_tokens_map.json +28 -0
- tokenizer.json +0 -0
- tokenizer_config.json +154 -0
- train_results.json +9 -0
- trainer_state.json +3739 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: HuggingFaceTB/SmolLM-360M-Instruct
|
| 5 |
+
tags:
|
| 6 |
+
- trl
|
| 7 |
+
- sft
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
datasets:
|
| 10 |
+
- generator
|
| 11 |
+
model-index:
|
| 12 |
+
- name: smollm-360M-aliases-5
|
| 13 |
+
results: []
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 17 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 18 |
+
|
| 19 |
+
# smollm-360M-aliases-5
|
| 20 |
+
|
| 21 |
+
This model is a fine-tuned version of [HuggingFaceTB/SmolLM-360M-Instruct](https://huggingface.co/HuggingFaceTB/SmolLM-360M-Instruct) on the generator dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 2.2808
|
| 24 |
+
|
| 25 |
+
## Model description
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Intended uses & limitations
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training and evaluation data
|
| 34 |
+
|
| 35 |
+
More information needed
|
| 36 |
+
|
| 37 |
+
## Training procedure
|
| 38 |
+
|
| 39 |
+
### Training hyperparameters
|
| 40 |
+
|
| 41 |
+
The following hyperparameters were used during training:
|
| 42 |
+
- learning_rate: 0.001
|
| 43 |
+
- train_batch_size: 4
|
| 44 |
+
- eval_batch_size: 4
|
| 45 |
+
- seed: 42
|
| 46 |
+
- distributed_type: multi-GPU
|
| 47 |
+
- gradient_accumulation_steps: 4
|
| 48 |
+
- total_train_batch_size: 16
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: cosine
|
| 51 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 1
|
| 53 |
+
|
| 54 |
+
### Training results
|
| 55 |
+
|
| 56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 57 |
+
|:-------------:|:------:|:----:|:---------------:|
|
| 58 |
+
| 0.7448 | 0.9997 | 2633 | 2.2808 |
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
### Framework versions
|
| 62 |
+
|
| 63 |
+
- Transformers 4.44.2
|
| 64 |
+
- Pytorch 2.4.0+cu121
|
| 65 |
+
- Datasets 2.21.0
|
| 66 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9997152349311818,
|
| 3 |
+
"total_flos": 32299470028800.0,
|
| 4 |
+
"train_loss": 1.3562805264158209,
|
| 5 |
+
"train_runtime": 3817.2623,
|
| 6 |
+
"train_samples": 750000,
|
| 7 |
+
"train_samples_per_second": 11.039,
|
| 8 |
+
"train_steps_per_second": 0.69
|
| 9 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "HuggingFaceTB/SmolLM-360M-Instruct",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "silu",
|
| 11 |
+
"hidden_size": 960,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 2560,
|
| 14 |
+
"max_position_embeddings": 2048,
|
| 15 |
+
"mlp_bias": false,
|
| 16 |
+
"model_type": "llama",
|
| 17 |
+
"num_attention_heads": 15,
|
| 18 |
+
"num_hidden_layers": 32,
|
| 19 |
+
"num_key_value_heads": 5,
|
| 20 |
+
"pad_token_id": 2,
|
| 21 |
+
"pretraining_tp": 1,
|
| 22 |
+
"rms_norm_eps": 1e-05,
|
| 23 |
+
"rope_scaling": null,
|
| 24 |
+
"rope_theta": 10000.0,
|
| 25 |
+
"tie_word_embeddings": true,
|
| 26 |
+
"torch_dtype": "bfloat16",
|
| 27 |
+
"transformers_version": "4.44.2",
|
| 28 |
+
"use_cache": false,
|
| 29 |
+
"vocab_size": 49152
|
| 30 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"max_new_tokens": 40,
|
| 6 |
+
"pad_token_id": 2,
|
| 7 |
+
"transformers_version": "4.44.2"
|
| 8 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:404eeb911bf2394c6c2e28a58c417a0b0fd0bdc700e6c0b8ca72d59c27ed89b3
|
| 3 |
+
size 723674912
|
runs/Aug26_20-32-44_GPU-4090s7/events.out.tfevents.1724697710.GPU-4090s7.3310924.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a958fe294d4e3b83b44e96a82afdee9c50957af306b95470a4f6ef5e52d6f38a
|
| 3 |
+
size 117169
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
{
|
| 4 |
+
"content": "<|im_start|>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
{
|
| 11 |
+
"content": "<|im_end|>",
|
| 12 |
+
"lstrip": false,
|
| 13 |
+
"normalized": false,
|
| 14 |
+
"rstrip": false,
|
| 15 |
+
"single_word": false
|
| 16 |
+
}
|
| 17 |
+
],
|
| 18 |
+
"bos_token": "<|im_start|>",
|
| 19 |
+
"eos_token": "<|im_end|>",
|
| 20 |
+
"pad_token": "<|im_end|>",
|
| 21 |
+
"unk_token": {
|
| 22 |
+
"content": "<|endoftext|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false
|
| 27 |
+
}
|
| 28 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,154 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<|im_start|>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "<|im_end|>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<repo_name>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"4": {
|
| 37 |
+
"content": "<reponame>",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
},
|
| 44 |
+
"5": {
|
| 45 |
+
"content": "<file_sep>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false,
|
| 50 |
+
"special": true
|
| 51 |
+
},
|
| 52 |
+
"6": {
|
| 53 |
+
"content": "<filename>",
|
| 54 |
+
"lstrip": false,
|
| 55 |
+
"normalized": false,
|
| 56 |
+
"rstrip": false,
|
| 57 |
+
"single_word": false,
|
| 58 |
+
"special": true
|
| 59 |
+
},
|
| 60 |
+
"7": {
|
| 61 |
+
"content": "<gh_stars>",
|
| 62 |
+
"lstrip": false,
|
| 63 |
+
"normalized": false,
|
| 64 |
+
"rstrip": false,
|
| 65 |
+
"single_word": false,
|
| 66 |
+
"special": true
|
| 67 |
+
},
|
| 68 |
+
"8": {
|
| 69 |
+
"content": "<issue_start>",
|
| 70 |
+
"lstrip": false,
|
| 71 |
+
"normalized": false,
|
| 72 |
+
"rstrip": false,
|
| 73 |
+
"single_word": false,
|
| 74 |
+
"special": true
|
| 75 |
+
},
|
| 76 |
+
"9": {
|
| 77 |
+
"content": "<issue_comment>",
|
| 78 |
+
"lstrip": false,
|
| 79 |
+
"normalized": false,
|
| 80 |
+
"rstrip": false,
|
| 81 |
+
"single_word": false,
|
| 82 |
+
"special": true
|
| 83 |
+
},
|
| 84 |
+
"10": {
|
| 85 |
+
"content": "<issue_closed>",
|
| 86 |
+
"lstrip": false,
|
| 87 |
+
"normalized": false,
|
| 88 |
+
"rstrip": false,
|
| 89 |
+
"single_word": false,
|
| 90 |
+
"special": true
|
| 91 |
+
},
|
| 92 |
+
"11": {
|
| 93 |
+
"content": "<jupyter_start>",
|
| 94 |
+
"lstrip": false,
|
| 95 |
+
"normalized": false,
|
| 96 |
+
"rstrip": false,
|
| 97 |
+
"single_word": false,
|
| 98 |
+
"special": true
|
| 99 |
+
},
|
| 100 |
+
"12": {
|
| 101 |
+
"content": "<jupyter_text>",
|
| 102 |
+
"lstrip": false,
|
| 103 |
+
"normalized": false,
|
| 104 |
+
"rstrip": false,
|
| 105 |
+
"single_word": false,
|
| 106 |
+
"special": true
|
| 107 |
+
},
|
| 108 |
+
"13": {
|
| 109 |
+
"content": "<jupyter_code>",
|
| 110 |
+
"lstrip": false,
|
| 111 |
+
"normalized": false,
|
| 112 |
+
"rstrip": false,
|
| 113 |
+
"single_word": false,
|
| 114 |
+
"special": true
|
| 115 |
+
},
|
| 116 |
+
"14": {
|
| 117 |
+
"content": "<jupyter_output>",
|
| 118 |
+
"lstrip": false,
|
| 119 |
+
"normalized": false,
|
| 120 |
+
"rstrip": false,
|
| 121 |
+
"single_word": false,
|
| 122 |
+
"special": true
|
| 123 |
+
},
|
| 124 |
+
"15": {
|
| 125 |
+
"content": "<jupyter_script>",
|
| 126 |
+
"lstrip": false,
|
| 127 |
+
"normalized": false,
|
| 128 |
+
"rstrip": false,
|
| 129 |
+
"single_word": false,
|
| 130 |
+
"special": true
|
| 131 |
+
},
|
| 132 |
+
"16": {
|
| 133 |
+
"content": "<empty_output>",
|
| 134 |
+
"lstrip": false,
|
| 135 |
+
"normalized": false,
|
| 136 |
+
"rstrip": false,
|
| 137 |
+
"single_word": false,
|
| 138 |
+
"special": true
|
| 139 |
+
}
|
| 140 |
+
},
|
| 141 |
+
"additional_special_tokens": [
|
| 142 |
+
"<|im_start|>",
|
| 143 |
+
"<|im_end|>"
|
| 144 |
+
],
|
| 145 |
+
"bos_token": "<|im_start|>",
|
| 146 |
+
"chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
|
| 147 |
+
"clean_up_tokenization_spaces": false,
|
| 148 |
+
"eos_token": "<|im_end|>",
|
| 149 |
+
"model_max_length": 2048,
|
| 150 |
+
"pad_token": "<|im_end|>",
|
| 151 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 152 |
+
"unk_token": "<|endoftext|>",
|
| 153 |
+
"vocab_size": 49152
|
| 154 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9997152349311818,
|
| 3 |
+
"total_flos": 32299470028800.0,
|
| 4 |
+
"train_loss": 1.3562805264158209,
|
| 5 |
+
"train_runtime": 3817.2623,
|
| 6 |
+
"train_samples": 750000,
|
| 7 |
+
"train_samples_per_second": 11.039,
|
| 8 |
+
"train_steps_per_second": 0.69
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3739 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9997152349311818,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2633,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.00037968675842429994,
|
| 13 |
+
"grad_norm": 4.039186105207878,
|
| 14 |
+
"learning_rate": 3.7878787878787882e-06,
|
| 15 |
+
"loss": 3.1037,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0018984337921214998,
|
| 20 |
+
"grad_norm": 3.978803446290656,
|
| 21 |
+
"learning_rate": 1.893939393939394e-05,
|
| 22 |
+
"loss": 3.0676,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0037968675842429997,
|
| 27 |
+
"grad_norm": 3.4024459371591336,
|
| 28 |
+
"learning_rate": 3.787878787878788e-05,
|
| 29 |
+
"loss": 3.0058,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.005695301376364499,
|
| 34 |
+
"grad_norm": 0.769937135761119,
|
| 35 |
+
"learning_rate": 5.681818181818182e-05,
|
| 36 |
+
"loss": 2.7987,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.007593735168485999,
|
| 41 |
+
"grad_norm": 0.7081139470071718,
|
| 42 |
+
"learning_rate": 7.575757575757576e-05,
|
| 43 |
+
"loss": 2.6742,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.009492168960607499,
|
| 48 |
+
"grad_norm": 0.5663597400653296,
|
| 49 |
+
"learning_rate": 9.46969696969697e-05,
|
| 50 |
+
"loss": 2.5936,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.011390602752728999,
|
| 55 |
+
"grad_norm": 0.3713010333300208,
|
| 56 |
+
"learning_rate": 0.00011363636363636364,
|
| 57 |
+
"loss": 2.5012,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.013289036544850499,
|
| 62 |
+
"grad_norm": 0.4205375561732573,
|
| 63 |
+
"learning_rate": 0.00013257575757575756,
|
| 64 |
+
"loss": 2.433,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.015187470336971999,
|
| 69 |
+
"grad_norm": 0.587958832384919,
|
| 70 |
+
"learning_rate": 0.00015151515151515152,
|
| 71 |
+
"loss": 2.4168,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.0170859041290935,
|
| 76 |
+
"grad_norm": 0.6677663506298054,
|
| 77 |
+
"learning_rate": 0.00017045454545454544,
|
| 78 |
+
"loss": 2.3662,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.018984337921214997,
|
| 83 |
+
"grad_norm": 0.7064561950273625,
|
| 84 |
+
"learning_rate": 0.0001893939393939394,
|
| 85 |
+
"loss": 2.359,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0208827717133365,
|
| 90 |
+
"grad_norm": 0.35917369578708264,
|
| 91 |
+
"learning_rate": 0.00020833333333333335,
|
| 92 |
+
"loss": 2.3354,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.022781205505457997,
|
| 97 |
+
"grad_norm": 0.39187189416502427,
|
| 98 |
+
"learning_rate": 0.00022727272727272727,
|
| 99 |
+
"loss": 2.2868,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.024679639297579496,
|
| 104 |
+
"grad_norm": 0.3470459858223155,
|
| 105 |
+
"learning_rate": 0.0002462121212121212,
|
| 106 |
+
"loss": 2.2955,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.026578073089700997,
|
| 111 |
+
"grad_norm": 0.6714289041162775,
|
| 112 |
+
"learning_rate": 0.0002651515151515151,
|
| 113 |
+
"loss": 2.2607,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.028476506881822496,
|
| 118 |
+
"grad_norm": 0.43889084397376665,
|
| 119 |
+
"learning_rate": 0.00028409090909090913,
|
| 120 |
+
"loss": 2.2774,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.030374940673943997,
|
| 125 |
+
"grad_norm": 0.35987531792307215,
|
| 126 |
+
"learning_rate": 0.00030303030303030303,
|
| 127 |
+
"loss": 2.2579,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.0322733744660655,
|
| 132 |
+
"grad_norm": 0.45583566123152425,
|
| 133 |
+
"learning_rate": 0.000321969696969697,
|
| 134 |
+
"loss": 2.2278,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.034171808258187,
|
| 139 |
+
"grad_norm": 0.5495955157286524,
|
| 140 |
+
"learning_rate": 0.0003409090909090909,
|
| 141 |
+
"loss": 2.2217,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.036070242050308496,
|
| 146 |
+
"grad_norm": 0.5181443647177656,
|
| 147 |
+
"learning_rate": 0.0003598484848484849,
|
| 148 |
+
"loss": 2.2183,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.037968675842429994,
|
| 153 |
+
"grad_norm": 0.969821145035109,
|
| 154 |
+
"learning_rate": 0.0003787878787878788,
|
| 155 |
+
"loss": 2.2112,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.03986710963455149,
|
| 160 |
+
"grad_norm": 0.2635501239040389,
|
| 161 |
+
"learning_rate": 0.00039772727272727274,
|
| 162 |
+
"loss": 2.2081,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.041765543426673,
|
| 167 |
+
"grad_norm": 0.5715349391832327,
|
| 168 |
+
"learning_rate": 0.0004166666666666667,
|
| 169 |
+
"loss": 2.1948,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.043663977218794496,
|
| 174 |
+
"grad_norm": 0.5248994310808021,
|
| 175 |
+
"learning_rate": 0.0004356060606060606,
|
| 176 |
+
"loss": 2.1855,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.045562411010915994,
|
| 181 |
+
"grad_norm": 0.5608654765973496,
|
| 182 |
+
"learning_rate": 0.00045454545454545455,
|
| 183 |
+
"loss": 2.1841,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.04746084480303749,
|
| 188 |
+
"grad_norm": 0.379235271981335,
|
| 189 |
+
"learning_rate": 0.0004734848484848485,
|
| 190 |
+
"loss": 2.1668,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.04935927859515899,
|
| 195 |
+
"grad_norm": 0.6934486639382003,
|
| 196 |
+
"learning_rate": 0.0004924242424242425,
|
| 197 |
+
"loss": 2.1799,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.051257712387280496,
|
| 202 |
+
"grad_norm": 0.32726936547789276,
|
| 203 |
+
"learning_rate": 0.0005113636363636364,
|
| 204 |
+
"loss": 2.1539,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.053156146179401995,
|
| 209 |
+
"grad_norm": 0.4103864066901133,
|
| 210 |
+
"learning_rate": 0.0005303030303030302,
|
| 211 |
+
"loss": 2.1559,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.05505457997152349,
|
| 216 |
+
"grad_norm": 0.6157950630213448,
|
| 217 |
+
"learning_rate": 0.0005492424242424242,
|
| 218 |
+
"loss": 2.1297,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.05695301376364499,
|
| 223 |
+
"grad_norm": 1.4548659360346616,
|
| 224 |
+
"learning_rate": 0.0005681818181818183,
|
| 225 |
+
"loss": 2.161,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.05885144755576649,
|
| 230 |
+
"grad_norm": 0.6872944738590898,
|
| 231 |
+
"learning_rate": 0.0005871212121212122,
|
| 232 |
+
"loss": 2.1462,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.060749881347887995,
|
| 237 |
+
"grad_norm": 1.1748112715233312,
|
| 238 |
+
"learning_rate": 0.0006060606060606061,
|
| 239 |
+
"loss": 2.1558,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.06264831514000949,
|
| 244 |
+
"grad_norm": 0.3342527623973018,
|
| 245 |
+
"learning_rate": 0.000625,
|
| 246 |
+
"loss": 2.1381,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.064546748932131,
|
| 251 |
+
"grad_norm": 0.3090958929703956,
|
| 252 |
+
"learning_rate": 0.000643939393939394,
|
| 253 |
+
"loss": 2.1311,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.0664451827242525,
|
| 258 |
+
"grad_norm": 0.4445838850369956,
|
| 259 |
+
"learning_rate": 0.0006628787878787878,
|
| 260 |
+
"loss": 2.109,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.068343616516374,
|
| 265 |
+
"grad_norm": 0.40447800943933176,
|
| 266 |
+
"learning_rate": 0.0006818181818181818,
|
| 267 |
+
"loss": 2.1338,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.0702420503084955,
|
| 272 |
+
"grad_norm": 0.5270438533785088,
|
| 273 |
+
"learning_rate": 0.0007007575757575758,
|
| 274 |
+
"loss": 2.1221,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.07214048410061699,
|
| 279 |
+
"grad_norm": 1.1988668583819744,
|
| 280 |
+
"learning_rate": 0.0007196969696969698,
|
| 281 |
+
"loss": 2.0986,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.07403891789273849,
|
| 286 |
+
"grad_norm": 0.5173087158849319,
|
| 287 |
+
"learning_rate": 0.0007386363636363636,
|
| 288 |
+
"loss": 2.097,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.07593735168485999,
|
| 293 |
+
"grad_norm": 0.8713149911927134,
|
| 294 |
+
"learning_rate": 0.0007575757575757576,
|
| 295 |
+
"loss": 2.1171,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.07783578547698149,
|
| 300 |
+
"grad_norm": 0.43763842661820435,
|
| 301 |
+
"learning_rate": 0.0007765151515151515,
|
| 302 |
+
"loss": 2.1231,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.07973421926910298,
|
| 307 |
+
"grad_norm": 0.37755079956108445,
|
| 308 |
+
"learning_rate": 0.0007954545454545455,
|
| 309 |
+
"loss": 2.0974,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.08163265306122448,
|
| 314 |
+
"grad_norm": 0.3202651052626521,
|
| 315 |
+
"learning_rate": 0.0008143939393939394,
|
| 316 |
+
"loss": 2.1061,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.083531086853346,
|
| 321 |
+
"grad_norm": 1.0697394978550765,
|
| 322 |
+
"learning_rate": 0.0008333333333333334,
|
| 323 |
+
"loss": 2.1058,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.0854295206454675,
|
| 328 |
+
"grad_norm": 0.9419134947997803,
|
| 329 |
+
"learning_rate": 0.0008522727272727273,
|
| 330 |
+
"loss": 2.0832,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.08732795443758899,
|
| 335 |
+
"grad_norm": 0.3177328584436459,
|
| 336 |
+
"learning_rate": 0.0008712121212121212,
|
| 337 |
+
"loss": 2.0936,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.08922638822971049,
|
| 342 |
+
"grad_norm": 0.4260910025738285,
|
| 343 |
+
"learning_rate": 0.0008901515151515151,
|
| 344 |
+
"loss": 2.0836,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.09112482202183199,
|
| 349 |
+
"grad_norm": 0.37197183427696595,
|
| 350 |
+
"learning_rate": 0.0009090909090909091,
|
| 351 |
+
"loss": 2.0705,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.09302325581395349,
|
| 356 |
+
"grad_norm": 0.3469428530219821,
|
| 357 |
+
"learning_rate": 0.000928030303030303,
|
| 358 |
+
"loss": 2.0851,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.09492168960607499,
|
| 363 |
+
"grad_norm": 0.5432864309293935,
|
| 364 |
+
"learning_rate": 0.000946969696969697,
|
| 365 |
+
"loss": 2.0601,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.09682012339819648,
|
| 370 |
+
"grad_norm": 0.30271756878451744,
|
| 371 |
+
"learning_rate": 0.000965909090909091,
|
| 372 |
+
"loss": 2.068,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.09871855719031798,
|
| 377 |
+
"grad_norm": 0.613099552832842,
|
| 378 |
+
"learning_rate": 0.000984848484848485,
|
| 379 |
+
"loss": 2.0889,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.1006169909824395,
|
| 384 |
+
"grad_norm": 0.2879329961644459,
|
| 385 |
+
"learning_rate": 0.000999999560347478,
|
| 386 |
+
"loss": 2.0728,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.10251542477456099,
|
| 391 |
+
"grad_norm": 0.3522360779077239,
|
| 392 |
+
"learning_rate": 0.000999984172590384,
|
| 393 |
+
"loss": 2.0778,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.10441385856668249,
|
| 398 |
+
"grad_norm": 0.8444049085073484,
|
| 399 |
+
"learning_rate": 0.0009999468029803513,
|
| 400 |
+
"loss": 2.0645,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.10631229235880399,
|
| 405 |
+
"grad_norm": 0.2340268385510596,
|
| 406 |
+
"learning_rate": 0.0009998874531603381,
|
| 407 |
+
"loss": 2.0424,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.10821072615092549,
|
| 412 |
+
"grad_norm": 0.3788090273684815,
|
| 413 |
+
"learning_rate": 0.0009998061257396652,
|
| 414 |
+
"loss": 2.0373,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.11010915994304699,
|
| 419 |
+
"grad_norm": 0.4712223089490736,
|
| 420 |
+
"learning_rate": 0.0009997028242939002,
|
| 421 |
+
"loss": 2.0535,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.11200759373516848,
|
| 426 |
+
"grad_norm": 0.5222779081868865,
|
| 427 |
+
"learning_rate": 0.0009995775533647015,
|
| 428 |
+
"loss": 2.0383,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.11390602752728998,
|
| 433 |
+
"grad_norm": 0.38807754757014135,
|
| 434 |
+
"learning_rate": 0.0009994303184596178,
|
| 435 |
+
"loss": 2.0536,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.11580446131941148,
|
| 440 |
+
"grad_norm": 0.5558959816060527,
|
| 441 |
+
"learning_rate": 0.0009992611260518462,
|
| 442 |
+
"loss": 2.0476,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.11770289511153298,
|
| 447 |
+
"grad_norm": 0.5297156243713772,
|
| 448 |
+
"learning_rate": 0.0009990699835799469,
|
| 449 |
+
"loss": 2.0182,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.11960132890365449,
|
| 454 |
+
"grad_norm": 0.8463883226148833,
|
| 455 |
+
"learning_rate": 0.0009988568994475178,
|
| 456 |
+
"loss": 2.0527,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.12149976269577599,
|
| 461 |
+
"grad_norm": 0.3181378787047677,
|
| 462 |
+
"learning_rate": 0.0009986218830228234,
|
| 463 |
+
"loss": 2.0253,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.12339819648789749,
|
| 468 |
+
"grad_norm": 0.3005739806261844,
|
| 469 |
+
"learning_rate": 0.0009983649446383836,
|
| 470 |
+
"loss": 2.023,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.12529663028001897,
|
| 475 |
+
"grad_norm": 0.7577732908772633,
|
| 476 |
+
"learning_rate": 0.00099808609559052,
|
| 477 |
+
"loss": 2.0081,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.12719506407214048,
|
| 482 |
+
"grad_norm": 0.3189810102860473,
|
| 483 |
+
"learning_rate": 0.0009977853481388575,
|
| 484 |
+
"loss": 2.0103,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.129093497864262,
|
| 489 |
+
"grad_norm": 0.6410939817239918,
|
| 490 |
+
"learning_rate": 0.0009974627155057878,
|
| 491 |
+
"loss": 2.0029,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.13099193165638348,
|
| 496 |
+
"grad_norm": 0.606178387164357,
|
| 497 |
+
"learning_rate": 0.000997118211875886,
|
| 498 |
+
"loss": 1.9762,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.132890365448505,
|
| 503 |
+
"grad_norm": 0.3588634601930927,
|
| 504 |
+
"learning_rate": 0.0009967518523952875,
|
| 505 |
+
"loss": 1.9911,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.13478879924062648,
|
| 510 |
+
"grad_norm": 0.28091052944488715,
|
| 511 |
+
"learning_rate": 0.0009963636531710228,
|
| 512 |
+
"loss": 1.9776,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.136687233032748,
|
| 517 |
+
"grad_norm": 0.38172615236997626,
|
| 518 |
+
"learning_rate": 0.0009959536312703085,
|
| 519 |
+
"loss": 1.9787,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.13858566682486947,
|
| 524 |
+
"grad_norm": 0.4360166845102846,
|
| 525 |
+
"learning_rate": 0.0009955218047197978,
|
| 526 |
+
"loss": 1.9909,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.140484100616991,
|
| 531 |
+
"grad_norm": 0.35275497258123356,
|
| 532 |
+
"learning_rate": 0.000995068192504787,
|
| 533 |
+
"loss": 1.9889,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.14238253440911247,
|
| 538 |
+
"grad_norm": 0.6317060282966646,
|
| 539 |
+
"learning_rate": 0.0009945928145683814,
|
| 540 |
+
"loss": 1.9551,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.14428096820123398,
|
| 545 |
+
"grad_norm": 0.3850672622738674,
|
| 546 |
+
"learning_rate": 0.0009940956918106183,
|
| 547 |
+
"loss": 1.9585,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.1461794019933555,
|
| 552 |
+
"grad_norm": 0.4220457511407704,
|
| 553 |
+
"learning_rate": 0.0009935768460875483,
|
| 554 |
+
"loss": 1.951,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.14807783578547698,
|
| 559 |
+
"grad_norm": 0.7783987596579233,
|
| 560 |
+
"learning_rate": 0.0009930363002102743,
|
| 561 |
+
"loss": 1.9724,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.1499762695775985,
|
| 566 |
+
"grad_norm": 0.32401219109209967,
|
| 567 |
+
"learning_rate": 0.0009924740779439483,
|
| 568 |
+
"loss": 1.9667,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.15187470336971998,
|
| 573 |
+
"grad_norm": 0.2857979164581651,
|
| 574 |
+
"learning_rate": 0.0009918902040067276,
|
| 575 |
+
"loss": 1.9462,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.1537731371618415,
|
| 580 |
+
"grad_norm": 0.6210777091831657,
|
| 581 |
+
"learning_rate": 0.000991284704068686,
|
| 582 |
+
"loss": 1.9618,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.15567157095396297,
|
| 587 |
+
"grad_norm": 0.4226037865904592,
|
| 588 |
+
"learning_rate": 0.0009906576047506884,
|
| 589 |
+
"loss": 1.9602,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.15757000474608449,
|
| 594 |
+
"grad_norm": 0.2508966539644492,
|
| 595 |
+
"learning_rate": 0.0009900089336232166,
|
| 596 |
+
"loss": 1.9518,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.15946843853820597,
|
| 601 |
+
"grad_norm": 0.28625033438978126,
|
| 602 |
+
"learning_rate": 0.0009893387192051607,
|
| 603 |
+
"loss": 1.937,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.16136687233032748,
|
| 608 |
+
"grad_norm": 0.5695158778386894,
|
| 609 |
+
"learning_rate": 0.0009886469909625624,
|
| 610 |
+
"loss": 1.9356,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.16326530612244897,
|
| 615 |
+
"grad_norm": 0.3444114501804719,
|
| 616 |
+
"learning_rate": 0.0009879337793073219,
|
| 617 |
+
"loss": 1.9292,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.16516373991457048,
|
| 622 |
+
"grad_norm": 0.3562145151770506,
|
| 623 |
+
"learning_rate": 0.000987199115595859,
|
| 624 |
+
"loss": 1.9446,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.167062173706692,
|
| 629 |
+
"grad_norm": 0.6662044045300556,
|
| 630 |
+
"learning_rate": 0.0009864430321277354,
|
| 631 |
+
"loss": 1.9123,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.16896060749881348,
|
| 636 |
+
"grad_norm": 0.4008170455675707,
|
| 637 |
+
"learning_rate": 0.0009856655621442347,
|
| 638 |
+
"loss": 1.9153,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.170859041290935,
|
| 643 |
+
"grad_norm": 0.5501520320546401,
|
| 644 |
+
"learning_rate": 0.0009848667398269005,
|
| 645 |
+
"loss": 1.9134,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.17275747508305647,
|
| 650 |
+
"grad_norm": 0.6229887015853178,
|
| 651 |
+
"learning_rate": 0.000984046600296034,
|
| 652 |
+
"loss": 1.9161,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.17465590887517798,
|
| 657 |
+
"grad_norm": 0.3263959879918121,
|
| 658 |
+
"learning_rate": 0.0009832051796091496,
|
| 659 |
+
"loss": 1.8977,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.17655434266729947,
|
| 664 |
+
"grad_norm": 0.32030075439432676,
|
| 665 |
+
"learning_rate": 0.00098234251475939,
|
| 666 |
+
"loss": 1.9385,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.17845277645942098,
|
| 671 |
+
"grad_norm": 0.3241713798628184,
|
| 672 |
+
"learning_rate": 0.0009814586436738997,
|
| 673 |
+
"loss": 1.8839,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.18035121025154247,
|
| 678 |
+
"grad_norm": 0.710599959000179,
|
| 679 |
+
"learning_rate": 0.0009805536052121568,
|
| 680 |
+
"loss": 1.8957,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.18224964404366398,
|
| 685 |
+
"grad_norm": 0.5685009329132018,
|
| 686 |
+
"learning_rate": 0.000979627439164266,
|
| 687 |
+
"loss": 1.9039,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.1841480778357855,
|
| 692 |
+
"grad_norm": 0.2883177054081686,
|
| 693 |
+
"learning_rate": 0.0009786801862492075,
|
| 694 |
+
"loss": 1.8939,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.18604651162790697,
|
| 699 |
+
"grad_norm": 0.3306718558813855,
|
| 700 |
+
"learning_rate": 0.0009777118881130484,
|
| 701 |
+
"loss": 1.8871,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.1879449454200285,
|
| 706 |
+
"grad_norm": 0.27771057670095894,
|
| 707 |
+
"learning_rate": 0.000976722587327111,
|
| 708 |
+
"loss": 1.8709,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.18984337921214997,
|
| 713 |
+
"grad_norm": 0.39767226694729324,
|
| 714 |
+
"learning_rate": 0.0009757123273861006,
|
| 715 |
+
"loss": 1.8707,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.19174181300427148,
|
| 720 |
+
"grad_norm": 0.3486146634838498,
|
| 721 |
+
"learning_rate": 0.0009746811527061942,
|
| 722 |
+
"loss": 1.854,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.19364024679639297,
|
| 727 |
+
"grad_norm": 0.4080559390296307,
|
| 728 |
+
"learning_rate": 0.000973629108623087,
|
| 729 |
+
"loss": 1.8698,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.19553868058851448,
|
| 734 |
+
"grad_norm": 0.33081602292257706,
|
| 735 |
+
"learning_rate": 0.0009725562413900002,
|
| 736 |
+
"loss": 1.8464,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.19743711438063596,
|
| 741 |
+
"grad_norm": 0.6037306322904349,
|
| 742 |
+
"learning_rate": 0.000971462598175646,
|
| 743 |
+
"loss": 1.8439,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.19933554817275748,
|
| 748 |
+
"grad_norm": 0.2879376497544211,
|
| 749 |
+
"learning_rate": 0.0009703482270621553,
|
| 750 |
+
"loss": 1.8713,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.201233981964879,
|
| 755 |
+
"grad_norm": 0.6788567524588369,
|
| 756 |
+
"learning_rate": 0.0009692131770429629,
|
| 757 |
+
"loss": 1.8453,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.20313241575700047,
|
| 762 |
+
"grad_norm": 0.3161414574102799,
|
| 763 |
+
"learning_rate": 0.0009680574980206534,
|
| 764 |
+
"loss": 1.8594,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.20503084954912199,
|
| 769 |
+
"grad_norm": 0.5645515053171333,
|
| 770 |
+
"learning_rate": 0.0009668812408047678,
|
| 771 |
+
"loss": 1.8492,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.20692928334124347,
|
| 776 |
+
"grad_norm": 0.5534701462686209,
|
| 777 |
+
"learning_rate": 0.0009656844571095696,
|
| 778 |
+
"loss": 1.8465,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.20882771713336498,
|
| 783 |
+
"grad_norm": 0.2769884694724677,
|
| 784 |
+
"learning_rate": 0.0009644671995517705,
|
| 785 |
+
"loss": 1.8512,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.21072615092548647,
|
| 790 |
+
"grad_norm": 0.3389677879460361,
|
| 791 |
+
"learning_rate": 0.0009632295216482181,
|
| 792 |
+
"loss": 1.8338,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.21262458471760798,
|
| 797 |
+
"grad_norm": 0.3665481797512662,
|
| 798 |
+
"learning_rate": 0.000961971477813542,
|
| 799 |
+
"loss": 1.818,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.21452301850972946,
|
| 804 |
+
"grad_norm": 0.5102461937389484,
|
| 805 |
+
"learning_rate": 0.0009606931233577622,
|
| 806 |
+
"loss": 1.8304,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.21642145230185098,
|
| 811 |
+
"grad_norm": 0.31696862342234344,
|
| 812 |
+
"learning_rate": 0.0009593945144838571,
|
| 813 |
+
"loss": 1.8491,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.21831988609397246,
|
| 818 |
+
"grad_norm": 0.474051209621524,
|
| 819 |
+
"learning_rate": 0.0009580757082852929,
|
| 820 |
+
"loss": 1.7842,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.22021831988609397,
|
| 825 |
+
"grad_norm": 0.28232130793234095,
|
| 826 |
+
"learning_rate": 0.0009567367627435121,
|
| 827 |
+
"loss": 1.8143,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.22211675367821548,
|
| 832 |
+
"grad_norm": 0.3429346941858972,
|
| 833 |
+
"learning_rate": 0.0009553777367253867,
|
| 834 |
+
"loss": 1.8144,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.22401518747033697,
|
| 839 |
+
"grad_norm": 0.4168698468178576,
|
| 840 |
+
"learning_rate": 0.0009539986899806281,
|
| 841 |
+
"loss": 1.8069,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.22591362126245848,
|
| 846 |
+
"grad_norm": 0.2584426546217752,
|
| 847 |
+
"learning_rate": 0.0009525996831391607,
|
| 848 |
+
"loss": 1.8121,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.22781205505457996,
|
| 853 |
+
"grad_norm": 0.27734758082061345,
|
| 854 |
+
"learning_rate": 0.0009511807777084571,
|
| 855 |
+
"loss": 1.8144,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.22971048884670148,
|
| 860 |
+
"grad_norm": 0.29149160986727496,
|
| 861 |
+
"learning_rate": 0.0009497420360708331,
|
| 862 |
+
"loss": 1.8072,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.23160892263882296,
|
| 867 |
+
"grad_norm": 0.27795468706904236,
|
| 868 |
+
"learning_rate": 0.0009482835214807049,
|
| 869 |
+
"loss": 1.7843,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.23350735643094447,
|
| 874 |
+
"grad_norm": 0.37668313657202124,
|
| 875 |
+
"learning_rate": 0.0009468052980618091,
|
| 876 |
+
"loss": 1.7858,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.23540579022306596,
|
| 881 |
+
"grad_norm": 0.3635355026922671,
|
| 882 |
+
"learning_rate": 0.0009453074308043822,
|
| 883 |
+
"loss": 1.7828,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.23730422401518747,
|
| 888 |
+
"grad_norm": 0.5492319693073675,
|
| 889 |
+
"learning_rate": 0.0009437899855623046,
|
| 890 |
+
"loss": 1.7805,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.23920265780730898,
|
| 895 |
+
"grad_norm": 0.5045520053744543,
|
| 896 |
+
"learning_rate": 0.0009422530290502045,
|
| 897 |
+
"loss": 1.784,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.24110109159943047,
|
| 902 |
+
"grad_norm": 0.29163131272581627,
|
| 903 |
+
"learning_rate": 0.0009406966288405248,
|
| 904 |
+
"loss": 1.7719,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.24299952539155198,
|
| 909 |
+
"grad_norm": 0.48505663312724284,
|
| 910 |
+
"learning_rate": 0.0009391208533605527,
|
| 911 |
+
"loss": 1.7866,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.24489795918367346,
|
| 916 |
+
"grad_norm": 0.5368154913313934,
|
| 917 |
+
"learning_rate": 0.0009375257718894107,
|
| 918 |
+
"loss": 1.7909,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.24679639297579498,
|
| 923 |
+
"grad_norm": 0.35856621494537494,
|
| 924 |
+
"learning_rate": 0.0009359114545550116,
|
| 925 |
+
"loss": 1.7504,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.24869482676791646,
|
| 930 |
+
"grad_norm": 0.3659739590782703,
|
| 931 |
+
"learning_rate": 0.0009342779723309745,
|
| 932 |
+
"loss": 1.7605,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.25059326056003794,
|
| 937 |
+
"grad_norm": 0.511785026274882,
|
| 938 |
+
"learning_rate": 0.0009326253970335046,
|
| 939 |
+
"loss": 1.7158,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.25249169435215946,
|
| 944 |
+
"grad_norm": 0.30511971139113303,
|
| 945 |
+
"learning_rate": 0.0009309538013182364,
|
| 946 |
+
"loss": 1.7637,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.25439012814428097,
|
| 951 |
+
"grad_norm": 0.28292845629768315,
|
| 952 |
+
"learning_rate": 0.0009292632586770384,
|
| 953 |
+
"loss": 1.7601,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.2562885619364025,
|
| 958 |
+
"grad_norm": 0.24747199414214469,
|
| 959 |
+
"learning_rate": 0.000927553843434783,
|
| 960 |
+
"loss": 1.7477,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.258186995728524,
|
| 965 |
+
"grad_norm": 0.30515982384727836,
|
| 966 |
+
"learning_rate": 0.0009258256307460781,
|
| 967 |
+
"loss": 1.757,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.26008542952064545,
|
| 972 |
+
"grad_norm": 0.4086074740784753,
|
| 973 |
+
"learning_rate": 0.000924078696591963,
|
| 974 |
+
"loss": 1.7437,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.26198386331276696,
|
| 979 |
+
"grad_norm": 0.3606712756700905,
|
| 980 |
+
"learning_rate": 0.0009223131177765685,
|
| 981 |
+
"loss": 1.7156,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.2638822971048885,
|
| 986 |
+
"grad_norm": 0.31644420804415135,
|
| 987 |
+
"learning_rate": 0.0009205289719237393,
|
| 988 |
+
"loss": 1.738,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.26578073089701,
|
| 993 |
+
"grad_norm": 0.47740932792504026,
|
| 994 |
+
"learning_rate": 0.0009187263374736221,
|
| 995 |
+
"loss": 1.7548,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.26767916468913144,
|
| 1000 |
+
"grad_norm": 0.2710760174578327,
|
| 1001 |
+
"learning_rate": 0.0009169052936792164,
|
| 1002 |
+
"loss": 1.7425,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.26957759848125296,
|
| 1007 |
+
"grad_norm": 0.27636232385404463,
|
| 1008 |
+
"learning_rate": 0.0009150659206028904,
|
| 1009 |
+
"loss": 1.7267,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.27147603227337447,
|
| 1014 |
+
"grad_norm": 0.4625279056619356,
|
| 1015 |
+
"learning_rate": 0.0009132082991128607,
|
| 1016 |
+
"loss": 1.7387,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.273374466065496,
|
| 1021 |
+
"grad_norm": 4.000898556896865,
|
| 1022 |
+
"learning_rate": 0.0009113325108796374,
|
| 1023 |
+
"loss": 1.745,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.2752728998576175,
|
| 1028 |
+
"grad_norm": 0.38141390080270937,
|
| 1029 |
+
"learning_rate": 0.0009094386383724332,
|
| 1030 |
+
"loss": 1.7522,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.27717133364973895,
|
| 1035 |
+
"grad_norm": 0.35177413794126794,
|
| 1036 |
+
"learning_rate": 0.0009075267648555378,
|
| 1037 |
+
"loss": 1.6868,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.27906976744186046,
|
| 1042 |
+
"grad_norm": 0.8220268286150999,
|
| 1043 |
+
"learning_rate": 0.000905596974384657,
|
| 1044 |
+
"loss": 1.8246,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.280968201233982,
|
| 1049 |
+
"grad_norm": 0.8599631648767855,
|
| 1050 |
+
"learning_rate": 0.0009036493518032172,
|
| 1051 |
+
"loss": 1.738,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.2828666350261035,
|
| 1056 |
+
"grad_norm": 1.0678364091016483,
|
| 1057 |
+
"learning_rate": 0.0009016839827386351,
|
| 1058 |
+
"loss": 1.7384,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.28476506881822494,
|
| 1063 |
+
"grad_norm": 1.8720283645771043,
|
| 1064 |
+
"learning_rate": 0.0008997009535985534,
|
| 1065 |
+
"loss": 1.8852,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.28666350261034645,
|
| 1070 |
+
"grad_norm": 1.1071369931471309,
|
| 1071 |
+
"learning_rate": 0.0008977003515670418,
|
| 1072 |
+
"loss": 1.7642,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.28856193640246797,
|
| 1077 |
+
"grad_norm": 0.9209513939367386,
|
| 1078 |
+
"learning_rate": 0.0008956822646007639,
|
| 1079 |
+
"loss": 1.8114,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.2904603701945895,
|
| 1084 |
+
"grad_norm": 0.4031852696574249,
|
| 1085 |
+
"learning_rate": 0.0008936467814251102,
|
| 1086 |
+
"loss": 1.749,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.292358803986711,
|
| 1091 |
+
"grad_norm": 0.482604929605048,
|
| 1092 |
+
"learning_rate": 0.0008915939915302969,
|
| 1093 |
+
"loss": 1.6956,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.29425723777883245,
|
| 1098 |
+
"grad_norm": 0.31688589446500226,
|
| 1099 |
+
"learning_rate": 0.000889523985167432,
|
| 1100 |
+
"loss": 1.7137,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.29615567157095396,
|
| 1105 |
+
"grad_norm": 0.2662426574337696,
|
| 1106 |
+
"learning_rate": 0.0008874368533445476,
|
| 1107 |
+
"loss": 1.716,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.29805410536307547,
|
| 1112 |
+
"grad_norm": 0.30925480824258833,
|
| 1113 |
+
"learning_rate": 0.0008853326878225978,
|
| 1114 |
+
"loss": 1.6941,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.299952539155197,
|
| 1119 |
+
"grad_norm": 0.3332600934262706,
|
| 1120 |
+
"learning_rate": 0.000883211581111425,
|
| 1121 |
+
"loss": 1.673,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.30185097294731844,
|
| 1126 |
+
"grad_norm": 0.4173857531935143,
|
| 1127 |
+
"learning_rate": 0.0008810736264656929,
|
| 1128 |
+
"loss": 1.677,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.30374940673943995,
|
| 1133 |
+
"grad_norm": 0.3528547476107368,
|
| 1134 |
+
"learning_rate": 0.0008789189178807862,
|
| 1135 |
+
"loss": 1.6878,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.30564784053156147,
|
| 1140 |
+
"grad_norm": 0.35241925687665077,
|
| 1141 |
+
"learning_rate": 0.0008767475500886777,
|
| 1142 |
+
"loss": 1.6895,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.307546274323683,
|
| 1147 |
+
"grad_norm": 0.3235829351373288,
|
| 1148 |
+
"learning_rate": 0.0008745596185537648,
|
| 1149 |
+
"loss": 1.6903,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.30944470811580443,
|
| 1154 |
+
"grad_norm": 0.2948128495705965,
|
| 1155 |
+
"learning_rate": 0.0008723552194686705,
|
| 1156 |
+
"loss": 1.6609,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.31134314190792595,
|
| 1161 |
+
"grad_norm": 0.36422304971890607,
|
| 1162 |
+
"learning_rate": 0.0008701344497500159,
|
| 1163 |
+
"loss": 1.6614,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.31324157570004746,
|
| 1168 |
+
"grad_norm": 0.2825304340578671,
|
| 1169 |
+
"learning_rate": 0.0008678974070341584,
|
| 1170 |
+
"loss": 1.6769,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.31514000949216897,
|
| 1175 |
+
"grad_norm": 0.26538936909594424,
|
| 1176 |
+
"learning_rate": 0.0008656441896728994,
|
| 1177 |
+
"loss": 1.6485,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.3170384432842905,
|
| 1182 |
+
"grad_norm": 0.31905462352993175,
|
| 1183 |
+
"learning_rate": 0.0008633748967291598,
|
| 1184 |
+
"loss": 1.6482,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.31893687707641194,
|
| 1189 |
+
"grad_norm": 0.39448662981546256,
|
| 1190 |
+
"learning_rate": 0.0008610896279726255,
|
| 1191 |
+
"loss": 1.6259,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.32083531086853345,
|
| 1196 |
+
"grad_norm": 0.31299399145300977,
|
| 1197 |
+
"learning_rate": 0.0008587884838753603,
|
| 1198 |
+
"loss": 1.6398,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.32273374466065496,
|
| 1203 |
+
"grad_norm": 0.3247022426656648,
|
| 1204 |
+
"learning_rate": 0.0008564715656073893,
|
| 1205 |
+
"loss": 1.6501,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.3246321784527765,
|
| 1210 |
+
"grad_norm": 0.2859709456607949,
|
| 1211 |
+
"learning_rate": 0.0008541389750322498,
|
| 1212 |
+
"loss": 1.6268,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.32653061224489793,
|
| 1217 |
+
"grad_norm": 0.28129496647987356,
|
| 1218 |
+
"learning_rate": 0.0008517908147025143,
|
| 1219 |
+
"loss": 1.6085,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.32842904603701945,
|
| 1224 |
+
"grad_norm": 0.3456433785224482,
|
| 1225 |
+
"learning_rate": 0.0008494271878552808,
|
| 1226 |
+
"loss": 1.6285,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.33032747982914096,
|
| 1231 |
+
"grad_norm": 0.49566297259294595,
|
| 1232 |
+
"learning_rate": 0.0008470481984076344,
|
| 1233 |
+
"loss": 1.6269,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.33222591362126247,
|
| 1238 |
+
"grad_norm": 0.34226796976284596,
|
| 1239 |
+
"learning_rate": 0.000844653950952078,
|
| 1240 |
+
"loss": 1.5934,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.334124347413384,
|
| 1245 |
+
"grad_norm": 0.2874518927936025,
|
| 1246 |
+
"learning_rate": 0.000842244550751935,
|
| 1247 |
+
"loss": 1.6233,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.33602278120550544,
|
| 1252 |
+
"grad_norm": 0.27909933392041014,
|
| 1253 |
+
"learning_rate": 0.0008398201037367201,
|
| 1254 |
+
"loss": 1.5986,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.33792121499762695,
|
| 1259 |
+
"grad_norm": 0.2755722277474001,
|
| 1260 |
+
"learning_rate": 0.0008373807164974832,
|
| 1261 |
+
"loss": 1.5831,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.33981964878974846,
|
| 1266 |
+
"grad_norm": 0.2647419405762948,
|
| 1267 |
+
"learning_rate": 0.0008349264962821219,
|
| 1268 |
+
"loss": 1.5928,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.34171808258187,
|
| 1273 |
+
"grad_norm": 0.34636953235762286,
|
| 1274 |
+
"learning_rate": 0.0008324575509906677,
|
| 1275 |
+
"loss": 1.5908,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.34361651637399143,
|
| 1280 |
+
"grad_norm": 0.3152387214237848,
|
| 1281 |
+
"learning_rate": 0.0008299739891705413,
|
| 1282 |
+
"loss": 1.5952,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.34551495016611294,
|
| 1287 |
+
"grad_norm": 0.2654394015701188,
|
| 1288 |
+
"learning_rate": 0.0008274759200117803,
|
| 1289 |
+
"loss": 1.5985,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.34741338395823446,
|
| 1294 |
+
"grad_norm": 0.6103144176291811,
|
| 1295 |
+
"learning_rate": 0.0008249634533422392,
|
| 1296 |
+
"loss": 1.5877,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.34931181775035597,
|
| 1301 |
+
"grad_norm": 0.3224755104739263,
|
| 1302 |
+
"learning_rate": 0.0008224366996227604,
|
| 1303 |
+
"loss": 1.5771,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.3512102515424775,
|
| 1308 |
+
"grad_norm": 0.5837396991238525,
|
| 1309 |
+
"learning_rate": 0.0008198957699423175,
|
| 1310 |
+
"loss": 1.5843,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.35310868533459894,
|
| 1315 |
+
"grad_norm": 0.32548897062944837,
|
| 1316 |
+
"learning_rate": 0.000817340776013132,
|
| 1317 |
+
"loss": 1.5851,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.35500711912672045,
|
| 1322 |
+
"grad_norm": 0.31829180968967913,
|
| 1323 |
+
"learning_rate": 0.0008147718301657612,
|
| 1324 |
+
"loss": 1.5942,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.35690555291884196,
|
| 1329 |
+
"grad_norm": 0.2731850877543467,
|
| 1330 |
+
"learning_rate": 0.0008121890453441602,
|
| 1331 |
+
"loss": 1.5694,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.3588039867109635,
|
| 1336 |
+
"grad_norm": 0.24810235775271594,
|
| 1337 |
+
"learning_rate": 0.0008095925351007156,
|
| 1338 |
+
"loss": 1.559,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.36070242050308493,
|
| 1343 |
+
"grad_norm": 0.28169507454726866,
|
| 1344 |
+
"learning_rate": 0.0008069824135912536,
|
| 1345 |
+
"loss": 1.5638,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.36260085429520644,
|
| 1350 |
+
"grad_norm": 0.24752038600316323,
|
| 1351 |
+
"learning_rate": 0.0008043587955700211,
|
| 1352 |
+
"loss": 1.5718,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.36449928808732796,
|
| 1357 |
+
"grad_norm": 0.27876047342351895,
|
| 1358 |
+
"learning_rate": 0.0008017217963846403,
|
| 1359 |
+
"loss": 1.5501,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.36639772187944947,
|
| 1364 |
+
"grad_norm": 0.40897383234845613,
|
| 1365 |
+
"learning_rate": 0.0007990715319710381,
|
| 1366 |
+
"loss": 1.5306,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.368296155671571,
|
| 1371 |
+
"grad_norm": 0.2858525738953562,
|
| 1372 |
+
"learning_rate": 0.0007964081188483476,
|
| 1373 |
+
"loss": 1.5283,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.37019458946369244,
|
| 1378 |
+
"grad_norm": 0.26144274136184015,
|
| 1379 |
+
"learning_rate": 0.0007937316741137871,
|
| 1380 |
+
"loss": 1.5446,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.37209302325581395,
|
| 1385 |
+
"grad_norm": 0.32273246789256727,
|
| 1386 |
+
"learning_rate": 0.0007910423154375101,
|
| 1387 |
+
"loss": 1.5672,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.37399145704793546,
|
| 1392 |
+
"grad_norm": 0.3490759385530748,
|
| 1393 |
+
"learning_rate": 0.0007883401610574337,
|
| 1394 |
+
"loss": 1.5401,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.375889890840057,
|
| 1399 |
+
"grad_norm": 0.28213557009744067,
|
| 1400 |
+
"learning_rate": 0.0007856253297740383,
|
| 1401 |
+
"loss": 1.5533,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.37778832463217843,
|
| 1406 |
+
"grad_norm": 0.33445422182184414,
|
| 1407 |
+
"learning_rate": 0.0007828979409451468,
|
| 1408 |
+
"loss": 1.5361,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.37968675842429994,
|
| 1413 |
+
"grad_norm": 0.3535301287121482,
|
| 1414 |
+
"learning_rate": 0.0007801581144806751,
|
| 1415 |
+
"loss": 1.5189,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.38158519221642145,
|
| 1420 |
+
"grad_norm": 0.3391675710324959,
|
| 1421 |
+
"learning_rate": 0.0007774059708373606,
|
| 1422 |
+
"loss": 1.5301,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.38348362600854297,
|
| 1427 |
+
"grad_norm": 0.6452174423374145,
|
| 1428 |
+
"learning_rate": 0.0007746416310134679,
|
| 1429 |
+
"loss": 1.5214,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.3853820598006645,
|
| 1434 |
+
"grad_norm": 0.43341843816494785,
|
| 1435 |
+
"learning_rate": 0.0007718652165434664,
|
| 1436 |
+
"loss": 1.5235,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.38728049359278593,
|
| 1441 |
+
"grad_norm": 0.3003373383959457,
|
| 1442 |
+
"learning_rate": 0.0007690768494926897,
|
| 1443 |
+
"loss": 1.5063,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.38917892738490745,
|
| 1448 |
+
"grad_norm": 0.3078153548021384,
|
| 1449 |
+
"learning_rate": 0.000766276652451967,
|
| 1450 |
+
"loss": 1.5174,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.39107736117702896,
|
| 1455 |
+
"grad_norm": 0.4798286034200326,
|
| 1456 |
+
"learning_rate": 0.0007634647485322347,
|
| 1457 |
+
"loss": 1.4958,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.39297579496915047,
|
| 1462 |
+
"grad_norm": 0.2962615700403029,
|
| 1463 |
+
"learning_rate": 0.0007606412613591228,
|
| 1464 |
+
"loss": 1.5187,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.39487422876127193,
|
| 1469 |
+
"grad_norm": 0.3904932019460365,
|
| 1470 |
+
"learning_rate": 0.0007578063150675206,
|
| 1471 |
+
"loss": 1.4828,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.39677266255339344,
|
| 1476 |
+
"grad_norm": 0.31811472424437215,
|
| 1477 |
+
"learning_rate": 0.0007549600342961183,
|
| 1478 |
+
"loss": 1.4752,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.39867109634551495,
|
| 1483 |
+
"grad_norm": 0.4803758991042241,
|
| 1484 |
+
"learning_rate": 0.0007521025441819278,
|
| 1485 |
+
"loss": 1.4969,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.40056953013763646,
|
| 1490 |
+
"grad_norm": 0.472851525192678,
|
| 1491 |
+
"learning_rate": 0.0007492339703547808,
|
| 1492 |
+
"loss": 1.492,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.402467963929758,
|
| 1497 |
+
"grad_norm": 0.3908915345807985,
|
| 1498 |
+
"learning_rate": 0.0007463544389318058,
|
| 1499 |
+
"loss": 1.4682,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.40436639772187943,
|
| 1504 |
+
"grad_norm": 0.3886806997641989,
|
| 1505 |
+
"learning_rate": 0.0007434640765118824,
|
| 1506 |
+
"loss": 1.4759,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.40626483151400095,
|
| 1511 |
+
"grad_norm": 0.5234433759967211,
|
| 1512 |
+
"learning_rate": 0.0007405630101700769,
|
| 1513 |
+
"loss": 1.4779,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.40816326530612246,
|
| 1518 |
+
"grad_norm": 0.3607186516700363,
|
| 1519 |
+
"learning_rate": 0.0007376513674520537,
|
| 1520 |
+
"loss": 1.4894,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.41006169909824397,
|
| 1525 |
+
"grad_norm": 0.3422699133158823,
|
| 1526 |
+
"learning_rate": 0.0007347292763684691,
|
| 1527 |
+
"loss": 1.4713,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.4119601328903654,
|
| 1532 |
+
"grad_norm": 0.456283403837333,
|
| 1533 |
+
"learning_rate": 0.000731796865389343,
|
| 1534 |
+
"loss": 1.4786,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.41385856668248694,
|
| 1539 |
+
"grad_norm": 0.29907259245021056,
|
| 1540 |
+
"learning_rate": 0.0007288542634384101,
|
| 1541 |
+
"loss": 1.4428,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.41575700047460845,
|
| 1546 |
+
"grad_norm": 0.4065968463089126,
|
| 1547 |
+
"learning_rate": 0.0007259015998874521,
|
| 1548 |
+
"loss": 1.4551,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.41765543426672996,
|
| 1553 |
+
"grad_norm": 0.2851166065497382,
|
| 1554 |
+
"learning_rate": 0.0007229390045506107,
|
| 1555 |
+
"loss": 1.4726,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.4195538680588514,
|
| 1560 |
+
"grad_norm": 0.34495921420404607,
|
| 1561 |
+
"learning_rate": 0.0007199666076786786,
|
| 1562 |
+
"loss": 1.4491,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.42145230185097293,
|
| 1567 |
+
"grad_norm": 0.2657939111249429,
|
| 1568 |
+
"learning_rate": 0.0007169845399533742,
|
| 1569 |
+
"loss": 1.4362,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.42335073564309444,
|
| 1574 |
+
"grad_norm": 0.4937795186011948,
|
| 1575 |
+
"learning_rate": 0.0007139929324815965,
|
| 1576 |
+
"loss": 1.4407,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.42524916943521596,
|
| 1581 |
+
"grad_norm": 0.4407673687752531,
|
| 1582 |
+
"learning_rate": 0.0007109919167896597,
|
| 1583 |
+
"loss": 1.4319,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.42714760322733747,
|
| 1588 |
+
"grad_norm": 0.5300468580153148,
|
| 1589 |
+
"learning_rate": 0.0007079816248175114,
|
| 1590 |
+
"loss": 1.4275,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.4290460370194589,
|
| 1595 |
+
"grad_norm": 0.3241335090242318,
|
| 1596 |
+
"learning_rate": 0.000704962188912932,
|
| 1597 |
+
"loss": 1.4426,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.43094447081158044,
|
| 1602 |
+
"grad_norm": 0.3968224596845394,
|
| 1603 |
+
"learning_rate": 0.0007019337418257159,
|
| 1604 |
+
"loss": 1.4398,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.43284290460370195,
|
| 1609 |
+
"grad_norm": 0.3781038122854968,
|
| 1610 |
+
"learning_rate": 0.0006988964167018346,
|
| 1611 |
+
"loss": 1.4298,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.43474133839582346,
|
| 1616 |
+
"grad_norm": 0.2842174258844352,
|
| 1617 |
+
"learning_rate": 0.0006958503470775836,
|
| 1618 |
+
"loss": 1.4358,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.4366397721879449,
|
| 1623 |
+
"grad_norm": 0.5316938383381906,
|
| 1624 |
+
"learning_rate": 0.0006927956668737115,
|
| 1625 |
+
"loss": 1.4152,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.43853820598006643,
|
| 1630 |
+
"grad_norm": 0.26129078645372833,
|
| 1631 |
+
"learning_rate": 0.000689732510389531,
|
| 1632 |
+
"loss": 1.4117,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.44043663977218794,
|
| 1637 |
+
"grad_norm": 0.45097710133639385,
|
| 1638 |
+
"learning_rate": 0.0006866610122970162,
|
| 1639 |
+
"loss": 1.3814,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.44233507356430946,
|
| 1644 |
+
"grad_norm": 0.34768384716485046,
|
| 1645 |
+
"learning_rate": 0.0006835813076348805,
|
| 1646 |
+
"loss": 1.409,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.44423350735643097,
|
| 1651 |
+
"grad_norm": 0.6973633265305714,
|
| 1652 |
+
"learning_rate": 0.0006804935318026396,
|
| 1653 |
+
"loss": 1.4211,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.4461319411485524,
|
| 1658 |
+
"grad_norm": 0.4342853868362056,
|
| 1659 |
+
"learning_rate": 0.0006773978205546597,
|
| 1660 |
+
"loss": 1.4037,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.44803037494067394,
|
| 1665 |
+
"grad_norm": 0.46738592482568664,
|
| 1666 |
+
"learning_rate": 0.0006742943099941876,
|
| 1667 |
+
"loss": 1.4309,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.44992880873279545,
|
| 1672 |
+
"grad_norm": 0.25387391273809895,
|
| 1673 |
+
"learning_rate": 0.000671183136567368,
|
| 1674 |
+
"loss": 1.3923,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.45182724252491696,
|
| 1679 |
+
"grad_norm": 0.46112835773656796,
|
| 1680 |
+
"learning_rate": 0.0006680644370572444,
|
| 1681 |
+
"loss": 1.4028,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.4537256763170384,
|
| 1686 |
+
"grad_norm": 0.2551607349467472,
|
| 1687 |
+
"learning_rate": 0.0006649383485777449,
|
| 1688 |
+
"loss": 1.3815,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.45562411010915993,
|
| 1693 |
+
"grad_norm": 0.37310127138077576,
|
| 1694 |
+
"learning_rate": 0.0006618050085676546,
|
| 1695 |
+
"loss": 1.3911,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.45752254390128144,
|
| 1700 |
+
"grad_norm": 0.4254847053830724,
|
| 1701 |
+
"learning_rate": 0.0006586645547845729,
|
| 1702 |
+
"loss": 1.3929,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.45942097769340295,
|
| 1707 |
+
"grad_norm": 0.7687175874517974,
|
| 1708 |
+
"learning_rate": 0.0006555171252988568,
|
| 1709 |
+
"loss": 1.3733,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.46131941148552447,
|
| 1714 |
+
"grad_norm": 0.3146929888691999,
|
| 1715 |
+
"learning_rate": 0.0006523628584875507,
|
| 1716 |
+
"loss": 1.3711,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.4632178452776459,
|
| 1721 |
+
"grad_norm": 0.27183821399250013,
|
| 1722 |
+
"learning_rate": 0.0006492018930283026,
|
| 1723 |
+
"loss": 1.3455,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.46511627906976744,
|
| 1728 |
+
"grad_norm": 0.31617364418439614,
|
| 1729 |
+
"learning_rate": 0.000646034367893268,
|
| 1730 |
+
"loss": 1.3619,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.46701471286188895,
|
| 1735 |
+
"grad_norm": 0.3833518263216662,
|
| 1736 |
+
"learning_rate": 0.0006428604223429979,
|
| 1737 |
+
"loss": 1.3452,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.46891314665401046,
|
| 1742 |
+
"grad_norm": 0.3267810331058041,
|
| 1743 |
+
"learning_rate": 0.0006396801959203186,
|
| 1744 |
+
"loss": 1.3627,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.4708115804461319,
|
| 1749 |
+
"grad_norm": 0.34041494057204974,
|
| 1750 |
+
"learning_rate": 0.0006364938284441949,
|
| 1751 |
+
"loss": 1.3418,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.47271001423825343,
|
| 1756 |
+
"grad_norm": 0.42237475709823624,
|
| 1757 |
+
"learning_rate": 0.0006333014600035838,
|
| 1758 |
+
"loss": 1.3594,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.47460844803037494,
|
| 1763 |
+
"grad_norm": 0.4449414736784289,
|
| 1764 |
+
"learning_rate": 0.0006301032309512754,
|
| 1765 |
+
"loss": 1.3293,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.47650688182249645,
|
| 1770 |
+
"grad_norm": 0.3046879446109181,
|
| 1771 |
+
"learning_rate": 0.0006268992818977221,
|
| 1772 |
+
"loss": 1.3438,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.47840531561461797,
|
| 1777 |
+
"grad_norm": 0.3963369420982114,
|
| 1778 |
+
"learning_rate": 0.0006236897537048566,
|
| 1779 |
+
"loss": 1.3418,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.4803037494067394,
|
| 1784 |
+
"grad_norm": 0.449251264603866,
|
| 1785 |
+
"learning_rate": 0.0006204747874798993,
|
| 1786 |
+
"loss": 1.3443,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.48220218319886093,
|
| 1791 |
+
"grad_norm": 0.30313300002195165,
|
| 1792 |
+
"learning_rate": 0.0006172545245691538,
|
| 1793 |
+
"loss": 1.3414,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.48410061699098245,
|
| 1798 |
+
"grad_norm": 0.33417145989869673,
|
| 1799 |
+
"learning_rate": 0.0006140291065517931,
|
| 1800 |
+
"loss": 1.3179,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.48599905078310396,
|
| 1805 |
+
"grad_norm": 0.3904784303862518,
|
| 1806 |
+
"learning_rate": 0.0006107986752336357,
|
| 1807 |
+
"loss": 1.3411,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.4878974845752254,
|
| 1812 |
+
"grad_norm": 0.3222639679317221,
|
| 1813 |
+
"learning_rate": 0.0006075633726409091,
|
| 1814 |
+
"loss": 1.3086,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.4897959183673469,
|
| 1819 |
+
"grad_norm": 0.3288048359935024,
|
| 1820 |
+
"learning_rate": 0.0006043233410140076,
|
| 1821 |
+
"loss": 1.3305,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.49169435215946844,
|
| 1826 |
+
"grad_norm": 0.2937886282400547,
|
| 1827 |
+
"learning_rate": 0.0006010787228012384,
|
| 1828 |
+
"loss": 1.3029,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.49359278595158995,
|
| 1833 |
+
"grad_norm": 0.29213405752561167,
|
| 1834 |
+
"learning_rate": 0.0005978296606525572,
|
| 1835 |
+
"loss": 1.3161,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.49549121974371146,
|
| 1840 |
+
"grad_norm": 0.34322670276805756,
|
| 1841 |
+
"learning_rate": 0.0005945762974132986,
|
| 1842 |
+
"loss": 1.3281,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.4973896535358329,
|
| 1847 |
+
"grad_norm": 0.28223010113720515,
|
| 1848 |
+
"learning_rate": 0.0005913187761178951,
|
| 1849 |
+
"loss": 1.2943,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.49928808732795443,
|
| 1854 |
+
"grad_norm": 0.29364231136684454,
|
| 1855 |
+
"learning_rate": 0.0005880572399835881,
|
| 1856 |
+
"loss": 1.2988,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.5011865211200759,
|
| 1861 |
+
"grad_norm": 0.3396537753038387,
|
| 1862 |
+
"learning_rate": 0.0005847918324041324,
|
| 1863 |
+
"loss": 1.3053,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.5030849549121974,
|
| 1868 |
+
"grad_norm": 0.3515005890654536,
|
| 1869 |
+
"learning_rate": 0.0005815226969434903,
|
| 1870 |
+
"loss": 1.2881,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.5049833887043189,
|
| 1875 |
+
"grad_norm": 0.32308285051579544,
|
| 1876 |
+
"learning_rate": 0.0005782499773295219,
|
| 1877 |
+
"loss": 1.31,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.5068818224964404,
|
| 1882 |
+
"grad_norm": 0.32615560955136413,
|
| 1883 |
+
"learning_rate": 0.0005749738174476639,
|
| 1884 |
+
"loss": 1.2659,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.5087802562885619,
|
| 1889 |
+
"grad_norm": 0.3566565149320192,
|
| 1890 |
+
"learning_rate": 0.0005716943613346059,
|
| 1891 |
+
"loss": 1.2746,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.5106786900806835,
|
| 1896 |
+
"grad_norm": 0.4244945485487363,
|
| 1897 |
+
"learning_rate": 0.0005684117531719551,
|
| 1898 |
+
"loss": 1.2741,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.512577123872805,
|
| 1903 |
+
"grad_norm": 0.31519325513997737,
|
| 1904 |
+
"learning_rate": 0.0005651261372799002,
|
| 1905 |
+
"loss": 1.2773,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.5144755576649265,
|
| 1910 |
+
"grad_norm": 0.4339192027972204,
|
| 1911 |
+
"learning_rate": 0.0005618376581108647,
|
| 1912 |
+
"loss": 1.2669,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.516373991457048,
|
| 1917 |
+
"grad_norm": 0.3419108710774503,
|
| 1918 |
+
"learning_rate": 0.0005585464602431556,
|
| 1919 |
+
"loss": 1.2508,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.5182724252491694,
|
| 1924 |
+
"grad_norm": 0.38978151919816467,
|
| 1925 |
+
"learning_rate": 0.0005552526883746087,
|
| 1926 |
+
"loss": 1.2745,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.5201708590412909,
|
| 1931 |
+
"grad_norm": 0.4895534717498465,
|
| 1932 |
+
"learning_rate": 0.0005519564873162257,
|
| 1933 |
+
"loss": 1.2461,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.5220692928334124,
|
| 1938 |
+
"grad_norm": 0.4461111757267344,
|
| 1939 |
+
"learning_rate": 0.0005486580019858075,
|
| 1940 |
+
"loss": 1.2704,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.5239677266255339,
|
| 1945 |
+
"grad_norm": 0.32031503363419567,
|
| 1946 |
+
"learning_rate": 0.0005453573774015837,
|
| 1947 |
+
"loss": 1.2638,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.5258661604176554,
|
| 1952 |
+
"grad_norm": 0.3643504769224511,
|
| 1953 |
+
"learning_rate": 0.0005420547586758364,
|
| 1954 |
+
"loss": 1.2595,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.527764594209777,
|
| 1959 |
+
"grad_norm": 0.32810368902662324,
|
| 1960 |
+
"learning_rate": 0.0005387502910085201,
|
| 1961 |
+
"loss": 1.2326,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.5296630280018985,
|
| 1966 |
+
"grad_norm": 0.3336242397751466,
|
| 1967 |
+
"learning_rate": 0.0005354441196808778,
|
| 1968 |
+
"loss": 1.2553,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.53156146179402,
|
| 1973 |
+
"grad_norm": 0.4348846619107913,
|
| 1974 |
+
"learning_rate": 0.000532136390049055,
|
| 1975 |
+
"loss": 1.2457,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.5334598955861415,
|
| 1980 |
+
"grad_norm": 0.3940477704170248,
|
| 1981 |
+
"learning_rate": 0.0005288272475377078,
|
| 1982 |
+
"loss": 1.2286,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.5353583293782629,
|
| 1987 |
+
"grad_norm": 0.3034212811156197,
|
| 1988 |
+
"learning_rate": 0.0005255168376336094,
|
| 1989 |
+
"loss": 1.209,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.5372567631703844,
|
| 1994 |
+
"grad_norm": 0.4307385167019054,
|
| 1995 |
+
"learning_rate": 0.0005222053058792543,
|
| 1996 |
+
"loss": 1.2476,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.5391551969625059,
|
| 2001 |
+
"grad_norm": 0.34543357053441875,
|
| 2002 |
+
"learning_rate": 0.0005188927978664594,
|
| 2003 |
+
"loss": 1.2149,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.5410536307546274,
|
| 2008 |
+
"grad_norm": 0.3611941956078568,
|
| 2009 |
+
"learning_rate": 0.0005155794592299626,
|
| 2010 |
+
"loss": 1.2407,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.5429520645467489,
|
| 2015 |
+
"grad_norm": 0.3375414745847685,
|
| 2016 |
+
"learning_rate": 0.0005122654356410205,
|
| 2017 |
+
"loss": 1.2412,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.5448504983388704,
|
| 2022 |
+
"grad_norm": 0.286479882843699,
|
| 2023 |
+
"learning_rate": 0.0005089508728010033,
|
| 2024 |
+
"loss": 1.2167,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.546748932130992,
|
| 2029 |
+
"grad_norm": 0.3385247369838371,
|
| 2030 |
+
"learning_rate": 0.0005056359164349902,
|
| 2031 |
+
"loss": 1.231,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.5486473659231135,
|
| 2036 |
+
"grad_norm": 0.30665827143092833,
|
| 2037 |
+
"learning_rate": 0.000502320712285361,
|
| 2038 |
+
"loss": 1.1893,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.550545799715235,
|
| 2043 |
+
"grad_norm": 0.3668246619215131,
|
| 2044 |
+
"learning_rate": 0.0004990054061053896,
|
| 2045 |
+
"loss": 1.1876,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.5524442335073564,
|
| 2050 |
+
"grad_norm": 0.3886963786863651,
|
| 2051 |
+
"learning_rate": 0.0004956901436528358,
|
| 2052 |
+
"loss": 1.1897,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.5543426672994779,
|
| 2057 |
+
"grad_norm": 0.33560559480266355,
|
| 2058 |
+
"learning_rate": 0.0004923750706835371,
|
| 2059 |
+
"loss": 1.2048,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.5562411010915994,
|
| 2064 |
+
"grad_norm": 0.35331444970790565,
|
| 2065 |
+
"learning_rate": 0.0004890603329449997,
|
| 2066 |
+
"loss": 1.1893,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.5581395348837209,
|
| 2071 |
+
"grad_norm": 0.31797300187844246,
|
| 2072 |
+
"learning_rate": 0.000485746076169992,
|
| 2073 |
+
"loss": 1.1653,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.5600379686758424,
|
| 2078 |
+
"grad_norm": 0.3697991594252187,
|
| 2079 |
+
"learning_rate": 0.00048243244607013654,
|
| 2080 |
+
"loss": 1.2023,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.561936402467964,
|
| 2085 |
+
"grad_norm": 0.5326439145387593,
|
| 2086 |
+
"learning_rate": 0.0004791195883295036,
|
| 2087 |
+
"loss": 1.1813,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.5638348362600855,
|
| 2092 |
+
"grad_norm": 0.371212386610569,
|
| 2093 |
+
"learning_rate": 0.0004758076485982076,
|
| 2094 |
+
"loss": 1.1702,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.565733270052207,
|
| 2099 |
+
"grad_norm": 0.3968881977650288,
|
| 2100 |
+
"learning_rate": 0.00047249677248600185,
|
| 2101 |
+
"loss": 1.1673,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.5676317038443285,
|
| 2106 |
+
"grad_norm": 2.9338387571685405,
|
| 2107 |
+
"learning_rate": 0.0004691871055558776,
|
| 2108 |
+
"loss": 1.1908,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.5695301376364499,
|
| 2113 |
+
"grad_norm": 0.49847319164775866,
|
| 2114 |
+
"learning_rate": 0.00046587879331766457,
|
| 2115 |
+
"loss": 1.1864,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.5714285714285714,
|
| 2120 |
+
"grad_norm": 0.4078724706306019,
|
| 2121 |
+
"learning_rate": 0.000462571981221633,
|
| 2122 |
+
"loss": 1.1663,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.5733270052206929,
|
| 2127 |
+
"grad_norm": 0.3228249575712461,
|
| 2128 |
+
"learning_rate": 0.0004592668146520994,
|
| 2129 |
+
"loss": 1.1702,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.5752254390128144,
|
| 2134 |
+
"grad_norm": 0.37770036266775847,
|
| 2135 |
+
"learning_rate": 0.00045596343892103443,
|
| 2136 |
+
"loss": 1.176,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.5771238728049359,
|
| 2141 |
+
"grad_norm": 0.3148160726043947,
|
| 2142 |
+
"learning_rate": 0.00045266199926167485,
|
| 2143 |
+
"loss": 1.1747,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.5790223065970574,
|
| 2148 |
+
"grad_norm": 0.3626958842571658,
|
| 2149 |
+
"learning_rate": 0.00044936264082213724,
|
| 2150 |
+
"loss": 1.143,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.580920740389179,
|
| 2155 |
+
"grad_norm": 0.3613417080127485,
|
| 2156 |
+
"learning_rate": 0.00044606550865903725,
|
| 2157 |
+
"loss": 1.1523,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.5828191741813005,
|
| 2162 |
+
"grad_norm": 0.36620946050669057,
|
| 2163 |
+
"learning_rate": 0.0004427707477311123,
|
| 2164 |
+
"loss": 1.1732,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.584717607973422,
|
| 2169 |
+
"grad_norm": 0.3418418756481974,
|
| 2170 |
+
"learning_rate": 0.000439478502892848,
|
| 2171 |
+
"loss": 1.1616,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.5866160417655434,
|
| 2176 |
+
"grad_norm": 0.33600686205594626,
|
| 2177 |
+
"learning_rate": 0.0004361889188881102,
|
| 2178 |
+
"loss": 1.1158,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.5885144755576649,
|
| 2183 |
+
"grad_norm": 0.3764722521246889,
|
| 2184 |
+
"learning_rate": 0.0004329021403437802,
|
| 2185 |
+
"loss": 1.154,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.5904129093497864,
|
| 2190 |
+
"grad_norm": 0.34240802002457055,
|
| 2191 |
+
"learning_rate": 0.000429618311763398,
|
| 2192 |
+
"loss": 1.139,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.5923113431419079,
|
| 2197 |
+
"grad_norm": 0.2902338020123924,
|
| 2198 |
+
"learning_rate": 0.00042633757752080727,
|
| 2199 |
+
"loss": 1.134,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.5942097769340294,
|
| 2204 |
+
"grad_norm": 0.3752689277444726,
|
| 2205 |
+
"learning_rate": 0.00042306008185380927,
|
| 2206 |
+
"loss": 1.1343,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.5961082107261509,
|
| 2211 |
+
"grad_norm": 0.31329902926972597,
|
| 2212 |
+
"learning_rate": 0.0004197859688578207,
|
| 2213 |
+
"loss": 1.1016,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.5980066445182725,
|
| 2218 |
+
"grad_norm": 0.3402332223232388,
|
| 2219 |
+
"learning_rate": 0.00041651538247953904,
|
| 2220 |
+
"loss": 1.1052,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.599905078310394,
|
| 2225 |
+
"grad_norm": 0.3781151786788464,
|
| 2226 |
+
"learning_rate": 0.0004132484665106135,
|
| 2227 |
+
"loss": 1.1191,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.6018035121025154,
|
| 2232 |
+
"grad_norm": 0.34137800960299325,
|
| 2233 |
+
"learning_rate": 0.0004099853645813235,
|
| 2234 |
+
"loss": 1.0947,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.6037019458946369,
|
| 2239 |
+
"grad_norm": 0.3136113067101357,
|
| 2240 |
+
"learning_rate": 0.00040672622015426363,
|
| 2241 |
+
"loss": 1.0856,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.6056003796867584,
|
| 2246 |
+
"grad_norm": 0.3358747164703956,
|
| 2247 |
+
"learning_rate": 0.00040347117651803703,
|
| 2248 |
+
"loss": 1.0948,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.6074988134788799,
|
| 2253 |
+
"grad_norm": 0.3020729530283223,
|
| 2254 |
+
"learning_rate": 0.00040022037678095454,
|
| 2255 |
+
"loss": 1.0802,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.6093972472710014,
|
| 2260 |
+
"grad_norm": 0.3904330186630717,
|
| 2261 |
+
"learning_rate": 0.00039697396386474394,
|
| 2262 |
+
"loss": 1.127,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.6112956810631229,
|
| 2267 |
+
"grad_norm": 0.4027533805204724,
|
| 2268 |
+
"learning_rate": 0.0003937320804982659,
|
| 2269 |
+
"loss": 1.1042,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.6131941148552444,
|
| 2274 |
+
"grad_norm": 0.3743040482483265,
|
| 2275 |
+
"learning_rate": 0.00039049486921123876,
|
| 2276 |
+
"loss": 1.0907,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.615092548647366,
|
| 2281 |
+
"grad_norm": 0.38060393763386724,
|
| 2282 |
+
"learning_rate": 0.000387262472327973,
|
| 2283 |
+
"loss": 1.1051,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.6169909824394875,
|
| 2288 |
+
"grad_norm": 0.34005085980267563,
|
| 2289 |
+
"learning_rate": 0.00038403503196111265,
|
| 2290 |
+
"loss": 1.0647,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.6188894162316089,
|
| 2295 |
+
"grad_norm": 0.30724882728663455,
|
| 2296 |
+
"learning_rate": 0.0003808126900053887,
|
| 2297 |
+
"loss": 1.0771,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.6207878500237304,
|
| 2302 |
+
"grad_norm": 0.306237705852109,
|
| 2303 |
+
"learning_rate": 0.0003775955881313797,
|
| 2304 |
+
"loss": 1.0831,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.6226862838158519,
|
| 2309 |
+
"grad_norm": 0.4486821484171895,
|
| 2310 |
+
"learning_rate": 0.0003743838677792833,
|
| 2311 |
+
"loss": 1.0767,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.6245847176079734,
|
| 2316 |
+
"grad_norm": 0.35675874358918835,
|
| 2317 |
+
"learning_rate": 0.0003711776701526982,
|
| 2318 |
+
"loss": 1.0841,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.6264831514000949,
|
| 2323 |
+
"grad_norm": 0.31338528181698055,
|
| 2324 |
+
"learning_rate": 0.00036797713621241615,
|
| 2325 |
+
"loss": 1.0851,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.6283815851922164,
|
| 2330 |
+
"grad_norm": 0.3687470434111214,
|
| 2331 |
+
"learning_rate": 0.000364782406670224,
|
| 2332 |
+
"loss": 1.0707,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.6302800189843379,
|
| 2337 |
+
"grad_norm": 0.31784199572514304,
|
| 2338 |
+
"learning_rate": 0.0003615936219827176,
|
| 2339 |
+
"loss": 1.0606,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.6321784527764595,
|
| 2344 |
+
"grad_norm": 0.42249881083466695,
|
| 2345 |
+
"learning_rate": 0.00035841092234512723,
|
| 2346 |
+
"loss": 1.0725,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.634076886568581,
|
| 2351 |
+
"grad_norm": 0.3459092803870884,
|
| 2352 |
+
"learning_rate": 0.0003552344476851531,
|
| 2353 |
+
"loss": 1.0574,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.6359753203607024,
|
| 2358 |
+
"grad_norm": 0.31341808911866365,
|
| 2359 |
+
"learning_rate": 0.00035206433765681334,
|
| 2360 |
+
"loss": 1.0399,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.6378737541528239,
|
| 2365 |
+
"grad_norm": 0.3619705130805933,
|
| 2366 |
+
"learning_rate": 0.00034890073163430503,
|
| 2367 |
+
"loss": 1.0295,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.6397721879449454,
|
| 2372 |
+
"grad_norm": 0.302930290181787,
|
| 2373 |
+
"learning_rate": 0.00034574376870587535,
|
| 2374 |
+
"loss": 1.0302,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.6416706217370669,
|
| 2379 |
+
"grad_norm": 0.32319117475207404,
|
| 2380 |
+
"learning_rate": 0.00034259358766770766,
|
| 2381 |
+
"loss": 1.0457,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.6435690555291884,
|
| 2386 |
+
"grad_norm": 0.35131830575911627,
|
| 2387 |
+
"learning_rate": 0.0003394503270178185,
|
| 2388 |
+
"loss": 1.0527,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.6454674893213099,
|
| 2393 |
+
"grad_norm": 0.35565194703093944,
|
| 2394 |
+
"learning_rate": 0.0003363141249499696,
|
| 2395 |
+
"loss": 1.0271,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.6473659231134314,
|
| 2400 |
+
"grad_norm": 0.3631695241713009,
|
| 2401 |
+
"learning_rate": 0.00033318511934759046,
|
| 2402 |
+
"loss": 1.044,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.649264356905553,
|
| 2407 |
+
"grad_norm": 0.3236231297783939,
|
| 2408 |
+
"learning_rate": 0.0003300634477777179,
|
| 2409 |
+
"loss": 1.0102,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.6511627906976745,
|
| 2414 |
+
"grad_norm": 0.3213681061725248,
|
| 2415 |
+
"learning_rate": 0.00032694924748494713,
|
| 2416 |
+
"loss": 1.0083,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.6530612244897959,
|
| 2421 |
+
"grad_norm": 0.6068864104166338,
|
| 2422 |
+
"learning_rate": 0.00032384265538539783,
|
| 2423 |
+
"loss": 0.9907,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.6549596582819174,
|
| 2428 |
+
"grad_norm": 0.41210521486813323,
|
| 2429 |
+
"learning_rate": 0.0003207438080606949,
|
| 2430 |
+
"loss": 1.0233,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.6568580920740389,
|
| 2435 |
+
"grad_norm": 0.46261139879081853,
|
| 2436 |
+
"learning_rate": 0.00031765284175196324,
|
| 2437 |
+
"loss": 1.0167,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.6587565258661604,
|
| 2442 |
+
"grad_norm": 0.3436327494287044,
|
| 2443 |
+
"learning_rate": 0.0003145698923538384,
|
| 2444 |
+
"loss": 1.0282,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.6606549596582819,
|
| 2449 |
+
"grad_norm": 0.35619308018504636,
|
| 2450 |
+
"learning_rate": 0.00031149509540849156,
|
| 2451 |
+
"loss": 0.99,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.6625533934504034,
|
| 2456 |
+
"grad_norm": 0.34309970855605526,
|
| 2457 |
+
"learning_rate": 0.0003084285860996704,
|
| 2458 |
+
"loss": 1.015,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.6644518272425249,
|
| 2463 |
+
"grad_norm": 0.393508338674986,
|
| 2464 |
+
"learning_rate": 0.0003053704992467558,
|
| 2465 |
+
"loss": 1.0116,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.6663502610346465,
|
| 2470 |
+
"grad_norm": 0.3397492909000668,
|
| 2471 |
+
"learning_rate": 0.0003023209692988349,
|
| 2472 |
+
"loss": 0.9983,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.668248694826768,
|
| 2477 |
+
"grad_norm": 0.30189558040021247,
|
| 2478 |
+
"learning_rate": 0.0002992801303287892,
|
| 2479 |
+
"loss": 1.0073,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.6701471286188894,
|
| 2484 |
+
"grad_norm": 0.33390373051452854,
|
| 2485 |
+
"learning_rate": 0.00029624811602740105,
|
| 2486 |
+
"loss": 0.9931,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.6720455624110109,
|
| 2491 |
+
"grad_norm": 0.3565911003751968,
|
| 2492 |
+
"learning_rate": 0.0002932250596974747,
|
| 2493 |
+
"loss": 0.9794,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.6739439962031324,
|
| 2498 |
+
"grad_norm": 0.30184049203060254,
|
| 2499 |
+
"learning_rate": 0.00029021109424797706,
|
| 2500 |
+
"loss": 0.9853,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.6758424299952539,
|
| 2505 |
+
"grad_norm": 0.2996692988493312,
|
| 2506 |
+
"learning_rate": 0.00028720635218819313,
|
| 2507 |
+
"loss": 0.9716,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.6777408637873754,
|
| 2512 |
+
"grad_norm": 0.3158967543323723,
|
| 2513 |
+
"learning_rate": 0.00028421096562190087,
|
| 2514 |
+
"loss": 0.971,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.6796392975794969,
|
| 2519 |
+
"grad_norm": 0.3259541143986902,
|
| 2520 |
+
"learning_rate": 0.00028122506624156287,
|
| 2521 |
+
"loss": 0.9667,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.6815377313716184,
|
| 2526 |
+
"grad_norm": 0.3050723057723714,
|
| 2527 |
+
"learning_rate": 0.00027824878532253675,
|
| 2528 |
+
"loss": 0.9698,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.68343616516374,
|
| 2533 |
+
"grad_norm": 0.3225081325517291,
|
| 2534 |
+
"learning_rate": 0.0002752822537173033,
|
| 2535 |
+
"loss": 0.9627,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.6853345989558615,
|
| 2540 |
+
"grad_norm": 0.3337812239040389,
|
| 2541 |
+
"learning_rate": 0.00027232560184971434,
|
| 2542 |
+
"loss": 0.9673,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.6872330327479829,
|
| 2547 |
+
"grad_norm": 0.4504964673114019,
|
| 2548 |
+
"learning_rate": 0.00026937895970925794,
|
| 2549 |
+
"loss": 0.9504,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.6891314665401044,
|
| 2554 |
+
"grad_norm": 0.6271532351806949,
|
| 2555 |
+
"learning_rate": 0.00026644245684534317,
|
| 2556 |
+
"loss": 0.9655,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.6910299003322259,
|
| 2561 |
+
"grad_norm": 0.42681861959807327,
|
| 2562 |
+
"learning_rate": 0.00026351622236160487,
|
| 2563 |
+
"loss": 0.9418,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.6929283341243474,
|
| 2568 |
+
"grad_norm": 0.4353835236159284,
|
| 2569 |
+
"learning_rate": 0.00026060038491022787,
|
| 2570 |
+
"loss": 0.9777,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.6948267679164689,
|
| 2575 |
+
"grad_norm": 0.3479809339329037,
|
| 2576 |
+
"learning_rate": 0.00025769507268628993,
|
| 2577 |
+
"loss": 0.9721,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.6967252017085904,
|
| 2582 |
+
"grad_norm": 0.31070434225155574,
|
| 2583 |
+
"learning_rate": 0.00025480041342212695,
|
| 2584 |
+
"loss": 0.9553,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.6986236355007119,
|
| 2589 |
+
"grad_norm": 0.3849172751179302,
|
| 2590 |
+
"learning_rate": 0.00025191653438171545,
|
| 2591 |
+
"loss": 0.9461,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.7005220692928334,
|
| 2596 |
+
"grad_norm": 0.4264474066939237,
|
| 2597 |
+
"learning_rate": 0.00024904356235507945,
|
| 2598 |
+
"loss": 0.9467,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.702420503084955,
|
| 2603 |
+
"grad_norm": 0.3548766629306175,
|
| 2604 |
+
"learning_rate": 0.0002461816236527141,
|
| 2605 |
+
"loss": 0.947,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.7043189368770764,
|
| 2610 |
+
"grad_norm": 0.3514101652953766,
|
| 2611 |
+
"learning_rate": 0.0002433308441000338,
|
| 2612 |
+
"loss": 0.9268,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.7062173706691979,
|
| 2617 |
+
"grad_norm": 0.380453955195146,
|
| 2618 |
+
"learning_rate": 0.00024049134903183955,
|
| 2619 |
+
"loss": 0.9293,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.7081158044613194,
|
| 2624 |
+
"grad_norm": 0.35298900401975053,
|
| 2625 |
+
"learning_rate": 0.00023766326328680958,
|
| 2626 |
+
"loss": 0.9348,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.7100142382534409,
|
| 2631 |
+
"grad_norm": 0.42441310564082074,
|
| 2632 |
+
"learning_rate": 0.00023484671120200935,
|
| 2633 |
+
"loss": 0.9276,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.7119126720455624,
|
| 2638 |
+
"grad_norm": 0.319038899787948,
|
| 2639 |
+
"learning_rate": 0.00023204181660742602,
|
| 2640 |
+
"loss": 0.9431,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.7138111058376839,
|
| 2645 |
+
"grad_norm": 0.4049851574728775,
|
| 2646 |
+
"learning_rate": 0.00022924870282052445,
|
| 2647 |
+
"loss": 0.906,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.7157095396298054,
|
| 2652 |
+
"grad_norm": 0.3513877016599214,
|
| 2653 |
+
"learning_rate": 0.00022646749264082478,
|
| 2654 |
+
"loss": 0.9095,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.717607973421927,
|
| 2659 |
+
"grad_norm": 0.4081310968482307,
|
| 2660 |
+
"learning_rate": 0.00022369830834450367,
|
| 2661 |
+
"loss": 0.9003,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.7195064072140485,
|
| 2666 |
+
"grad_norm": 0.35805106514821416,
|
| 2667 |
+
"learning_rate": 0.00022094127167901934,
|
| 2668 |
+
"loss": 0.9099,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.7214048410061699,
|
| 2673 |
+
"grad_norm": 0.3631465368774142,
|
| 2674 |
+
"learning_rate": 0.0002181965038577577,
|
| 2675 |
+
"loss": 0.9092,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.7233032747982914,
|
| 2680 |
+
"grad_norm": 0.3650905218048075,
|
| 2681 |
+
"learning_rate": 0.0002154641255547038,
|
| 2682 |
+
"loss": 0.9252,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.7252017085904129,
|
| 2687 |
+
"grad_norm": 0.31898731584804263,
|
| 2688 |
+
"learning_rate": 0.00021274425689913617,
|
| 2689 |
+
"loss": 0.9071,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.7271001423825344,
|
| 2694 |
+
"grad_norm": 0.31603070803205635,
|
| 2695 |
+
"learning_rate": 0.00021003701747034616,
|
| 2696 |
+
"loss": 0.9165,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.7289985761746559,
|
| 2701 |
+
"grad_norm": 0.30904969427001383,
|
| 2702 |
+
"learning_rate": 0.00020734252629237893,
|
| 2703 |
+
"loss": 0.8971,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.7308970099667774,
|
| 2708 |
+
"grad_norm": 0.32433517710502985,
|
| 2709 |
+
"learning_rate": 0.00020466090182880248,
|
| 2710 |
+
"loss": 0.8908,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.7327954437588989,
|
| 2715 |
+
"grad_norm": 0.3168746074632352,
|
| 2716 |
+
"learning_rate": 0.00020199226197749792,
|
| 2717 |
+
"loss": 0.9011,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.7346938775510204,
|
| 2722 |
+
"grad_norm": 0.31832881987489503,
|
| 2723 |
+
"learning_rate": 0.00019933672406547665,
|
| 2724 |
+
"loss": 0.9138,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.736592311343142,
|
| 2729 |
+
"grad_norm": 0.301051575919513,
|
| 2730 |
+
"learning_rate": 0.00019669440484372213,
|
| 2731 |
+
"loss": 0.8897,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.7384907451352634,
|
| 2736 |
+
"grad_norm": 0.33220133814920705,
|
| 2737 |
+
"learning_rate": 0.00019406542048205666,
|
| 2738 |
+
"loss": 0.8868,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.7403891789273849,
|
| 2743 |
+
"grad_norm": 0.31331199745305405,
|
| 2744 |
+
"learning_rate": 0.0001914498865640344,
|
| 2745 |
+
"loss": 0.8921,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.7422876127195064,
|
| 2750 |
+
"grad_norm": 0.3150890748673222,
|
| 2751 |
+
"learning_rate": 0.00018884791808185947,
|
| 2752 |
+
"loss": 0.8925,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.7441860465116279,
|
| 2757 |
+
"grad_norm": 0.31284305447491684,
|
| 2758 |
+
"learning_rate": 0.0001862596294313299,
|
| 2759 |
+
"loss": 0.8779,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.7460844803037494,
|
| 2764 |
+
"grad_norm": 0.33272829874860427,
|
| 2765 |
+
"learning_rate": 0.00018368513440680884,
|
| 2766 |
+
"loss": 0.8799,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.7479829140958709,
|
| 2771 |
+
"grad_norm": 0.32686823493529116,
|
| 2772 |
+
"learning_rate": 0.0001811245461962212,
|
| 2773 |
+
"loss": 0.8837,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.7498813478879924,
|
| 2778 |
+
"grad_norm": 0.32129617206658423,
|
| 2779 |
+
"learning_rate": 0.0001785779773760775,
|
| 2780 |
+
"loss": 0.891,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.751779781680114,
|
| 2785 |
+
"grad_norm": 0.29507710361264766,
|
| 2786 |
+
"learning_rate": 0.0001760455399065246,
|
| 2787 |
+
"loss": 0.87,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.7536782154722355,
|
| 2792 |
+
"grad_norm": 0.3392524599510225,
|
| 2793 |
+
"learning_rate": 0.00017352734512642276,
|
| 2794 |
+
"loss": 0.8614,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.7555766492643569,
|
| 2799 |
+
"grad_norm": 0.34922494756037853,
|
| 2800 |
+
"learning_rate": 0.00017102350374845155,
|
| 2801 |
+
"loss": 0.8739,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.7574750830564784,
|
| 2806 |
+
"grad_norm": 0.38856477147363533,
|
| 2807 |
+
"learning_rate": 0.00016853412585424128,
|
| 2808 |
+
"loss": 0.8608,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.7593735168485999,
|
| 2813 |
+
"grad_norm": 0.3667333097382861,
|
| 2814 |
+
"learning_rate": 0.00016605932088953397,
|
| 2815 |
+
"loss": 0.8524,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.7612719506407214,
|
| 2820 |
+
"grad_norm": 0.4112130915608781,
|
| 2821 |
+
"learning_rate": 0.00016359919765937149,
|
| 2822 |
+
"loss": 0.851,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.7631703844328429,
|
| 2827 |
+
"grad_norm": 0.35792666839319404,
|
| 2828 |
+
"learning_rate": 0.00016115386432331147,
|
| 2829 |
+
"loss": 0.8444,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.7650688182249644,
|
| 2834 |
+
"grad_norm": 0.35486084734332507,
|
| 2835 |
+
"learning_rate": 0.00015872342839067305,
|
| 2836 |
+
"loss": 0.8561,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.7669672520170859,
|
| 2841 |
+
"grad_norm": 0.34325463219420715,
|
| 2842 |
+
"learning_rate": 0.0001563079967158088,
|
| 2843 |
+
"loss": 0.8549,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.7688656858092074,
|
| 2848 |
+
"grad_norm": 0.33043895406245294,
|
| 2849 |
+
"learning_rate": 0.0001539076754934084,
|
| 2850 |
+
"loss": 0.8576,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.770764119601329,
|
| 2855 |
+
"grad_norm": 0.30878788182142447,
|
| 2856 |
+
"learning_rate": 0.00015152257025382844,
|
| 2857 |
+
"loss": 0.8559,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.7726625533934504,
|
| 2862 |
+
"grad_norm": 0.35444741515541456,
|
| 2863 |
+
"learning_rate": 0.00014915278585845348,
|
| 2864 |
+
"loss": 0.8415,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.7745609871855719,
|
| 2869 |
+
"grad_norm": 0.30847085513086014,
|
| 2870 |
+
"learning_rate": 0.00014679842649508568,
|
| 2871 |
+
"loss": 0.8259,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.7764594209776934,
|
| 2876 |
+
"grad_norm": 0.31252404771224196,
|
| 2877 |
+
"learning_rate": 0.00014445959567336441,
|
| 2878 |
+
"loss": 0.859,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.7783578547698149,
|
| 2883 |
+
"grad_norm": 0.3240169921265319,
|
| 2884 |
+
"learning_rate": 0.0001421363962202149,
|
| 2885 |
+
"loss": 0.8301,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.7802562885619364,
|
| 2890 |
+
"grad_norm": 0.32322492788162155,
|
| 2891 |
+
"learning_rate": 0.00013982893027532757,
|
| 2892 |
+
"loss": 0.822,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.7821547223540579,
|
| 2897 |
+
"grad_norm": 0.3516161817741868,
|
| 2898 |
+
"learning_rate": 0.00013753729928666825,
|
| 2899 |
+
"loss": 0.8374,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.7840531561461794,
|
| 2904 |
+
"grad_norm": 0.31124890284462753,
|
| 2905 |
+
"learning_rate": 0.00013526160400601682,
|
| 2906 |
+
"loss": 0.8293,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.7859515899383009,
|
| 2911 |
+
"grad_norm": 0.307376831052294,
|
| 2912 |
+
"learning_rate": 0.00013300194448453818,
|
| 2913 |
+
"loss": 0.8314,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.7878500237304225,
|
| 2918 |
+
"grad_norm": 0.33386681520064726,
|
| 2919 |
+
"learning_rate": 0.00013075842006838407,
|
| 2920 |
+
"loss": 0.8485,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.7897484575225439,
|
| 2925 |
+
"grad_norm": 0.3223752393505699,
|
| 2926 |
+
"learning_rate": 0.0001285311293943241,
|
| 2927 |
+
"loss": 0.8424,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.7916468913146654,
|
| 2932 |
+
"grad_norm": 0.3575122610804609,
|
| 2933 |
+
"learning_rate": 0.00012632017038541026,
|
| 2934 |
+
"loss": 0.8326,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.7935453251067869,
|
| 2939 |
+
"grad_norm": 0.32346828366821057,
|
| 2940 |
+
"learning_rate": 0.0001241256402466709,
|
| 2941 |
+
"loss": 0.8422,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.7954437588989084,
|
| 2946 |
+
"grad_norm": 0.3341501772850559,
|
| 2947 |
+
"learning_rate": 0.00012194763546083803,
|
| 2948 |
+
"loss": 0.8155,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.7973421926910299,
|
| 2953 |
+
"grad_norm": 0.33689872208599597,
|
| 2954 |
+
"learning_rate": 0.00011978625178410434,
|
| 2955 |
+
"loss": 0.8179,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.7992406264831514,
|
| 2960 |
+
"grad_norm": 0.3640068937736561,
|
| 2961 |
+
"learning_rate": 0.00011764158424191435,
|
| 2962 |
+
"loss": 0.8107,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.8011390602752729,
|
| 2967 |
+
"grad_norm": 0.31299920707259116,
|
| 2968 |
+
"learning_rate": 0.00011551372712478575,
|
| 2969 |
+
"loss": 0.8013,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.8030374940673944,
|
| 2974 |
+
"grad_norm": 0.3184880942321623,
|
| 2975 |
+
"learning_rate": 0.0001134027739841642,
|
| 2976 |
+
"loss": 0.8012,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.804935927859516,
|
| 2981 |
+
"grad_norm": 0.3162681368592125,
|
| 2982 |
+
"learning_rate": 0.00011130881762831069,
|
| 2983 |
+
"loss": 0.8156,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.8068343616516374,
|
| 2988 |
+
"grad_norm": 0.3211345805440541,
|
| 2989 |
+
"learning_rate": 0.00010923195011822058,
|
| 2990 |
+
"loss": 0.8051,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.8087327954437589,
|
| 2995 |
+
"grad_norm": 0.3249567779384397,
|
| 2996 |
+
"learning_rate": 0.00010717226276357667,
|
| 2997 |
+
"loss": 0.8024,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.8106312292358804,
|
| 3002 |
+
"grad_norm": 0.31173448153562644,
|
| 3003 |
+
"learning_rate": 0.00010512984611873466,
|
| 3004 |
+
"loss": 0.8028,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.8125296630280019,
|
| 3009 |
+
"grad_norm": 0.34264672110601546,
|
| 3010 |
+
"learning_rate": 0.00010310478997874162,
|
| 3011 |
+
"loss": 0.7893,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.8144280968201234,
|
| 3016 |
+
"grad_norm": 0.36225713373489543,
|
| 3017 |
+
"learning_rate": 0.0001010971833753882,
|
| 3018 |
+
"loss": 0.7941,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.8163265306122449,
|
| 3023 |
+
"grad_norm": 0.39829779195110515,
|
| 3024 |
+
"learning_rate": 9.910711457329479e-05,
|
| 3025 |
+
"loss": 0.784,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.8182249644043664,
|
| 3030 |
+
"grad_norm": 0.4154842554468083,
|
| 3031 |
+
"learning_rate": 9.713467106603024e-05,
|
| 3032 |
+
"loss": 0.7973,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.8201233981964879,
|
| 3037 |
+
"grad_norm": 0.35334120375262246,
|
| 3038 |
+
"learning_rate": 9.517993957226612e-05,
|
| 3039 |
+
"loss": 0.7887,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.8220218319886095,
|
| 3044 |
+
"grad_norm": 0.342997739965322,
|
| 3045 |
+
"learning_rate": 9.3243006031963e-05,
|
| 3046 |
+
"loss": 0.7884,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.8239202657807309,
|
| 3051 |
+
"grad_norm": 0.345714194018433,
|
| 3052 |
+
"learning_rate": 9.132395560259337e-05,
|
| 3053 |
+
"loss": 0.7894,
|
| 3054 |
+
"step": 2170
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.8258186995728524,
|
| 3058 |
+
"grad_norm": 0.34821270209325056,
|
| 3059 |
+
"learning_rate": 8.942287265539639e-05,
|
| 3060 |
+
"loss": 0.796,
|
| 3061 |
+
"step": 2175
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.8277171333649739,
|
| 3065 |
+
"grad_norm": 0.32502745681436357,
|
| 3066 |
+
"learning_rate": 8.753984077166937e-05,
|
| 3067 |
+
"loss": 0.8015,
|
| 3068 |
+
"step": 2180
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.8296155671570954,
|
| 3072 |
+
"grad_norm": 0.388775004922499,
|
| 3073 |
+
"learning_rate": 8.567494273909277e-05,
|
| 3074 |
+
"loss": 0.8025,
|
| 3075 |
+
"step": 2185
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.8315140009492169,
|
| 3079 |
+
"grad_norm": 0.3127036551166177,
|
| 3080 |
+
"learning_rate": 8.382826054809079e-05,
|
| 3081 |
+
"loss": 0.7895,
|
| 3082 |
+
"step": 2190
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.8334124347413384,
|
| 3086 |
+
"grad_norm": 0.29839600378854597,
|
| 3087 |
+
"learning_rate": 8.1999875388226e-05,
|
| 3088 |
+
"loss": 0.7818,
|
| 3089 |
+
"step": 2195
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.8353108685334599,
|
| 3093 |
+
"grad_norm": 0.31004929243565,
|
| 3094 |
+
"learning_rate": 8.018986764463032e-05,
|
| 3095 |
+
"loss": 0.7881,
|
| 3096 |
+
"step": 2200
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.8372093023255814,
|
| 3100 |
+
"grad_norm": 0.32248432401397453,
|
| 3101 |
+
"learning_rate": 7.839831689447102e-05,
|
| 3102 |
+
"loss": 0.7961,
|
| 3103 |
+
"step": 2205
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.8391077361177028,
|
| 3107 |
+
"grad_norm": 0.32300916569046356,
|
| 3108 |
+
"learning_rate": 7.662530190345157e-05,
|
| 3109 |
+
"loss": 0.7859,
|
| 3110 |
+
"step": 2210
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.8410061699098244,
|
| 3114 |
+
"grad_norm": 0.3437848320718501,
|
| 3115 |
+
"learning_rate": 7.487090062234898e-05,
|
| 3116 |
+
"loss": 0.7706,
|
| 3117 |
+
"step": 2215
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.8429046037019459,
|
| 3121 |
+
"grad_norm": 0.3104343754734283,
|
| 3122 |
+
"learning_rate": 7.313519018358695e-05,
|
| 3123 |
+
"loss": 0.7715,
|
| 3124 |
+
"step": 2220
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.8448030374940674,
|
| 3128 |
+
"grad_norm": 0.3171733947505178,
|
| 3129 |
+
"learning_rate": 7.141824689784421e-05,
|
| 3130 |
+
"loss": 0.7487,
|
| 3131 |
+
"step": 2225
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.8467014712861889,
|
| 3135 |
+
"grad_norm": 0.3114906950878784,
|
| 3136 |
+
"learning_rate": 6.972014625069984e-05,
|
| 3137 |
+
"loss": 0.7557,
|
| 3138 |
+
"step": 2230
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.8485999050783104,
|
| 3142 |
+
"grad_norm": 0.3427221492524242,
|
| 3143 |
+
"learning_rate": 6.804096289931443e-05,
|
| 3144 |
+
"loss": 0.7505,
|
| 3145 |
+
"step": 2235
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.8504983388704319,
|
| 3149 |
+
"grad_norm": 0.3257908676508504,
|
| 3150 |
+
"learning_rate": 6.638077066914811e-05,
|
| 3151 |
+
"loss": 0.7691,
|
| 3152 |
+
"step": 2240
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.8523967726625534,
|
| 3156 |
+
"grad_norm": 0.3245771425045967,
|
| 3157 |
+
"learning_rate": 6.473964255071418e-05,
|
| 3158 |
+
"loss": 0.7712,
|
| 3159 |
+
"step": 2245
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.8542952064546749,
|
| 3163 |
+
"grad_norm": 0.30642317659601426,
|
| 3164 |
+
"learning_rate": 6.311765069637037e-05,
|
| 3165 |
+
"loss": 0.7508,
|
| 3166 |
+
"step": 2250
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.8561936402467963,
|
| 3170 |
+
"grad_norm": 0.35717316205581434,
|
| 3171 |
+
"learning_rate": 6.151486641714705e-05,
|
| 3172 |
+
"loss": 0.7585,
|
| 3173 |
+
"step": 2255
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.8580920740389179,
|
| 3177 |
+
"grad_norm": 0.29834051216293683,
|
| 3178 |
+
"learning_rate": 5.993136017961143e-05,
|
| 3179 |
+
"loss": 0.7564,
|
| 3180 |
+
"step": 2260
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.8599905078310394,
|
| 3184 |
+
"grad_norm": 0.31967989226553994,
|
| 3185 |
+
"learning_rate": 5.83672016027697e-05,
|
| 3186 |
+
"loss": 0.7402,
|
| 3187 |
+
"step": 2265
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.8618889416231609,
|
| 3191 |
+
"grad_norm": 0.3316739821901055,
|
| 3192 |
+
"learning_rate": 5.6822459455006246e-05,
|
| 3193 |
+
"loss": 0.7617,
|
| 3194 |
+
"step": 2270
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.8637873754152824,
|
| 3198 |
+
"grad_norm": 0.33111810076533643,
|
| 3199 |
+
"learning_rate": 5.529720165106056e-05,
|
| 3200 |
+
"loss": 0.7702,
|
| 3201 |
+
"step": 2275
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.8656858092074039,
|
| 3205 |
+
"grad_norm": 0.31953471953308643,
|
| 3206 |
+
"learning_rate": 5.3791495249040644e-05,
|
| 3207 |
+
"loss": 0.7585,
|
| 3208 |
+
"step": 2280
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.8675842429995254,
|
| 3212 |
+
"grad_norm": 0.34435784933473257,
|
| 3213 |
+
"learning_rate": 5.2305406447475504e-05,
|
| 3214 |
+
"loss": 0.7539,
|
| 3215 |
+
"step": 2285
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.8694826767916469,
|
| 3219 |
+
"grad_norm": 0.33531663584742627,
|
| 3220 |
+
"learning_rate": 5.083900058240437e-05,
|
| 3221 |
+
"loss": 0.7574,
|
| 3222 |
+
"step": 2290
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.8713811105837684,
|
| 3226 |
+
"grad_norm": 0.32627115970227594,
|
| 3227 |
+
"learning_rate": 4.939234212450405e-05,
|
| 3228 |
+
"loss": 0.7502,
|
| 3229 |
+
"step": 2295
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.8732795443758898,
|
| 3233 |
+
"grad_norm": 0.34710941754901015,
|
| 3234 |
+
"learning_rate": 4.796549467625494e-05,
|
| 3235 |
+
"loss": 0.7517,
|
| 3236 |
+
"step": 2300
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.8751779781680114,
|
| 3240 |
+
"grad_norm": 0.3259737091035039,
|
| 3241 |
+
"learning_rate": 4.6558520969144205e-05,
|
| 3242 |
+
"loss": 0.7531,
|
| 3243 |
+
"step": 2305
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.8770764119601329,
|
| 3247 |
+
"grad_norm": 0.3543527295348622,
|
| 3248 |
+
"learning_rate": 4.517148286090822e-05,
|
| 3249 |
+
"loss": 0.7395,
|
| 3250 |
+
"step": 2310
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.8789748457522544,
|
| 3254 |
+
"grad_norm": 0.3411443809878121,
|
| 3255 |
+
"learning_rate": 4.3804441332812915e-05,
|
| 3256 |
+
"loss": 0.7493,
|
| 3257 |
+
"step": 2315
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.8808732795443759,
|
| 3261 |
+
"grad_norm": 0.323002963765091,
|
| 3262 |
+
"learning_rate": 4.245745648697241e-05,
|
| 3263 |
+
"loss": 0.7562,
|
| 3264 |
+
"step": 2320
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.8827717133364974,
|
| 3268 |
+
"grad_norm": 0.3369758289265976,
|
| 3269 |
+
"learning_rate": 4.1130587543706796e-05,
|
| 3270 |
+
"loss": 0.7466,
|
| 3271 |
+
"step": 2325
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.8846701471286189,
|
| 3275 |
+
"grad_norm": 0.3152326006804494,
|
| 3276 |
+
"learning_rate": 3.982389283893878e-05,
|
| 3277 |
+
"loss": 0.7458,
|
| 3278 |
+
"step": 2330
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.8865685809207404,
|
| 3282 |
+
"grad_norm": 0.3504098491203572,
|
| 3283 |
+
"learning_rate": 3.853742982162839e-05,
|
| 3284 |
+
"loss": 0.7503,
|
| 3285 |
+
"step": 2335
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.8884670147128619,
|
| 3289 |
+
"grad_norm": 0.33687850279262355,
|
| 3290 |
+
"learning_rate": 3.72712550512479e-05,
|
| 3291 |
+
"loss": 0.7651,
|
| 3292 |
+
"step": 2340
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.8903654485049833,
|
| 3296 |
+
"grad_norm": 0.3386005467292844,
|
| 3297 |
+
"learning_rate": 3.602542419529453e-05,
|
| 3298 |
+
"loss": 0.766,
|
| 3299 |
+
"step": 2345
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.8922638822971048,
|
| 3303 |
+
"grad_norm": 0.3110137804067388,
|
| 3304 |
+
"learning_rate": 3.479999202684353e-05,
|
| 3305 |
+
"loss": 0.7397,
|
| 3306 |
+
"step": 2350
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.8941623160892264,
|
| 3310 |
+
"grad_norm": 0.32583342601441784,
|
| 3311 |
+
"learning_rate": 3.359501242213981e-05,
|
| 3312 |
+
"loss": 0.7346,
|
| 3313 |
+
"step": 2355
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.8960607498813479,
|
| 3317 |
+
"grad_norm": 0.3187549346009566,
|
| 3318 |
+
"learning_rate": 3.24105383582291e-05,
|
| 3319 |
+
"loss": 0.7492,
|
| 3320 |
+
"step": 2360
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.8979591836734694,
|
| 3324 |
+
"grad_norm": 0.3218971294090315,
|
| 3325 |
+
"learning_rate": 3.1246621910629323e-05,
|
| 3326 |
+
"loss": 0.7504,
|
| 3327 |
+
"step": 2365
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.8998576174655909,
|
| 3331 |
+
"grad_norm": 0.3509832182230347,
|
| 3332 |
+
"learning_rate": 3.0103314251040683e-05,
|
| 3333 |
+
"loss": 0.7474,
|
| 3334 |
+
"step": 2370
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.9017560512577124,
|
| 3338 |
+
"grad_norm": 0.3045155962925135,
|
| 3339 |
+
"learning_rate": 2.8980665645095993e-05,
|
| 3340 |
+
"loss": 0.7457,
|
| 3341 |
+
"step": 2375
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.9036544850498339,
|
| 3345 |
+
"grad_norm": 0.33768892145147017,
|
| 3346 |
+
"learning_rate": 2.787872545015069e-05,
|
| 3347 |
+
"loss": 0.7369,
|
| 3348 |
+
"step": 2380
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.9055529188419554,
|
| 3352 |
+
"grad_norm": 0.30591307319847055,
|
| 3353 |
+
"learning_rate": 2.679754211311314e-05,
|
| 3354 |
+
"loss": 0.7349,
|
| 3355 |
+
"step": 2385
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.9074513526340768,
|
| 3359 |
+
"grad_norm": 0.31814695669489196,
|
| 3360 |
+
"learning_rate": 2.5737163168314093e-05,
|
| 3361 |
+
"loss": 0.7465,
|
| 3362 |
+
"step": 2390
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.9093497864261983,
|
| 3366 |
+
"grad_norm": 0.33557857522275575,
|
| 3367 |
+
"learning_rate": 2.4697635235417403e-05,
|
| 3368 |
+
"loss": 0.7304,
|
| 3369 |
+
"step": 2395
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.9112482202183199,
|
| 3373 |
+
"grad_norm": 0.31077407798760204,
|
| 3374 |
+
"learning_rate": 2.3679004017370165e-05,
|
| 3375 |
+
"loss": 0.7322,
|
| 3376 |
+
"step": 2400
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.9131466540104414,
|
| 3380 |
+
"grad_norm": 0.3696076542535449,
|
| 3381 |
+
"learning_rate": 2.2681314298393208e-05,
|
| 3382 |
+
"loss": 0.7415,
|
| 3383 |
+
"step": 2405
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.9150450878025629,
|
| 3387 |
+
"grad_norm": 0.34756329848102396,
|
| 3388 |
+
"learning_rate": 2.1704609942012344e-05,
|
| 3389 |
+
"loss": 0.7437,
|
| 3390 |
+
"step": 2410
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.9169435215946844,
|
| 3394 |
+
"grad_norm": 0.3105145840944022,
|
| 3395 |
+
"learning_rate": 2.074893388912996e-05,
|
| 3396 |
+
"loss": 0.7261,
|
| 3397 |
+
"step": 2415
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.9188419553868059,
|
| 3401 |
+
"grad_norm": 0.3385109368325919,
|
| 3402 |
+
"learning_rate": 1.9814328156136986e-05,
|
| 3403 |
+
"loss": 0.7507,
|
| 3404 |
+
"step": 2420
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.9207403891789274,
|
| 3408 |
+
"grad_norm": 0.30326901789443217,
|
| 3409 |
+
"learning_rate": 1.8900833833065622e-05,
|
| 3410 |
+
"loss": 0.7368,
|
| 3411 |
+
"step": 2425
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.9226388229710489,
|
| 3415 |
+
"grad_norm": 0.31849955729705076,
|
| 3416 |
+
"learning_rate": 1.800849108178304e-05,
|
| 3417 |
+
"loss": 0.732,
|
| 3418 |
+
"step": 2430
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.9245372567631703,
|
| 3422 |
+
"grad_norm": 0.3023724936604926,
|
| 3423 |
+
"learning_rate": 1.7137339134225326e-05,
|
| 3424 |
+
"loss": 0.7395,
|
| 3425 |
+
"step": 2435
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.9264356905552918,
|
| 3429 |
+
"grad_norm": 0.3045797697728719,
|
| 3430 |
+
"learning_rate": 1.628741629067282e-05,
|
| 3431 |
+
"loss": 0.7372,
|
| 3432 |
+
"step": 2440
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.9283341243474134,
|
| 3436 |
+
"grad_norm": 0.3072440159247055,
|
| 3437 |
+
"learning_rate": 1.5458759918066333e-05,
|
| 3438 |
+
"loss": 0.7271,
|
| 3439 |
+
"step": 2445
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.9302325581395349,
|
| 3443 |
+
"grad_norm": 0.2973843731076206,
|
| 3444 |
+
"learning_rate": 1.4651406448364046e-05,
|
| 3445 |
+
"loss": 0.7286,
|
| 3446 |
+
"step": 2450
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.9321309919316564,
|
| 3450 |
+
"grad_norm": 0.33315733476007137,
|
| 3451 |
+
"learning_rate": 1.3865391376940151e-05,
|
| 3452 |
+
"loss": 0.7245,
|
| 3453 |
+
"step": 2455
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.9340294257237779,
|
| 3457 |
+
"grad_norm": 0.35918744134846453,
|
| 3458 |
+
"learning_rate": 1.3100749261024003e-05,
|
| 3459 |
+
"loss": 0.7233,
|
| 3460 |
+
"step": 2460
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.9359278595158994,
|
| 3464 |
+
"grad_norm": 0.2959597147751244,
|
| 3465 |
+
"learning_rate": 1.2357513718180724e-05,
|
| 3466 |
+
"loss": 0.7406,
|
| 3467 |
+
"step": 2465
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 0.9378262933080209,
|
| 3471 |
+
"grad_norm": 0.6025608542347619,
|
| 3472 |
+
"learning_rate": 1.1635717424833602e-05,
|
| 3473 |
+
"loss": 0.7319,
|
| 3474 |
+
"step": 2470
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 0.9397247271001424,
|
| 3478 |
+
"grad_norm": 0.31728158924540384,
|
| 3479 |
+
"learning_rate": 1.0935392114827026e-05,
|
| 3480 |
+
"loss": 0.7142,
|
| 3481 |
+
"step": 2475
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 0.9416231608922638,
|
| 3485 |
+
"grad_norm": 0.3202471419277355,
|
| 3486 |
+
"learning_rate": 1.0256568578031533e-05,
|
| 3487 |
+
"loss": 0.7234,
|
| 3488 |
+
"step": 2480
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.9435215946843853,
|
| 3492 |
+
"grad_norm": 0.33730636346394754,
|
| 3493 |
+
"learning_rate": 9.599276658990353e-06,
|
| 3494 |
+
"loss": 0.7266,
|
| 3495 |
+
"step": 2485
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 0.9454200284765069,
|
| 3499 |
+
"grad_norm": 0.3105080016851193,
|
| 3500 |
+
"learning_rate": 8.963545255606664e-06,
|
| 3501 |
+
"loss": 0.7309,
|
| 3502 |
+
"step": 2490
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 0.9473184622686284,
|
| 3506 |
+
"grad_norm": 0.31936758390953845,
|
| 3507 |
+
"learning_rate": 8.349402317873788e-06,
|
| 3508 |
+
"loss": 0.7385,
|
| 3509 |
+
"step": 2495
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 0.9492168960607499,
|
| 3513 |
+
"grad_norm": 0.35011965244445487,
|
| 3514 |
+
"learning_rate": 7.756874846645834e-06,
|
| 3515 |
+
"loss": 0.7298,
|
| 3516 |
+
"step": 2500
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 0.9511153298528714,
|
| 3520 |
+
"grad_norm": 0.3189446744264921,
|
| 3521 |
+
"learning_rate": 7.185988892450923e-06,
|
| 3522 |
+
"loss": 0.7345,
|
| 3523 |
+
"step": 2505
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 0.9530137636449929,
|
| 3527 |
+
"grad_norm": 0.3463481807220011,
|
| 3528 |
+
"learning_rate": 6.636769554345778e-06,
|
| 3529 |
+
"loss": 0.7296,
|
| 3530 |
+
"step": 2510
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.9549121974371144,
|
| 3534 |
+
"grad_norm": 0.3144784342698344,
|
| 3535 |
+
"learning_rate": 6.109240978812047e-06,
|
| 3536 |
+
"loss": 0.733,
|
| 3537 |
+
"step": 2515
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 0.9568106312292359,
|
| 3541 |
+
"grad_norm": 0.32358238607927053,
|
| 3542 |
+
"learning_rate": 5.603426358695207e-06,
|
| 3543 |
+
"loss": 0.7227,
|
| 3544 |
+
"step": 2520
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 0.9587090650213573,
|
| 3548 |
+
"grad_norm": 0.3070502390360057,
|
| 3549 |
+
"learning_rate": 5.11934793218427e-06,
|
| 3550 |
+
"loss": 0.7379,
|
| 3551 |
+
"step": 2525
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 0.9606074988134788,
|
| 3555 |
+
"grad_norm": 0.30930683987481156,
|
| 3556 |
+
"learning_rate": 4.657026981834622e-06,
|
| 3557 |
+
"loss": 0.7244,
|
| 3558 |
+
"step": 2530
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 0.9625059326056004,
|
| 3562 |
+
"grad_norm": 0.3184603698723896,
|
| 3563 |
+
"learning_rate": 4.216483833631879e-06,
|
| 3564 |
+
"loss": 0.7374,
|
| 3565 |
+
"step": 2535
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 0.9644043663977219,
|
| 3569 |
+
"grad_norm": 0.313081890733716,
|
| 3570 |
+
"learning_rate": 3.7977378560985487e-06,
|
| 3571 |
+
"loss": 0.7253,
|
| 3572 |
+
"step": 2540
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 0.9663028001898434,
|
| 3576 |
+
"grad_norm": 0.33414880766899385,
|
| 3577 |
+
"learning_rate": 3.4008074594423233e-06,
|
| 3578 |
+
"loss": 0.7246,
|
| 3579 |
+
"step": 2545
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 0.9682012339819649,
|
| 3583 |
+
"grad_norm": 0.30264216877470684,
|
| 3584 |
+
"learning_rate": 3.0257100947470027e-06,
|
| 3585 |
+
"loss": 0.7274,
|
| 3586 |
+
"step": 2550
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 0.9700996677740864,
|
| 3590 |
+
"grad_norm": 0.34059297981030995,
|
| 3591 |
+
"learning_rate": 2.672462253204666e-06,
|
| 3592 |
+
"loss": 0.7429,
|
| 3593 |
+
"step": 2555
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 0.9719981015662079,
|
| 3597 |
+
"grad_norm": 0.3255229667279891,
|
| 3598 |
+
"learning_rate": 2.3410794653911936e-06,
|
| 3599 |
+
"loss": 0.7246,
|
| 3600 |
+
"step": 2560
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 0.9738965353583294,
|
| 3604 |
+
"grad_norm": 0.32454959915539006,
|
| 3605 |
+
"learning_rate": 2.03157630058326e-06,
|
| 3606 |
+
"loss": 0.7202,
|
| 3607 |
+
"step": 2565
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 0.9757949691504508,
|
| 3611 |
+
"grad_norm": 0.3205716386887886,
|
| 3612 |
+
"learning_rate": 1.7439663661176219e-06,
|
| 3613 |
+
"loss": 0.7363,
|
| 3614 |
+
"step": 2570
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 0.9776934029425723,
|
| 3618 |
+
"grad_norm": 0.29763916460228407,
|
| 3619 |
+
"learning_rate": 1.478262306793099e-06,
|
| 3620 |
+
"loss": 0.7319,
|
| 3621 |
+
"step": 2575
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 0.9795918367346939,
|
| 3625 |
+
"grad_norm": 0.3316371045620813,
|
| 3626 |
+
"learning_rate": 1.234475804314683e-06,
|
| 3627 |
+
"loss": 0.7519,
|
| 3628 |
+
"step": 2580
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 0.9814902705268154,
|
| 3632 |
+
"grad_norm": 0.3105059150408721,
|
| 3633 |
+
"learning_rate": 1.012617576779673e-06,
|
| 3634 |
+
"loss": 0.7261,
|
| 3635 |
+
"step": 2585
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 0.9833887043189369,
|
| 3639 |
+
"grad_norm": 0.30466660031856074,
|
| 3640 |
+
"learning_rate": 8.126973782067171e-07,
|
| 3641 |
+
"loss": 0.7321,
|
| 3642 |
+
"step": 2590
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 0.9852871381110584,
|
| 3646 |
+
"grad_norm": 0.3373827673378842,
|
| 3647 |
+
"learning_rate": 6.347239981068231e-07,
|
| 3648 |
+
"loss": 0.7395,
|
| 3649 |
+
"step": 2595
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.9871855719031799,
|
| 3653 |
+
"grad_norm": 0.31614437085864816,
|
| 3654 |
+
"learning_rate": 4.787052610970566e-07,
|
| 3655 |
+
"loss": 0.7253,
|
| 3656 |
+
"step": 2600
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 0.9890840056953014,
|
| 3660 |
+
"grad_norm": 0.3300895421062348,
|
| 3661 |
+
"learning_rate": 3.446480265563712e-07,
|
| 3662 |
+
"loss": 0.7329,
|
| 3663 |
+
"step": 2605
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 0.9909824394874229,
|
| 3667 |
+
"grad_norm": 0.33291802253713565,
|
| 3668 |
+
"learning_rate": 2.3255818832423892e-07,
|
| 3669 |
+
"loss": 0.7245,
|
| 3670 |
+
"step": 2610
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 0.9928808732795443,
|
| 3674 |
+
"grad_norm": 0.32675535698611996,
|
| 3675 |
+
"learning_rate": 1.4244067444124652e-07,
|
| 3676 |
+
"loss": 0.7218,
|
| 3677 |
+
"step": 2615
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 0.9947793070716658,
|
| 3681 |
+
"grad_norm": 0.32024826764191183,
|
| 3682 |
+
"learning_rate": 7.429944693276847e-08,
|
| 3683 |
+
"loss": 0.736,
|
| 3684 |
+
"step": 2620
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 0.9966777408637874,
|
| 3688 |
+
"grad_norm": 0.3176106426665133,
|
| 3689 |
+
"learning_rate": 2.8137501634439844e-08,
|
| 3690 |
+
"loss": 0.7406,
|
| 3691 |
+
"step": 2625
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.9985761746559089,
|
| 3695 |
+
"grad_norm": 0.3213640893094606,
|
| 3696 |
+
"learning_rate": 3.956868060761565e-09,
|
| 3697 |
+
"loss": 0.7448,
|
| 3698 |
+
"step": 2630
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 0.9997152349311818,
|
| 3702 |
+
"eval_loss": 2.2808492183685303,
|
| 3703 |
+
"eval_runtime": 8.5401,
|
| 3704 |
+
"eval_samples_per_second": 46.018,
|
| 3705 |
+
"eval_steps_per_second": 11.592,
|
| 3706 |
+
"step": 2633
|
| 3707 |
+
},
|
| 3708 |
+
{
|
| 3709 |
+
"epoch": 0.9997152349311818,
|
| 3710 |
+
"step": 2633,
|
| 3711 |
+
"total_flos": 32299470028800.0,
|
| 3712 |
+
"train_loss": 1.3562805264158209,
|
| 3713 |
+
"train_runtime": 3817.2623,
|
| 3714 |
+
"train_samples_per_second": 11.039,
|
| 3715 |
+
"train_steps_per_second": 0.69
|
| 3716 |
+
}
|
| 3717 |
+
],
|
| 3718 |
+
"logging_steps": 5,
|
| 3719 |
+
"max_steps": 2633,
|
| 3720 |
+
"num_input_tokens_seen": 0,
|
| 3721 |
+
"num_train_epochs": 1,
|
| 3722 |
+
"save_steps": 500,
|
| 3723 |
+
"stateful_callbacks": {
|
| 3724 |
+
"TrainerControl": {
|
| 3725 |
+
"args": {
|
| 3726 |
+
"should_epoch_stop": false,
|
| 3727 |
+
"should_evaluate": false,
|
| 3728 |
+
"should_log": false,
|
| 3729 |
+
"should_save": false,
|
| 3730 |
+
"should_training_stop": false
|
| 3731 |
+
},
|
| 3732 |
+
"attributes": {}
|
| 3733 |
+
}
|
| 3734 |
+
},
|
| 3735 |
+
"total_flos": 32299470028800.0,
|
| 3736 |
+
"train_batch_size": 4,
|
| 3737 |
+
"trial_name": null,
|
| 3738 |
+
"trial_params": null
|
| 3739 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:20dff1e7a19286b7de77e2b35ef736ac905578db3f64b18524ad993ee1d18735
|
| 3 |
+
size 6776
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|