Upload 9 files
Browse files- all_results.json +9 -0
- merges.txt +0 -0
- special_tokens_map.json +15 -0
- tokenizer.json +0 -0
- tokenizer_config.json +57 -0
- train_results.json +9 -0
- trainer_state.json +2998 -0
- training_args.bin +3 -0
- vocab.json +0 -0
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9998235190305312,
|
| 3 |
+
"total_flos": 1.037665224400896e+16,
|
| 4 |
+
"train_loss": 0.1519955188758403,
|
| 5 |
+
"train_runtime": 5739.131,
|
| 6 |
+
"train_samples": 67996,
|
| 7 |
+
"train_samples_per_second": 11.848,
|
| 8 |
+
"train_steps_per_second": 0.74
|
| 9 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": {
|
| 6 |
+
"content": "<mask>",
|
| 7 |
+
"lstrip": true,
|
| 8 |
+
"normalized": true,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false
|
| 11 |
+
},
|
| 12 |
+
"pad_token": "<pad>",
|
| 13 |
+
"sep_token": "</s>",
|
| 14 |
+
"unk_token": "<unk>"
|
| 15 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<s>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": true,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<pad>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "</s>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": true,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<unk>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": true,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"50264": {
|
| 37 |
+
"content": "<mask>",
|
| 38 |
+
"lstrip": true,
|
| 39 |
+
"normalized": true,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"bos_token": "<s>",
|
| 46 |
+
"clean_up_tokenization_spaces": true,
|
| 47 |
+
"cls_token": "<s>",
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"errors": "replace",
|
| 50 |
+
"mask_token": "<mask>",
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"pad_token": "<pad>",
|
| 53 |
+
"sep_token": "</s>",
|
| 54 |
+
"tokenizer_class": "BartTokenizer",
|
| 55 |
+
"trim_offsets": true,
|
| 56 |
+
"unk_token": "<unk>"
|
| 57 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9998235190305312,
|
| 3 |
+
"total_flos": 1.037665224400896e+16,
|
| 4 |
+
"train_loss": 0.1519955188758403,
|
| 5 |
+
"train_runtime": 5739.131,
|
| 6 |
+
"train_samples": 67996,
|
| 7 |
+
"train_samples_per_second": 11.848,
|
| 8 |
+
"train_steps_per_second": 0.74
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2998 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9998235190305312,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 4249,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0023530795929172306,
|
| 13 |
+
"grad_norm": 4.500667572021484,
|
| 14 |
+
"learning_rate": 0.0002992939515180042,
|
| 15 |
+
"loss": 2.9988,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.004706159185834461,
|
| 20 |
+
"grad_norm": 2.7827706336975098,
|
| 21 |
+
"learning_rate": 0.00029858790303600844,
|
| 22 |
+
"loss": 1.0643,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.007059238778751691,
|
| 27 |
+
"grad_norm": 2.746577739715576,
|
| 28 |
+
"learning_rate": 0.0002978818545540127,
|
| 29 |
+
"loss": 0.7978,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.009412318371668922,
|
| 34 |
+
"grad_norm": 2.567692279815674,
|
| 35 |
+
"learning_rate": 0.0002971758060720169,
|
| 36 |
+
"loss": 0.6546,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.011765397964586153,
|
| 41 |
+
"grad_norm": 2.2394003868103027,
|
| 42 |
+
"learning_rate": 0.00029646975759002115,
|
| 43 |
+
"loss": 0.5361,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.014118477557503383,
|
| 48 |
+
"grad_norm": 2.1666100025177,
|
| 49 |
+
"learning_rate": 0.0002957637091080254,
|
| 50 |
+
"loss": 0.5137,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.01647155715042061,
|
| 55 |
+
"grad_norm": 1.922058343887329,
|
| 56 |
+
"learning_rate": 0.0002950576606260296,
|
| 57 |
+
"loss": 0.446,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.018824636743337845,
|
| 62 |
+
"grad_norm": 1.732611060142517,
|
| 63 |
+
"learning_rate": 0.00029435161214403386,
|
| 64 |
+
"loss": 0.4405,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.021177716336255075,
|
| 69 |
+
"grad_norm": 2.2046239376068115,
|
| 70 |
+
"learning_rate": 0.0002936455636620381,
|
| 71 |
+
"loss": 0.3959,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.023530795929172305,
|
| 76 |
+
"grad_norm": 2.083113670349121,
|
| 77 |
+
"learning_rate": 0.0002929395151800423,
|
| 78 |
+
"loss": 0.4114,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.025883875522089535,
|
| 83 |
+
"grad_norm": 1.671247124671936,
|
| 84 |
+
"learning_rate": 0.00029223346669804656,
|
| 85 |
+
"loss": 0.3554,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.028236955115006766,
|
| 90 |
+
"grad_norm": 2.001924514770508,
|
| 91 |
+
"learning_rate": 0.0002915274182160508,
|
| 92 |
+
"loss": 0.3577,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.030590034707923996,
|
| 97 |
+
"grad_norm": 2.07259202003479,
|
| 98 |
+
"learning_rate": 0.00029082136973405503,
|
| 99 |
+
"loss": 0.3422,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.03294311430084122,
|
| 104 |
+
"grad_norm": 1.7239247560501099,
|
| 105 |
+
"learning_rate": 0.00029011532125205927,
|
| 106 |
+
"loss": 0.3079,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.03529619389375846,
|
| 111 |
+
"grad_norm": 1.7430157661437988,
|
| 112 |
+
"learning_rate": 0.0002894092727700635,
|
| 113 |
+
"loss": 0.3304,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03764927348667569,
|
| 118 |
+
"grad_norm": 1.1152617931365967,
|
| 119 |
+
"learning_rate": 0.00028870322428806774,
|
| 120 |
+
"loss": 0.3009,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.04000235307959292,
|
| 125 |
+
"grad_norm": 1.5272759199142456,
|
| 126 |
+
"learning_rate": 0.00028799717580607197,
|
| 127 |
+
"loss": 0.3027,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04235543267251015,
|
| 132 |
+
"grad_norm": 1.3934285640716553,
|
| 133 |
+
"learning_rate": 0.0002872911273240762,
|
| 134 |
+
"loss": 0.2514,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04470851226542738,
|
| 139 |
+
"grad_norm": 1.7138372659683228,
|
| 140 |
+
"learning_rate": 0.00028658507884208044,
|
| 141 |
+
"loss": 0.2556,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04706159185834461,
|
| 146 |
+
"grad_norm": 1.7979109287261963,
|
| 147 |
+
"learning_rate": 0.00028587903036008473,
|
| 148 |
+
"loss": 0.2696,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.04941467145126184,
|
| 153 |
+
"grad_norm": 1.342785358428955,
|
| 154 |
+
"learning_rate": 0.0002851729818780889,
|
| 155 |
+
"loss": 0.2496,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.05176775104417907,
|
| 160 |
+
"grad_norm": 1.5516395568847656,
|
| 161 |
+
"learning_rate": 0.0002844669333960932,
|
| 162 |
+
"loss": 0.2727,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.0541208306370963,
|
| 167 |
+
"grad_norm": 6.922358989715576,
|
| 168 |
+
"learning_rate": 0.0002837608849140974,
|
| 169 |
+
"loss": 0.2492,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.05647391023001353,
|
| 174 |
+
"grad_norm": 1.5551228523254395,
|
| 175 |
+
"learning_rate": 0.00028305483643210167,
|
| 176 |
+
"loss": 0.2451,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05882698982293076,
|
| 181 |
+
"grad_norm": 1.300445318222046,
|
| 182 |
+
"learning_rate": 0.00028234878795010585,
|
| 183 |
+
"loss": 0.253,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.06118006941584799,
|
| 188 |
+
"grad_norm": 1.4355467557907104,
|
| 189 |
+
"learning_rate": 0.00028164273946811014,
|
| 190 |
+
"loss": 0.2453,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06353314900876522,
|
| 195 |
+
"grad_norm": 15.9704008102417,
|
| 196 |
+
"learning_rate": 0.0002809366909861143,
|
| 197 |
+
"loss": 0.2072,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06588622860168244,
|
| 202 |
+
"grad_norm": 1.7124171257019043,
|
| 203 |
+
"learning_rate": 0.0002802306425041186,
|
| 204 |
+
"loss": 0.4367,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06823930819459968,
|
| 209 |
+
"grad_norm": 1.6787582635879517,
|
| 210 |
+
"learning_rate": 0.0002795245940221228,
|
| 211 |
+
"loss": 0.2716,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.07059238778751692,
|
| 216 |
+
"grad_norm": 1.2618638277053833,
|
| 217 |
+
"learning_rate": 0.0002788185455401271,
|
| 218 |
+
"loss": 0.2821,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.07294546738043414,
|
| 223 |
+
"grad_norm": 2.927347421646118,
|
| 224 |
+
"learning_rate": 0.00027811249705813126,
|
| 225 |
+
"loss": 0.2712,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07529854697335138,
|
| 230 |
+
"grad_norm": 1.9304898977279663,
|
| 231 |
+
"learning_rate": 0.00027740644857613555,
|
| 232 |
+
"loss": 0.2637,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.0776516265662686,
|
| 237 |
+
"grad_norm": 1.2599807977676392,
|
| 238 |
+
"learning_rate": 0.0002767004000941398,
|
| 239 |
+
"loss": 0.2389,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.08000470615918584,
|
| 244 |
+
"grad_norm": 1.4264953136444092,
|
| 245 |
+
"learning_rate": 0.000275994351612144,
|
| 246 |
+
"loss": 0.2143,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.08235778575210306,
|
| 251 |
+
"grad_norm": 1.4229093790054321,
|
| 252 |
+
"learning_rate": 0.00027528830313014826,
|
| 253 |
+
"loss": 0.2571,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.0847108653450203,
|
| 258 |
+
"grad_norm": 1.743034839630127,
|
| 259 |
+
"learning_rate": 0.0002745822546481525,
|
| 260 |
+
"loss": 0.2292,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08706394493793752,
|
| 265 |
+
"grad_norm": 1.3582898378372192,
|
| 266 |
+
"learning_rate": 0.00027387620616615673,
|
| 267 |
+
"loss": 0.2314,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.08941702453085476,
|
| 272 |
+
"grad_norm": 1.2714539766311646,
|
| 273 |
+
"learning_rate": 0.00027317015768416096,
|
| 274 |
+
"loss": 0.2694,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.09177010412377198,
|
| 279 |
+
"grad_norm": 1.0213568210601807,
|
| 280 |
+
"learning_rate": 0.0002724641092021652,
|
| 281 |
+
"loss": 0.2269,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.09412318371668922,
|
| 286 |
+
"grad_norm": 0.8783596754074097,
|
| 287 |
+
"learning_rate": 0.00027175806072016943,
|
| 288 |
+
"loss": 0.2488,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09647626330960644,
|
| 293 |
+
"grad_norm": 1.842328667640686,
|
| 294 |
+
"learning_rate": 0.00027105201223817367,
|
| 295 |
+
"loss": 0.2175,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09882934290252368,
|
| 300 |
+
"grad_norm": 1.4185247421264648,
|
| 301 |
+
"learning_rate": 0.0002703459637561779,
|
| 302 |
+
"loss": 0.2049,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.1011824224954409,
|
| 307 |
+
"grad_norm": 1.3057924509048462,
|
| 308 |
+
"learning_rate": 0.00026963991527418214,
|
| 309 |
+
"loss": 0.1819,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.10353550208835814,
|
| 314 |
+
"grad_norm": 1.563916802406311,
|
| 315 |
+
"learning_rate": 0.0002689338667921864,
|
| 316 |
+
"loss": 0.2042,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.10588858168127536,
|
| 321 |
+
"grad_norm": 0.9588648080825806,
|
| 322 |
+
"learning_rate": 0.0002682278183101906,
|
| 323 |
+
"loss": 0.1977,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.1082416612741926,
|
| 328 |
+
"grad_norm": 1.3258203268051147,
|
| 329 |
+
"learning_rate": 0.00026752176982819485,
|
| 330 |
+
"loss": 0.1984,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.11059474086710983,
|
| 335 |
+
"grad_norm": 1.6783477067947388,
|
| 336 |
+
"learning_rate": 0.0002668157213461991,
|
| 337 |
+
"loss": 0.2086,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11294782046002706,
|
| 342 |
+
"grad_norm": 1.820469617843628,
|
| 343 |
+
"learning_rate": 0.0002661096728642033,
|
| 344 |
+
"loss": 0.2128,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.11530090005294429,
|
| 349 |
+
"grad_norm": 1.1493395566940308,
|
| 350 |
+
"learning_rate": 0.00026540362438220755,
|
| 351 |
+
"loss": 0.2022,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.11765397964586152,
|
| 356 |
+
"grad_norm": 1.3134245872497559,
|
| 357 |
+
"learning_rate": 0.0002646975759002118,
|
| 358 |
+
"loss": 0.2199,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.12000705923877875,
|
| 363 |
+
"grad_norm": 4.0345988273620605,
|
| 364 |
+
"learning_rate": 0.000263991527418216,
|
| 365 |
+
"loss": 0.2181,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.12236013883169598,
|
| 370 |
+
"grad_norm": 1.7995352745056152,
|
| 371 |
+
"learning_rate": 0.00026328547893622026,
|
| 372 |
+
"loss": 0.2233,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.1247132184246132,
|
| 377 |
+
"grad_norm": 1.6303889751434326,
|
| 378 |
+
"learning_rate": 0.0002625794304542245,
|
| 379 |
+
"loss": 0.2161,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.12706629801753044,
|
| 384 |
+
"grad_norm": 1.1785417795181274,
|
| 385 |
+
"learning_rate": 0.0002618733819722287,
|
| 386 |
+
"loss": 0.1952,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.12941937761044767,
|
| 391 |
+
"grad_norm": 2.5084645748138428,
|
| 392 |
+
"learning_rate": 0.00026116733349023296,
|
| 393 |
+
"loss": 0.204,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.1317724572033649,
|
| 398 |
+
"grad_norm": 0.9784884452819824,
|
| 399 |
+
"learning_rate": 0.0002604612850082372,
|
| 400 |
+
"loss": 0.1598,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.13412553679628214,
|
| 405 |
+
"grad_norm": 0.8020937442779541,
|
| 406 |
+
"learning_rate": 0.00025975523652624143,
|
| 407 |
+
"loss": 0.2045,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.13647861638919936,
|
| 412 |
+
"grad_norm": 0.9151997566223145,
|
| 413 |
+
"learning_rate": 0.00025904918804424567,
|
| 414 |
+
"loss": 0.1744,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.1388316959821166,
|
| 419 |
+
"grad_norm": 1.55955171585083,
|
| 420 |
+
"learning_rate": 0.0002583431395622499,
|
| 421 |
+
"loss": 0.1878,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.14118477557503384,
|
| 426 |
+
"grad_norm": 23.52423858642578,
|
| 427 |
+
"learning_rate": 0.0002576370910802542,
|
| 428 |
+
"loss": 0.1779,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.14353785516795106,
|
| 433 |
+
"grad_norm": 1.1516189575195312,
|
| 434 |
+
"learning_rate": 0.0002569310425982584,
|
| 435 |
+
"loss": 0.195,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.14589093476086828,
|
| 440 |
+
"grad_norm": 1.0912541151046753,
|
| 441 |
+
"learning_rate": 0.00025622499411626266,
|
| 442 |
+
"loss": 0.205,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.1482440143537855,
|
| 447 |
+
"grad_norm": 1.2680310010910034,
|
| 448 |
+
"learning_rate": 0.00025551894563426684,
|
| 449 |
+
"loss": 0.1654,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.15059709394670276,
|
| 454 |
+
"grad_norm": 1.2099932432174683,
|
| 455 |
+
"learning_rate": 0.00025481289715227113,
|
| 456 |
+
"loss": 0.1698,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.15295017353961998,
|
| 461 |
+
"grad_norm": 1.1155511140823364,
|
| 462 |
+
"learning_rate": 0.0002541068486702753,
|
| 463 |
+
"loss": 0.1883,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.1553032531325372,
|
| 468 |
+
"grad_norm": 1.2237110137939453,
|
| 469 |
+
"learning_rate": 0.0002534008001882796,
|
| 470 |
+
"loss": 0.1739,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.15765633272545443,
|
| 475 |
+
"grad_norm": 2.2334392070770264,
|
| 476 |
+
"learning_rate": 0.0002526947517062838,
|
| 477 |
+
"loss": 0.2149,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.16000941231837168,
|
| 482 |
+
"grad_norm": 1.0051536560058594,
|
| 483 |
+
"learning_rate": 0.0002519887032242881,
|
| 484 |
+
"loss": 0.1755,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.1623624919112889,
|
| 489 |
+
"grad_norm": 1.5381518602371216,
|
| 490 |
+
"learning_rate": 0.00025128265474229225,
|
| 491 |
+
"loss": 0.1814,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.16471557150420613,
|
| 496 |
+
"grad_norm": 1.3390990495681763,
|
| 497 |
+
"learning_rate": 0.00025057660626029654,
|
| 498 |
+
"loss": 0.1866,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.16706865109712335,
|
| 503 |
+
"grad_norm": 1.4517531394958496,
|
| 504 |
+
"learning_rate": 0.0002498705577783007,
|
| 505 |
+
"loss": 0.1928,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.1694217306900406,
|
| 510 |
+
"grad_norm": 1.4081028699874878,
|
| 511 |
+
"learning_rate": 0.000249164509296305,
|
| 512 |
+
"loss": 0.1714,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.17177481028295782,
|
| 517 |
+
"grad_norm": 1.357934832572937,
|
| 518 |
+
"learning_rate": 0.00024845846081430925,
|
| 519 |
+
"loss": 0.1856,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.17412788987587505,
|
| 524 |
+
"grad_norm": 1.043090581893921,
|
| 525 |
+
"learning_rate": 0.0002477524123323135,
|
| 526 |
+
"loss": 0.1495,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.17648096946879227,
|
| 531 |
+
"grad_norm": 1.2053163051605225,
|
| 532 |
+
"learning_rate": 0.0002470463638503177,
|
| 533 |
+
"loss": 0.1517,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.17883404906170952,
|
| 538 |
+
"grad_norm": 2.3474409580230713,
|
| 539 |
+
"learning_rate": 0.00024634031536832195,
|
| 540 |
+
"loss": 0.1882,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.18118712865462674,
|
| 545 |
+
"grad_norm": 0.8380926847457886,
|
| 546 |
+
"learning_rate": 0.0002456342668863262,
|
| 547 |
+
"loss": 0.1796,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18354020824754397,
|
| 552 |
+
"grad_norm": 1.3997254371643066,
|
| 553 |
+
"learning_rate": 0.0002449282184043304,
|
| 554 |
+
"loss": 0.1843,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.1858932878404612,
|
| 559 |
+
"grad_norm": 1.3143609762191772,
|
| 560 |
+
"learning_rate": 0.00024422216992233466,
|
| 561 |
+
"loss": 0.1629,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.18824636743337844,
|
| 566 |
+
"grad_norm": 0.9414114952087402,
|
| 567 |
+
"learning_rate": 0.0002435161214403389,
|
| 568 |
+
"loss": 0.1583,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.19059944702629567,
|
| 573 |
+
"grad_norm": 1.0523838996887207,
|
| 574 |
+
"learning_rate": 0.0002428100729583431,
|
| 575 |
+
"loss": 0.194,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.1929525266192129,
|
| 580 |
+
"grad_norm": 1.0871750116348267,
|
| 581 |
+
"learning_rate": 0.00024210402447634737,
|
| 582 |
+
"loss": 0.15,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.1953056062121301,
|
| 587 |
+
"grad_norm": 1.515932321548462,
|
| 588 |
+
"learning_rate": 0.0002413979759943516,
|
| 589 |
+
"loss": 0.1895,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.19765868580504736,
|
| 594 |
+
"grad_norm": 0.7211456298828125,
|
| 595 |
+
"learning_rate": 0.00024069192751235584,
|
| 596 |
+
"loss": 0.1685,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.20001176539796459,
|
| 601 |
+
"grad_norm": 0.7664592862129211,
|
| 602 |
+
"learning_rate": 0.00023998587903036007,
|
| 603 |
+
"loss": 0.1658,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.2023648449908818,
|
| 608 |
+
"grad_norm": 0.8728657960891724,
|
| 609 |
+
"learning_rate": 0.0002392798305483643,
|
| 610 |
+
"loss": 0.1468,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.20471792458379906,
|
| 615 |
+
"grad_norm": 1.3027325868606567,
|
| 616 |
+
"learning_rate": 0.00023857378206636854,
|
| 617 |
+
"loss": 0.1633,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.20707100417671628,
|
| 622 |
+
"grad_norm": 1.1061084270477295,
|
| 623 |
+
"learning_rate": 0.00023786773358437278,
|
| 624 |
+
"loss": 0.1861,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.2094240837696335,
|
| 629 |
+
"grad_norm": 1.176365613937378,
|
| 630 |
+
"learning_rate": 0.000237161685102377,
|
| 631 |
+
"loss": 0.1589,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.21177716336255073,
|
| 636 |
+
"grad_norm": 0.8307468295097351,
|
| 637 |
+
"learning_rate": 0.00023645563662038127,
|
| 638 |
+
"loss": 0.172,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.21413024295546798,
|
| 643 |
+
"grad_norm": 1.2759816646575928,
|
| 644 |
+
"learning_rate": 0.00023574958813838548,
|
| 645 |
+
"loss": 0.1475,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.2164833225483852,
|
| 650 |
+
"grad_norm": 1.661071538925171,
|
| 651 |
+
"learning_rate": 0.00023504353965638974,
|
| 652 |
+
"loss": 0.2048,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.21883640214130243,
|
| 657 |
+
"grad_norm": 1.3144210577011108,
|
| 658 |
+
"learning_rate": 0.00023433749117439395,
|
| 659 |
+
"loss": 0.1582,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.22118948173421965,
|
| 664 |
+
"grad_norm": 1.1830146312713623,
|
| 665 |
+
"learning_rate": 0.00023363144269239821,
|
| 666 |
+
"loss": 0.1567,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.2235425613271369,
|
| 671 |
+
"grad_norm": 0.7755473256111145,
|
| 672 |
+
"learning_rate": 0.00023292539421040242,
|
| 673 |
+
"loss": 0.1369,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.22589564092005412,
|
| 678 |
+
"grad_norm": 0.708152711391449,
|
| 679 |
+
"learning_rate": 0.00023221934572840666,
|
| 680 |
+
"loss": 0.1477,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.22824872051297135,
|
| 685 |
+
"grad_norm": 0.9567592144012451,
|
| 686 |
+
"learning_rate": 0.0002315132972464109,
|
| 687 |
+
"loss": 0.1685,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.23060180010588857,
|
| 692 |
+
"grad_norm": 1.019717812538147,
|
| 693 |
+
"learning_rate": 0.00023080724876441513,
|
| 694 |
+
"loss": 0.1485,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.23295487969880582,
|
| 699 |
+
"grad_norm": 3.704050302505493,
|
| 700 |
+
"learning_rate": 0.00023010120028241936,
|
| 701 |
+
"loss": 0.182,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.23530795929172305,
|
| 706 |
+
"grad_norm": 1.3113001585006714,
|
| 707 |
+
"learning_rate": 0.0002293951518004236,
|
| 708 |
+
"loss": 0.2033,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.23766103888464027,
|
| 713 |
+
"grad_norm": 1.586300253868103,
|
| 714 |
+
"learning_rate": 0.00022868910331842783,
|
| 715 |
+
"loss": 0.1658,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.2400141184775575,
|
| 720 |
+
"grad_norm": 0.7029755711555481,
|
| 721 |
+
"learning_rate": 0.00022798305483643207,
|
| 722 |
+
"loss": 0.1675,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.24236719807047474,
|
| 727 |
+
"grad_norm": 0.9558175802230835,
|
| 728 |
+
"learning_rate": 0.00022727700635443633,
|
| 729 |
+
"loss": 0.1292,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.24472027766339197,
|
| 734 |
+
"grad_norm": 0.552598774433136,
|
| 735 |
+
"learning_rate": 0.00022657095787244054,
|
| 736 |
+
"loss": 0.1271,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.2470733572563092,
|
| 741 |
+
"grad_norm": 1.160657525062561,
|
| 742 |
+
"learning_rate": 0.0002258649093904448,
|
| 743 |
+
"loss": 0.1406,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.2494264368492264,
|
| 748 |
+
"grad_norm": 0.9359754323959351,
|
| 749 |
+
"learning_rate": 0.000225158860908449,
|
| 750 |
+
"loss": 0.1456,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.25177951644214364,
|
| 755 |
+
"grad_norm": 0.6799198985099792,
|
| 756 |
+
"learning_rate": 0.00022445281242645327,
|
| 757 |
+
"loss": 0.1235,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.2541325960350609,
|
| 762 |
+
"grad_norm": 0.97700434923172,
|
| 763 |
+
"learning_rate": 0.00022374676394445748,
|
| 764 |
+
"loss": 0.1721,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.25648567562797814,
|
| 769 |
+
"grad_norm": 0.6762118935585022,
|
| 770 |
+
"learning_rate": 0.00022304071546246174,
|
| 771 |
+
"loss": 0.2006,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.25883875522089533,
|
| 776 |
+
"grad_norm": 0.6071228384971619,
|
| 777 |
+
"learning_rate": 0.00022233466698046595,
|
| 778 |
+
"loss": 0.1647,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.2611918348138126,
|
| 783 |
+
"grad_norm": 0.7097590565681458,
|
| 784 |
+
"learning_rate": 0.0002216286184984702,
|
| 785 |
+
"loss": 0.1662,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.2635449144067298,
|
| 790 |
+
"grad_norm": 0.48786184191703796,
|
| 791 |
+
"learning_rate": 0.00022092257001647442,
|
| 792 |
+
"loss": 0.1378,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.26589799399964703,
|
| 797 |
+
"grad_norm": 0.7238913178443909,
|
| 798 |
+
"learning_rate": 0.00022021652153447868,
|
| 799 |
+
"loss": 0.163,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.2682510735925643,
|
| 804 |
+
"grad_norm": 1.3571726083755493,
|
| 805 |
+
"learning_rate": 0.00021951047305248292,
|
| 806 |
+
"loss": 0.1545,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.2706041531854815,
|
| 811 |
+
"grad_norm": 0.6683372259140015,
|
| 812 |
+
"learning_rate": 0.00021880442457048715,
|
| 813 |
+
"loss": 0.1375,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.27295723277839873,
|
| 818 |
+
"grad_norm": 1.9159690141677856,
|
| 819 |
+
"learning_rate": 0.0002180983760884914,
|
| 820 |
+
"loss": 0.1604,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.275310312371316,
|
| 825 |
+
"grad_norm": 1.6136759519577026,
|
| 826 |
+
"learning_rate": 0.00021739232760649562,
|
| 827 |
+
"loss": 0.1827,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.2776633919642332,
|
| 832 |
+
"grad_norm": 1.2445416450500488,
|
| 833 |
+
"learning_rate": 0.00021668627912449986,
|
| 834 |
+
"loss": 0.1283,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2800164715571504,
|
| 839 |
+
"grad_norm": 1.143410563468933,
|
| 840 |
+
"learning_rate": 0.0002159802306425041,
|
| 841 |
+
"loss": 0.1571,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.2823695511500677,
|
| 846 |
+
"grad_norm": 0.641952633857727,
|
| 847 |
+
"learning_rate": 0.00021527418216050833,
|
| 848 |
+
"loss": 0.1511,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.2847226307429849,
|
| 853 |
+
"grad_norm": 0.9618122577667236,
|
| 854 |
+
"learning_rate": 0.00021456813367851256,
|
| 855 |
+
"loss": 0.1251,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.2870757103359021,
|
| 860 |
+
"grad_norm": 1.040390133857727,
|
| 861 |
+
"learning_rate": 0.0002138620851965168,
|
| 862 |
+
"loss": 0.1481,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.2894287899288193,
|
| 867 |
+
"grad_norm": 2.470360279083252,
|
| 868 |
+
"learning_rate": 0.00021315603671452106,
|
| 869 |
+
"loss": 0.1523,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.29178186952173657,
|
| 874 |
+
"grad_norm": 1.15378737449646,
|
| 875 |
+
"learning_rate": 0.00021244998823252527,
|
| 876 |
+
"loss": 0.1526,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.2941349491146538,
|
| 881 |
+
"grad_norm": 1.2236779928207397,
|
| 882 |
+
"learning_rate": 0.00021174393975052953,
|
| 883 |
+
"loss": 0.15,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.296488028707571,
|
| 888 |
+
"grad_norm": 0.6974225640296936,
|
| 889 |
+
"learning_rate": 0.00021103789126853374,
|
| 890 |
+
"loss": 0.1529,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.29884110830048827,
|
| 895 |
+
"grad_norm": 1.2019627094268799,
|
| 896 |
+
"learning_rate": 0.000210331842786538,
|
| 897 |
+
"loss": 0.1534,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.3011941878934055,
|
| 902 |
+
"grad_norm": 1.5245829820632935,
|
| 903 |
+
"learning_rate": 0.0002096257943045422,
|
| 904 |
+
"loss": 0.1452,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.3035472674863227,
|
| 909 |
+
"grad_norm": 1.5062931776046753,
|
| 910 |
+
"learning_rate": 0.00020891974582254647,
|
| 911 |
+
"loss": 0.1617,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.30590034707923996,
|
| 916 |
+
"grad_norm": 0.5989176034927368,
|
| 917 |
+
"learning_rate": 0.00020821369734055068,
|
| 918 |
+
"loss": 0.1567,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.30825342667215716,
|
| 923 |
+
"grad_norm": 1.1063286066055298,
|
| 924 |
+
"learning_rate": 0.00020750764885855494,
|
| 925 |
+
"loss": 0.1651,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.3106065062650744,
|
| 930 |
+
"grad_norm": 0.9815717935562134,
|
| 931 |
+
"learning_rate": 0.00020680160037655915,
|
| 932 |
+
"loss": 0.1485,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.31295958585799166,
|
| 937 |
+
"grad_norm": 1.218807578086853,
|
| 938 |
+
"learning_rate": 0.0002060955518945634,
|
| 939 |
+
"loss": 0.1151,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.31531266545090886,
|
| 944 |
+
"grad_norm": 1.1629014015197754,
|
| 945 |
+
"learning_rate": 0.00020538950341256765,
|
| 946 |
+
"loss": 0.1406,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.3176657450438261,
|
| 951 |
+
"grad_norm": 0.6818956732749939,
|
| 952 |
+
"learning_rate": 0.00020468345493057188,
|
| 953 |
+
"loss": 0.1465,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.32001882463674336,
|
| 958 |
+
"grad_norm": 0.7869308590888977,
|
| 959 |
+
"learning_rate": 0.00020397740644857612,
|
| 960 |
+
"loss": 0.1515,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.32237190422966056,
|
| 965 |
+
"grad_norm": 1.023478627204895,
|
| 966 |
+
"learning_rate": 0.00020327135796658035,
|
| 967 |
+
"loss": 0.1781,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.3247249838225778,
|
| 972 |
+
"grad_norm": 1.0383384227752686,
|
| 973 |
+
"learning_rate": 0.0002025653094845846,
|
| 974 |
+
"loss": 0.1195,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.327078063415495,
|
| 979 |
+
"grad_norm": 1.5291595458984375,
|
| 980 |
+
"learning_rate": 0.00020185926100258882,
|
| 981 |
+
"loss": 0.1334,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32943114300841225,
|
| 986 |
+
"grad_norm": 0.9488996267318726,
|
| 987 |
+
"learning_rate": 0.00020115321252059306,
|
| 988 |
+
"loss": 0.1368,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.3317842226013295,
|
| 993 |
+
"grad_norm": 1.1703331470489502,
|
| 994 |
+
"learning_rate": 0.0002004471640385973,
|
| 995 |
+
"loss": 0.131,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.3341373021942467,
|
| 1000 |
+
"grad_norm": 0.6122885346412659,
|
| 1001 |
+
"learning_rate": 0.00019974111555660153,
|
| 1002 |
+
"loss": 0.1356,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.33649038178716395,
|
| 1007 |
+
"grad_norm": 0.7869921326637268,
|
| 1008 |
+
"learning_rate": 0.0001990350670746058,
|
| 1009 |
+
"loss": 0.1817,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3388434613800812,
|
| 1014 |
+
"grad_norm": 0.691066324710846,
|
| 1015 |
+
"learning_rate": 0.00019832901859261,
|
| 1016 |
+
"loss": 0.131,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.3411965409729984,
|
| 1021 |
+
"grad_norm": 1.4205127954483032,
|
| 1022 |
+
"learning_rate": 0.00019762297011061426,
|
| 1023 |
+
"loss": 0.1366,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.34354962056591565,
|
| 1028 |
+
"grad_norm": 0.47127053141593933,
|
| 1029 |
+
"learning_rate": 0.00019691692162861847,
|
| 1030 |
+
"loss": 0.1498,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3459027001588329,
|
| 1035 |
+
"grad_norm": 0.9336820840835571,
|
| 1036 |
+
"learning_rate": 0.00019621087314662273,
|
| 1037 |
+
"loss": 0.1512,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.3482557797517501,
|
| 1042 |
+
"grad_norm": 0.8124200105667114,
|
| 1043 |
+
"learning_rate": 0.00019550482466462694,
|
| 1044 |
+
"loss": 0.1319,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.35060885934466735,
|
| 1049 |
+
"grad_norm": 0.6921178698539734,
|
| 1050 |
+
"learning_rate": 0.0001947987761826312,
|
| 1051 |
+
"loss": 0.1279,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.35296193893758454,
|
| 1056 |
+
"grad_norm": 1.336229681968689,
|
| 1057 |
+
"learning_rate": 0.0001940927277006354,
|
| 1058 |
+
"loss": 0.1251,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3553150185305018,
|
| 1063 |
+
"grad_norm": 0.9984803795814514,
|
| 1064 |
+
"learning_rate": 0.00019338667921863967,
|
| 1065 |
+
"loss": 0.1299,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.35766809812341904,
|
| 1070 |
+
"grad_norm": 1.0903042554855347,
|
| 1071 |
+
"learning_rate": 0.00019268063073664388,
|
| 1072 |
+
"loss": 0.1528,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.36002117771633624,
|
| 1077 |
+
"grad_norm": 0.666950523853302,
|
| 1078 |
+
"learning_rate": 0.00019197458225464814,
|
| 1079 |
+
"loss": 0.1446,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.3623742573092535,
|
| 1084 |
+
"grad_norm": 0.8104845285415649,
|
| 1085 |
+
"learning_rate": 0.00019126853377265238,
|
| 1086 |
+
"loss": 0.1221,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.36472733690217074,
|
| 1091 |
+
"grad_norm": 0.5904582738876343,
|
| 1092 |
+
"learning_rate": 0.00019056248529065661,
|
| 1093 |
+
"loss": 0.1164,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.36708041649508794,
|
| 1098 |
+
"grad_norm": 0.7703972458839417,
|
| 1099 |
+
"learning_rate": 0.00018985643680866085,
|
| 1100 |
+
"loss": 0.0978,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.3694334960880052,
|
| 1105 |
+
"grad_norm": 1.9245415925979614,
|
| 1106 |
+
"learning_rate": 0.00018915038832666508,
|
| 1107 |
+
"loss": 0.1624,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.3717865756809224,
|
| 1112 |
+
"grad_norm": 1.6459194421768188,
|
| 1113 |
+
"learning_rate": 0.00018844433984466932,
|
| 1114 |
+
"loss": 0.1289,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.37413965527383963,
|
| 1119 |
+
"grad_norm": 1.6774044036865234,
|
| 1120 |
+
"learning_rate": 0.00018773829136267355,
|
| 1121 |
+
"loss": 0.1468,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.3764927348667569,
|
| 1126 |
+
"grad_norm": 1.5878580808639526,
|
| 1127 |
+
"learning_rate": 0.0001870322428806778,
|
| 1128 |
+
"loss": 0.1318,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.3788458144596741,
|
| 1133 |
+
"grad_norm": 0.7039738297462463,
|
| 1134 |
+
"learning_rate": 0.00018632619439868203,
|
| 1135 |
+
"loss": 0.1242,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.38119889405259133,
|
| 1140 |
+
"grad_norm": 1.1770200729370117,
|
| 1141 |
+
"learning_rate": 0.00018562014591668626,
|
| 1142 |
+
"loss": 0.1321,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.3835519736455086,
|
| 1147 |
+
"grad_norm": 2.2201638221740723,
|
| 1148 |
+
"learning_rate": 0.00018491409743469052,
|
| 1149 |
+
"loss": 0.1214,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.3859050532384258,
|
| 1154 |
+
"grad_norm": 0.756149411201477,
|
| 1155 |
+
"learning_rate": 0.00018420804895269473,
|
| 1156 |
+
"loss": 0.1219,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.38825813283134303,
|
| 1161 |
+
"grad_norm": 0.5444088578224182,
|
| 1162 |
+
"learning_rate": 0.000183502000470699,
|
| 1163 |
+
"loss": 0.1346,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3906112124242602,
|
| 1168 |
+
"grad_norm": 0.7643070816993713,
|
| 1169 |
+
"learning_rate": 0.0001827959519887032,
|
| 1170 |
+
"loss": 0.1324,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.3929642920171775,
|
| 1175 |
+
"grad_norm": 0.885362446308136,
|
| 1176 |
+
"learning_rate": 0.00018208990350670746,
|
| 1177 |
+
"loss": 0.1166,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.3953173716100947,
|
| 1182 |
+
"grad_norm": 0.7135679721832275,
|
| 1183 |
+
"learning_rate": 0.00018138385502471167,
|
| 1184 |
+
"loss": 0.1364,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.3976704512030119,
|
| 1189 |
+
"grad_norm": 0.5533025860786438,
|
| 1190 |
+
"learning_rate": 0.00018067780654271593,
|
| 1191 |
+
"loss": 0.1137,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.40002353079592917,
|
| 1196 |
+
"grad_norm": 0.5916281342506409,
|
| 1197 |
+
"learning_rate": 0.00017997175806072014,
|
| 1198 |
+
"loss": 0.1131,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.4023766103888464,
|
| 1203 |
+
"grad_norm": 0.8299354314804077,
|
| 1204 |
+
"learning_rate": 0.0001792657095787244,
|
| 1205 |
+
"loss": 0.1331,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.4047296899817636,
|
| 1210 |
+
"grad_norm": 0.7944399118423462,
|
| 1211 |
+
"learning_rate": 0.0001785596610967286,
|
| 1212 |
+
"loss": 0.1049,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.40708276957468087,
|
| 1217 |
+
"grad_norm": 0.6967952251434326,
|
| 1218 |
+
"learning_rate": 0.00017785361261473287,
|
| 1219 |
+
"loss": 0.0997,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.4094358491675981,
|
| 1224 |
+
"grad_norm": 0.42431318759918213,
|
| 1225 |
+
"learning_rate": 0.0001771475641327371,
|
| 1226 |
+
"loss": 0.0964,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.4117889287605153,
|
| 1231 |
+
"grad_norm": 0.6767364740371704,
|
| 1232 |
+
"learning_rate": 0.00017644151565074134,
|
| 1233 |
+
"loss": 0.1627,
|
| 1234 |
+
"step": 1750
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.41414200835343257,
|
| 1238 |
+
"grad_norm": 1.0430301427841187,
|
| 1239 |
+
"learning_rate": 0.00017573546716874558,
|
| 1240 |
+
"loss": 0.1173,
|
| 1241 |
+
"step": 1760
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.41649508794634976,
|
| 1245 |
+
"grad_norm": 0.6168161034584045,
|
| 1246 |
+
"learning_rate": 0.00017502941868674981,
|
| 1247 |
+
"loss": 0.1229,
|
| 1248 |
+
"step": 1770
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.418848167539267,
|
| 1252 |
+
"grad_norm": 1.9067519903182983,
|
| 1253 |
+
"learning_rate": 0.00017432337020475405,
|
| 1254 |
+
"loss": 0.1369,
|
| 1255 |
+
"step": 1780
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.42120124713218426,
|
| 1259 |
+
"grad_norm": 1.5157831907272339,
|
| 1260 |
+
"learning_rate": 0.00017361732172275829,
|
| 1261 |
+
"loss": 0.1243,
|
| 1262 |
+
"step": 1790
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.42355432672510146,
|
| 1266 |
+
"grad_norm": 1.5152102708816528,
|
| 1267 |
+
"learning_rate": 0.00017291127324076252,
|
| 1268 |
+
"loss": 0.1395,
|
| 1269 |
+
"step": 1800
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.4259074063180187,
|
| 1273 |
+
"grad_norm": 0.8262742161750793,
|
| 1274 |
+
"learning_rate": 0.00017220522475876676,
|
| 1275 |
+
"loss": 0.1467,
|
| 1276 |
+
"step": 1810
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.42826048591093596,
|
| 1280 |
+
"grad_norm": 0.5484256744384766,
|
| 1281 |
+
"learning_rate": 0.000171499176276771,
|
| 1282 |
+
"loss": 0.1405,
|
| 1283 |
+
"step": 1820
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.43061356550385316,
|
| 1287 |
+
"grad_norm": 0.7796267867088318,
|
| 1288 |
+
"learning_rate": 0.00017079312779477525,
|
| 1289 |
+
"loss": 0.1508,
|
| 1290 |
+
"step": 1830
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.4329666450967704,
|
| 1294 |
+
"grad_norm": 0.7360082268714905,
|
| 1295 |
+
"learning_rate": 0.00017008707931277946,
|
| 1296 |
+
"loss": 0.1332,
|
| 1297 |
+
"step": 1840
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.4353197246896876,
|
| 1301 |
+
"grad_norm": 0.8352281451225281,
|
| 1302 |
+
"learning_rate": 0.00016938103083078372,
|
| 1303 |
+
"loss": 0.1343,
|
| 1304 |
+
"step": 1850
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.43767280428260485,
|
| 1308 |
+
"grad_norm": 0.6898388862609863,
|
| 1309 |
+
"learning_rate": 0.00016867498234878793,
|
| 1310 |
+
"loss": 0.0983,
|
| 1311 |
+
"step": 1860
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.4400258838755221,
|
| 1315 |
+
"grad_norm": 0.3843238651752472,
|
| 1316 |
+
"learning_rate": 0.0001679689338667922,
|
| 1317 |
+
"loss": 0.1091,
|
| 1318 |
+
"step": 1870
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.4423789634684393,
|
| 1322 |
+
"grad_norm": 0.7791532278060913,
|
| 1323 |
+
"learning_rate": 0.0001672628853847964,
|
| 1324 |
+
"loss": 0.1321,
|
| 1325 |
+
"step": 1880
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.44473204306135655,
|
| 1329 |
+
"grad_norm": 0.9906323552131653,
|
| 1330 |
+
"learning_rate": 0.00016655683690280064,
|
| 1331 |
+
"loss": 0.1125,
|
| 1332 |
+
"step": 1890
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.4470851226542738,
|
| 1336 |
+
"grad_norm": 0.631594181060791,
|
| 1337 |
+
"learning_rate": 0.00016585078842080487,
|
| 1338 |
+
"loss": 0.1328,
|
| 1339 |
+
"step": 1900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.449438202247191,
|
| 1343 |
+
"grad_norm": 1.4922380447387695,
|
| 1344 |
+
"learning_rate": 0.0001651447399388091,
|
| 1345 |
+
"loss": 0.1441,
|
| 1346 |
+
"step": 1910
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.45179128184010825,
|
| 1350 |
+
"grad_norm": 0.6896445751190186,
|
| 1351 |
+
"learning_rate": 0.00016443869145681334,
|
| 1352 |
+
"loss": 0.146,
|
| 1353 |
+
"step": 1920
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.45414436143302545,
|
| 1357 |
+
"grad_norm": 0.6470409035682678,
|
| 1358 |
+
"learning_rate": 0.00016373264297481758,
|
| 1359 |
+
"loss": 0.1123,
|
| 1360 |
+
"step": 1930
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.4564974410259427,
|
| 1364 |
+
"grad_norm": 1.4532804489135742,
|
| 1365 |
+
"learning_rate": 0.00016302659449282184,
|
| 1366 |
+
"loss": 0.1204,
|
| 1367 |
+
"step": 1940
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.45885052061885995,
|
| 1371 |
+
"grad_norm": 1.5582534074783325,
|
| 1372 |
+
"learning_rate": 0.00016232054601082605,
|
| 1373 |
+
"loss": 0.1275,
|
| 1374 |
+
"step": 1950
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.46120360021177714,
|
| 1378 |
+
"grad_norm": 0.7568921446800232,
|
| 1379 |
+
"learning_rate": 0.0001616144975288303,
|
| 1380 |
+
"loss": 0.1373,
|
| 1381 |
+
"step": 1960
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4635566798046944,
|
| 1385 |
+
"grad_norm": 0.7904714941978455,
|
| 1386 |
+
"learning_rate": 0.00016090844904683452,
|
| 1387 |
+
"loss": 0.1281,
|
| 1388 |
+
"step": 1970
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.46590975939761164,
|
| 1392 |
+
"grad_norm": 0.48104897141456604,
|
| 1393 |
+
"learning_rate": 0.00016020240056483878,
|
| 1394 |
+
"loss": 0.1173,
|
| 1395 |
+
"step": 1980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.46826283899052884,
|
| 1399 |
+
"grad_norm": 0.6676899194717407,
|
| 1400 |
+
"learning_rate": 0.000159496352082843,
|
| 1401 |
+
"loss": 0.1297,
|
| 1402 |
+
"step": 1990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.4706159185834461,
|
| 1406 |
+
"grad_norm": 0.7035501599311829,
|
| 1407 |
+
"learning_rate": 0.00015879030360084725,
|
| 1408 |
+
"loss": 0.1246,
|
| 1409 |
+
"step": 2000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.47296899817636334,
|
| 1413 |
+
"grad_norm": 1.289421796798706,
|
| 1414 |
+
"learning_rate": 0.00015808425511885146,
|
| 1415 |
+
"loss": 0.1501,
|
| 1416 |
+
"step": 2010
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.47532207776928054,
|
| 1420 |
+
"grad_norm": 0.6186831593513489,
|
| 1421 |
+
"learning_rate": 0.00015737820663685572,
|
| 1422 |
+
"loss": 0.1156,
|
| 1423 |
+
"step": 2020
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.4776751573621978,
|
| 1427 |
+
"grad_norm": 0.7897233963012695,
|
| 1428 |
+
"learning_rate": 0.00015667215815485993,
|
| 1429 |
+
"loss": 0.1317,
|
| 1430 |
+
"step": 2030
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.480028236955115,
|
| 1434 |
+
"grad_norm": 1.1652599573135376,
|
| 1435 |
+
"learning_rate": 0.0001559661096728642,
|
| 1436 |
+
"loss": 0.1325,
|
| 1437 |
+
"step": 2040
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.48238131654803224,
|
| 1441 |
+
"grad_norm": 0.6400769948959351,
|
| 1442 |
+
"learning_rate": 0.0001552600611908684,
|
| 1443 |
+
"loss": 0.1002,
|
| 1444 |
+
"step": 2050
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4847343961409495,
|
| 1448 |
+
"grad_norm": 0.5541133880615234,
|
| 1449 |
+
"learning_rate": 0.00015455401270887266,
|
| 1450 |
+
"loss": 0.1232,
|
| 1451 |
+
"step": 2060
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.4870874757338667,
|
| 1455 |
+
"grad_norm": 0.605411946773529,
|
| 1456 |
+
"learning_rate": 0.0001538479642268769,
|
| 1457 |
+
"loss": 0.1102,
|
| 1458 |
+
"step": 2070
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.48944055532678393,
|
| 1462 |
+
"grad_norm": 0.49058374762535095,
|
| 1463 |
+
"learning_rate": 0.00015314191574488113,
|
| 1464 |
+
"loss": 0.1228,
|
| 1465 |
+
"step": 2080
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.4917936349197012,
|
| 1469 |
+
"grad_norm": 0.7565241456031799,
|
| 1470 |
+
"learning_rate": 0.00015243586726288537,
|
| 1471 |
+
"loss": 0.1009,
|
| 1472 |
+
"step": 2090
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.4941467145126184,
|
| 1476 |
+
"grad_norm": 0.4517477750778198,
|
| 1477 |
+
"learning_rate": 0.0001517298187808896,
|
| 1478 |
+
"loss": 0.129,
|
| 1479 |
+
"step": 2100
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.49649979410553563,
|
| 1483 |
+
"grad_norm": 0.7871853709220886,
|
| 1484 |
+
"learning_rate": 0.00015102377029889384,
|
| 1485 |
+
"loss": 0.1049,
|
| 1486 |
+
"step": 2110
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.4988528736984528,
|
| 1490 |
+
"grad_norm": 0.4314168691635132,
|
| 1491 |
+
"learning_rate": 0.00015031772181689807,
|
| 1492 |
+
"loss": 0.117,
|
| 1493 |
+
"step": 2120
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.5012059532913701,
|
| 1497 |
+
"grad_norm": 0.8347052335739136,
|
| 1498 |
+
"learning_rate": 0.0001496116733349023,
|
| 1499 |
+
"loss": 0.1336,
|
| 1500 |
+
"step": 2130
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.5035590328842873,
|
| 1504 |
+
"grad_norm": 0.42039480805397034,
|
| 1505 |
+
"learning_rate": 0.00014890562485290657,
|
| 1506 |
+
"loss": 0.1176,
|
| 1507 |
+
"step": 2140
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.5059121124772046,
|
| 1511 |
+
"grad_norm": 1.1371684074401855,
|
| 1512 |
+
"learning_rate": 0.0001481995763709108,
|
| 1513 |
+
"loss": 0.1362,
|
| 1514 |
+
"step": 2150
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.5082651920701218,
|
| 1518 |
+
"grad_norm": 0.8690921664237976,
|
| 1519 |
+
"learning_rate": 0.00014749352788891504,
|
| 1520 |
+
"loss": 0.1444,
|
| 1521 |
+
"step": 2160
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.510618271663039,
|
| 1525 |
+
"grad_norm": 0.3952578604221344,
|
| 1526 |
+
"learning_rate": 0.00014678747940691928,
|
| 1527 |
+
"loss": 0.1411,
|
| 1528 |
+
"step": 2170
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.5129713512559563,
|
| 1532 |
+
"grad_norm": 1.0104624032974243,
|
| 1533 |
+
"learning_rate": 0.0001460814309249235,
|
| 1534 |
+
"loss": 0.1127,
|
| 1535 |
+
"step": 2180
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.5153244308488735,
|
| 1539 |
+
"grad_norm": 0.7708560824394226,
|
| 1540 |
+
"learning_rate": 0.00014537538244292775,
|
| 1541 |
+
"loss": 0.1364,
|
| 1542 |
+
"step": 2190
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5176775104417907,
|
| 1546 |
+
"grad_norm": 3.323113203048706,
|
| 1547 |
+
"learning_rate": 0.00014466933396093198,
|
| 1548 |
+
"loss": 0.1326,
|
| 1549 |
+
"step": 2200
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.520030590034708,
|
| 1553 |
+
"grad_norm": 0.5021951198577881,
|
| 1554 |
+
"learning_rate": 0.00014396328547893622,
|
| 1555 |
+
"loss": 0.1075,
|
| 1556 |
+
"step": 2210
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.5223836696276252,
|
| 1560 |
+
"grad_norm": 0.5558544397354126,
|
| 1561 |
+
"learning_rate": 0.00014325723699694045,
|
| 1562 |
+
"loss": 0.1019,
|
| 1563 |
+
"step": 2220
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.5247367492205424,
|
| 1567 |
+
"grad_norm": 0.7476164102554321,
|
| 1568 |
+
"learning_rate": 0.0001425511885149447,
|
| 1569 |
+
"loss": 0.108,
|
| 1570 |
+
"step": 2230
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.5270898288134596,
|
| 1574 |
+
"grad_norm": 0.8783542513847351,
|
| 1575 |
+
"learning_rate": 0.00014184514003294892,
|
| 1576 |
+
"loss": 0.1182,
|
| 1577 |
+
"step": 2240
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.5294429084063769,
|
| 1581 |
+
"grad_norm": 0.5716719627380371,
|
| 1582 |
+
"learning_rate": 0.00014113909155095316,
|
| 1583 |
+
"loss": 0.1048,
|
| 1584 |
+
"step": 2250
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.5317959879992941,
|
| 1588 |
+
"grad_norm": 0.41919055581092834,
|
| 1589 |
+
"learning_rate": 0.0001404330430689574,
|
| 1590 |
+
"loss": 0.1071,
|
| 1591 |
+
"step": 2260
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.5341490675922113,
|
| 1595 |
+
"grad_norm": 0.672885537147522,
|
| 1596 |
+
"learning_rate": 0.00013972699458696163,
|
| 1597 |
+
"loss": 0.1333,
|
| 1598 |
+
"step": 2270
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.5365021471851286,
|
| 1602 |
+
"grad_norm": 0.7414030432701111,
|
| 1603 |
+
"learning_rate": 0.00013902094610496586,
|
| 1604 |
+
"loss": 0.1288,
|
| 1605 |
+
"step": 2280
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.5388552267780458,
|
| 1609 |
+
"grad_norm": 1.1601518392562866,
|
| 1610 |
+
"learning_rate": 0.0001383148976229701,
|
| 1611 |
+
"loss": 0.1099,
|
| 1612 |
+
"step": 2290
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.541208306370963,
|
| 1616 |
+
"grad_norm": 0.4423375129699707,
|
| 1617 |
+
"learning_rate": 0.00013760884914097433,
|
| 1618 |
+
"loss": 0.1049,
|
| 1619 |
+
"step": 2300
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.5435613859638803,
|
| 1623 |
+
"grad_norm": 0.9248809218406677,
|
| 1624 |
+
"learning_rate": 0.00013690280065897857,
|
| 1625 |
+
"loss": 0.1172,
|
| 1626 |
+
"step": 2310
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.5459144655567975,
|
| 1630 |
+
"grad_norm": 1.3502943515777588,
|
| 1631 |
+
"learning_rate": 0.0001361967521769828,
|
| 1632 |
+
"loss": 0.1303,
|
| 1633 |
+
"step": 2320
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.5482675451497147,
|
| 1637 |
+
"grad_norm": 1.488297939300537,
|
| 1638 |
+
"learning_rate": 0.00013549070369498704,
|
| 1639 |
+
"loss": 0.1182,
|
| 1640 |
+
"step": 2330
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.550620624742632,
|
| 1644 |
+
"grad_norm": 0.6636572480201721,
|
| 1645 |
+
"learning_rate": 0.0001347846552129913,
|
| 1646 |
+
"loss": 0.1233,
|
| 1647 |
+
"step": 2340
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.5529737043355492,
|
| 1651 |
+
"grad_norm": 0.5864549279212952,
|
| 1652 |
+
"learning_rate": 0.00013407860673099554,
|
| 1653 |
+
"loss": 0.1102,
|
| 1654 |
+
"step": 2350
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.5553267839284663,
|
| 1658 |
+
"grad_norm": 1.9224406480789185,
|
| 1659 |
+
"learning_rate": 0.00013337255824899977,
|
| 1660 |
+
"loss": 0.1449,
|
| 1661 |
+
"step": 2360
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.5576798635213837,
|
| 1665 |
+
"grad_norm": 1.1239560842514038,
|
| 1666 |
+
"learning_rate": 0.00013266650976700398,
|
| 1667 |
+
"loss": 0.1155,
|
| 1668 |
+
"step": 2370
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.5600329431143009,
|
| 1672 |
+
"grad_norm": 0.6336050629615784,
|
| 1673 |
+
"learning_rate": 0.00013196046128500821,
|
| 1674 |
+
"loss": 0.1193,
|
| 1675 |
+
"step": 2380
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.562386022707218,
|
| 1679 |
+
"grad_norm": 0.9129360914230347,
|
| 1680 |
+
"learning_rate": 0.00013125441280301245,
|
| 1681 |
+
"loss": 0.1121,
|
| 1682 |
+
"step": 2390
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.5647391023001354,
|
| 1686 |
+
"grad_norm": 0.6220555305480957,
|
| 1687 |
+
"learning_rate": 0.00013054836432101668,
|
| 1688 |
+
"loss": 0.1172,
|
| 1689 |
+
"step": 2400
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5670921818930525,
|
| 1693 |
+
"grad_norm": 0.8981531262397766,
|
| 1694 |
+
"learning_rate": 0.00012984231583902092,
|
| 1695 |
+
"loss": 0.1184,
|
| 1696 |
+
"step": 2410
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5694452614859697,
|
| 1700 |
+
"grad_norm": 0.7610392570495605,
|
| 1701 |
+
"learning_rate": 0.00012913626735702515,
|
| 1702 |
+
"loss": 0.1204,
|
| 1703 |
+
"step": 2420
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5717983410788869,
|
| 1707 |
+
"grad_norm": 0.5133729577064514,
|
| 1708 |
+
"learning_rate": 0.0001284302188750294,
|
| 1709 |
+
"loss": 0.1081,
|
| 1710 |
+
"step": 2430
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5741514206718042,
|
| 1714 |
+
"grad_norm": 0.8097817897796631,
|
| 1715 |
+
"learning_rate": 0.00012772417039303363,
|
| 1716 |
+
"loss": 0.1142,
|
| 1717 |
+
"step": 2440
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5765045002647214,
|
| 1721 |
+
"grad_norm": 1.8712083101272583,
|
| 1722 |
+
"learning_rate": 0.00012701812191103786,
|
| 1723 |
+
"loss": 0.1234,
|
| 1724 |
+
"step": 2450
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5788575798576386,
|
| 1728 |
+
"grad_norm": 0.8425026535987854,
|
| 1729 |
+
"learning_rate": 0.00012631207342904212,
|
| 1730 |
+
"loss": 0.1027,
|
| 1731 |
+
"step": 2460
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5812106594505559,
|
| 1735 |
+
"grad_norm": 0.5562009811401367,
|
| 1736 |
+
"learning_rate": 0.00012560602494704636,
|
| 1737 |
+
"loss": 0.0916,
|
| 1738 |
+
"step": 2470
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5835637390434731,
|
| 1742 |
+
"grad_norm": 0.45057183504104614,
|
| 1743 |
+
"learning_rate": 0.0001248999764650506,
|
| 1744 |
+
"loss": 0.1166,
|
| 1745 |
+
"step": 2480
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5859168186363903,
|
| 1749 |
+
"grad_norm": 0.5411068797111511,
|
| 1750 |
+
"learning_rate": 0.00012419392798305483,
|
| 1751 |
+
"loss": 0.1254,
|
| 1752 |
+
"step": 2490
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5882698982293076,
|
| 1756 |
+
"grad_norm": 0.9400952458381653,
|
| 1757 |
+
"learning_rate": 0.00012348787950105906,
|
| 1758 |
+
"loss": 0.1411,
|
| 1759 |
+
"step": 2500
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5906229778222248,
|
| 1763 |
+
"grad_norm": 0.4275170564651489,
|
| 1764 |
+
"learning_rate": 0.0001227818310190633,
|
| 1765 |
+
"loss": 0.1089,
|
| 1766 |
+
"step": 2510
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.592976057415142,
|
| 1770 |
+
"grad_norm": 1.2033214569091797,
|
| 1771 |
+
"learning_rate": 0.00012207578253706753,
|
| 1772 |
+
"loss": 0.108,
|
| 1773 |
+
"step": 2520
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5953291370080593,
|
| 1777 |
+
"grad_norm": 1.257379412651062,
|
| 1778 |
+
"learning_rate": 0.00012136973405507177,
|
| 1779 |
+
"loss": 0.1207,
|
| 1780 |
+
"step": 2530
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5976822166009765,
|
| 1784 |
+
"grad_norm": 0.7070032954216003,
|
| 1785 |
+
"learning_rate": 0.000120663685573076,
|
| 1786 |
+
"loss": 0.0879,
|
| 1787 |
+
"step": 2540
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.6000352961938937,
|
| 1791 |
+
"grad_norm": 0.8550868034362793,
|
| 1792 |
+
"learning_rate": 0.00011995763709108024,
|
| 1793 |
+
"loss": 0.1087,
|
| 1794 |
+
"step": 2550
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.602388375786811,
|
| 1798 |
+
"grad_norm": 0.8301357626914978,
|
| 1799 |
+
"learning_rate": 0.00011925158860908447,
|
| 1800 |
+
"loss": 0.1266,
|
| 1801 |
+
"step": 2560
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.6047414553797282,
|
| 1805 |
+
"grad_norm": 0.4070800542831421,
|
| 1806 |
+
"learning_rate": 0.00011854554012708871,
|
| 1807 |
+
"loss": 0.1062,
|
| 1808 |
+
"step": 2570
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.6070945349726454,
|
| 1812 |
+
"grad_norm": 1.1967391967773438,
|
| 1813 |
+
"learning_rate": 0.00011783949164509294,
|
| 1814 |
+
"loss": 0.1147,
|
| 1815 |
+
"step": 2580
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.6094476145655627,
|
| 1819 |
+
"grad_norm": 0.5281302332878113,
|
| 1820 |
+
"learning_rate": 0.00011713344316309718,
|
| 1821 |
+
"loss": 0.0912,
|
| 1822 |
+
"step": 2590
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.6118006941584799,
|
| 1826 |
+
"grad_norm": 0.5271784067153931,
|
| 1827 |
+
"learning_rate": 0.00011642739468110142,
|
| 1828 |
+
"loss": 0.1084,
|
| 1829 |
+
"step": 2600
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.6141537737513971,
|
| 1833 |
+
"grad_norm": 0.4973151683807373,
|
| 1834 |
+
"learning_rate": 0.00011572134619910566,
|
| 1835 |
+
"loss": 0.1242,
|
| 1836 |
+
"step": 2610
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.6165068533443143,
|
| 1840 |
+
"grad_norm": 0.4281303882598877,
|
| 1841 |
+
"learning_rate": 0.0001150152977171099,
|
| 1842 |
+
"loss": 0.1101,
|
| 1843 |
+
"step": 2620
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.6188599329372316,
|
| 1847 |
+
"grad_norm": 0.5142689347267151,
|
| 1848 |
+
"learning_rate": 0.00011430924923511413,
|
| 1849 |
+
"loss": 0.1182,
|
| 1850 |
+
"step": 2630
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6212130125301488,
|
| 1854 |
+
"grad_norm": 0.5125661492347717,
|
| 1855 |
+
"learning_rate": 0.00011360320075311837,
|
| 1856 |
+
"loss": 0.0943,
|
| 1857 |
+
"step": 2640
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.623566092123066,
|
| 1861 |
+
"grad_norm": 0.43077680468559265,
|
| 1862 |
+
"learning_rate": 0.0001128971522711226,
|
| 1863 |
+
"loss": 0.0948,
|
| 1864 |
+
"step": 2650
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6259191717159833,
|
| 1868 |
+
"grad_norm": 0.5074141621589661,
|
| 1869 |
+
"learning_rate": 0.00011219110378912684,
|
| 1870 |
+
"loss": 0.0853,
|
| 1871 |
+
"step": 2660
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.6282722513089005,
|
| 1875 |
+
"grad_norm": 0.8260855674743652,
|
| 1876 |
+
"learning_rate": 0.00011148505530713107,
|
| 1877 |
+
"loss": 0.1104,
|
| 1878 |
+
"step": 2670
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.6306253309018177,
|
| 1882 |
+
"grad_norm": 0.7819215059280396,
|
| 1883 |
+
"learning_rate": 0.00011077900682513531,
|
| 1884 |
+
"loss": 0.1256,
|
| 1885 |
+
"step": 2680
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.632978410494735,
|
| 1889 |
+
"grad_norm": 0.46884438395500183,
|
| 1890 |
+
"learning_rate": 0.00011007295834313955,
|
| 1891 |
+
"loss": 0.1027,
|
| 1892 |
+
"step": 2690
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6353314900876522,
|
| 1896 |
+
"grad_norm": 0.9515593647956848,
|
| 1897 |
+
"learning_rate": 0.00010936690986114378,
|
| 1898 |
+
"loss": 0.112,
|
| 1899 |
+
"step": 2700
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6376845696805694,
|
| 1903 |
+
"grad_norm": 0.3602767586708069,
|
| 1904 |
+
"learning_rate": 0.00010866086137914803,
|
| 1905 |
+
"loss": 0.1053,
|
| 1906 |
+
"step": 2710
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.6400376492734867,
|
| 1910 |
+
"grad_norm": 0.7740781903266907,
|
| 1911 |
+
"learning_rate": 0.00010795481289715226,
|
| 1912 |
+
"loss": 0.112,
|
| 1913 |
+
"step": 2720
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.6423907288664039,
|
| 1917 |
+
"grad_norm": 0.5003033876419067,
|
| 1918 |
+
"learning_rate": 0.0001072487644151565,
|
| 1919 |
+
"loss": 0.0985,
|
| 1920 |
+
"step": 2730
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.6447438084593211,
|
| 1924 |
+
"grad_norm": 0.4092664122581482,
|
| 1925 |
+
"learning_rate": 0.00010654271593316073,
|
| 1926 |
+
"loss": 0.1113,
|
| 1927 |
+
"step": 2740
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6470968880522384,
|
| 1931 |
+
"grad_norm": 0.446584552526474,
|
| 1932 |
+
"learning_rate": 0.00010583666745116497,
|
| 1933 |
+
"loss": 0.0909,
|
| 1934 |
+
"step": 2750
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.6494499676451556,
|
| 1938 |
+
"grad_norm": 0.3130131661891937,
|
| 1939 |
+
"learning_rate": 0.0001051306189691692,
|
| 1940 |
+
"loss": 0.0954,
|
| 1941 |
+
"step": 2760
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.6518030472380728,
|
| 1945 |
+
"grad_norm": 0.7232083082199097,
|
| 1946 |
+
"learning_rate": 0.00010442457048717344,
|
| 1947 |
+
"loss": 0.1132,
|
| 1948 |
+
"step": 2770
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.65415612683099,
|
| 1952 |
+
"grad_norm": 0.5579691529273987,
|
| 1953 |
+
"learning_rate": 0.00010371852200517768,
|
| 1954 |
+
"loss": 0.1045,
|
| 1955 |
+
"step": 2780
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.6565092064239073,
|
| 1959 |
+
"grad_norm": 0.5319089889526367,
|
| 1960 |
+
"learning_rate": 0.00010301247352318191,
|
| 1961 |
+
"loss": 0.1215,
|
| 1962 |
+
"step": 2790
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6588622860168245,
|
| 1966 |
+
"grad_norm": 0.516445517539978,
|
| 1967 |
+
"learning_rate": 0.00010230642504118615,
|
| 1968 |
+
"loss": 0.111,
|
| 1969 |
+
"step": 2800
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6612153656097417,
|
| 1973 |
+
"grad_norm": 0.25264236330986023,
|
| 1974 |
+
"learning_rate": 0.0001016003765591904,
|
| 1975 |
+
"loss": 0.1126,
|
| 1976 |
+
"step": 2810
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.663568445202659,
|
| 1980 |
+
"grad_norm": 0.7910987138748169,
|
| 1981 |
+
"learning_rate": 0.00010089432807719463,
|
| 1982 |
+
"loss": 0.1306,
|
| 1983 |
+
"step": 2820
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.6659215247955762,
|
| 1987 |
+
"grad_norm": 0.7823461890220642,
|
| 1988 |
+
"learning_rate": 0.00010018827959519886,
|
| 1989 |
+
"loss": 0.0967,
|
| 1990 |
+
"step": 2830
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6682746043884934,
|
| 1994 |
+
"grad_norm": 0.7126127481460571,
|
| 1995 |
+
"learning_rate": 9.94822311132031e-05,
|
| 1996 |
+
"loss": 0.1296,
|
| 1997 |
+
"step": 2840
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6706276839814107,
|
| 2001 |
+
"grad_norm": 0.9327739477157593,
|
| 2002 |
+
"learning_rate": 9.877618263120733e-05,
|
| 2003 |
+
"loss": 0.1115,
|
| 2004 |
+
"step": 2850
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.6729807635743279,
|
| 2008 |
+
"grad_norm": 0.7680268883705139,
|
| 2009 |
+
"learning_rate": 9.807013414921157e-05,
|
| 2010 |
+
"loss": 0.1055,
|
| 2011 |
+
"step": 2860
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.6753338431672451,
|
| 2015 |
+
"grad_norm": 0.7711540460586548,
|
| 2016 |
+
"learning_rate": 9.73640856672158e-05,
|
| 2017 |
+
"loss": 0.0951,
|
| 2018 |
+
"step": 2870
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.6776869227601624,
|
| 2022 |
+
"grad_norm": 0.5041959881782532,
|
| 2023 |
+
"learning_rate": 9.665803718522004e-05,
|
| 2024 |
+
"loss": 0.1087,
|
| 2025 |
+
"step": 2880
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6800400023530796,
|
| 2029 |
+
"grad_norm": 0.5102591514587402,
|
| 2030 |
+
"learning_rate": 9.595198870322428e-05,
|
| 2031 |
+
"loss": 0.1176,
|
| 2032 |
+
"step": 2890
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6823930819459968,
|
| 2036 |
+
"grad_norm": 0.7100384831428528,
|
| 2037 |
+
"learning_rate": 9.524594022122851e-05,
|
| 2038 |
+
"loss": 0.1091,
|
| 2039 |
+
"step": 2900
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6847461615389141,
|
| 2043 |
+
"grad_norm": 0.6806867122650146,
|
| 2044 |
+
"learning_rate": 9.453989173923276e-05,
|
| 2045 |
+
"loss": 0.1251,
|
| 2046 |
+
"step": 2910
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6870992411318313,
|
| 2050 |
+
"grad_norm": 0.6659530401229858,
|
| 2051 |
+
"learning_rate": 9.3833843257237e-05,
|
| 2052 |
+
"loss": 0.0835,
|
| 2053 |
+
"step": 2920
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6894523207247485,
|
| 2057 |
+
"grad_norm": 0.4317012429237366,
|
| 2058 |
+
"learning_rate": 9.312779477524123e-05,
|
| 2059 |
+
"loss": 0.0893,
|
| 2060 |
+
"step": 2930
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6918054003176658,
|
| 2064 |
+
"grad_norm": 0.5916824340820312,
|
| 2065 |
+
"learning_rate": 9.242174629324546e-05,
|
| 2066 |
+
"loss": 0.0905,
|
| 2067 |
+
"step": 2940
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.694158479910583,
|
| 2071 |
+
"grad_norm": 0.7429795265197754,
|
| 2072 |
+
"learning_rate": 9.17156978112497e-05,
|
| 2073 |
+
"loss": 0.1063,
|
| 2074 |
+
"step": 2950
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6965115595035002,
|
| 2078 |
+
"grad_norm": 0.87420254945755,
|
| 2079 |
+
"learning_rate": 9.100964932925394e-05,
|
| 2080 |
+
"loss": 0.1042,
|
| 2081 |
+
"step": 2960
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6988646390964174,
|
| 2085 |
+
"grad_norm": 0.49567267298698425,
|
| 2086 |
+
"learning_rate": 9.030360084725817e-05,
|
| 2087 |
+
"loss": 0.1045,
|
| 2088 |
+
"step": 2970
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.7012177186893347,
|
| 2092 |
+
"grad_norm": 2.633138418197632,
|
| 2093 |
+
"learning_rate": 8.95975523652624e-05,
|
| 2094 |
+
"loss": 0.0884,
|
| 2095 |
+
"step": 2980
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.7035707982822519,
|
| 2099 |
+
"grad_norm": 0.33752286434173584,
|
| 2100 |
+
"learning_rate": 8.889150388326664e-05,
|
| 2101 |
+
"loss": 0.0848,
|
| 2102 |
+
"step": 2990
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.7059238778751691,
|
| 2106 |
+
"grad_norm": 0.5974826812744141,
|
| 2107 |
+
"learning_rate": 8.818545540127088e-05,
|
| 2108 |
+
"loss": 0.0971,
|
| 2109 |
+
"step": 3000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.7082769574680864,
|
| 2113 |
+
"grad_norm": 0.43427976965904236,
|
| 2114 |
+
"learning_rate": 8.747940691927512e-05,
|
| 2115 |
+
"loss": 0.1165,
|
| 2116 |
+
"step": 3010
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.7106300370610036,
|
| 2120 |
+
"grad_norm": 0.7770646810531616,
|
| 2121 |
+
"learning_rate": 8.677335843727936e-05,
|
| 2122 |
+
"loss": 0.1084,
|
| 2123 |
+
"step": 3020
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.7129831166539208,
|
| 2127 |
+
"grad_norm": 0.5276495218276978,
|
| 2128 |
+
"learning_rate": 8.60673099552836e-05,
|
| 2129 |
+
"loss": 0.111,
|
| 2130 |
+
"step": 3030
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.7153361962468381,
|
| 2134 |
+
"grad_norm": 0.9737383127212524,
|
| 2135 |
+
"learning_rate": 8.536126147328783e-05,
|
| 2136 |
+
"loss": 0.0972,
|
| 2137 |
+
"step": 3040
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.7176892758397553,
|
| 2141 |
+
"grad_norm": 0.36562997102737427,
|
| 2142 |
+
"learning_rate": 8.465521299129207e-05,
|
| 2143 |
+
"loss": 0.093,
|
| 2144 |
+
"step": 3050
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.7200423554326725,
|
| 2148 |
+
"grad_norm": 0.8244528770446777,
|
| 2149 |
+
"learning_rate": 8.39491645092963e-05,
|
| 2150 |
+
"loss": 0.1263,
|
| 2151 |
+
"step": 3060
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.7223954350255898,
|
| 2155 |
+
"grad_norm": 1.9532008171081543,
|
| 2156 |
+
"learning_rate": 8.324311602730054e-05,
|
| 2157 |
+
"loss": 0.1251,
|
| 2158 |
+
"step": 3070
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.724748514618507,
|
| 2162 |
+
"grad_norm": 0.572896420955658,
|
| 2163 |
+
"learning_rate": 8.253706754530477e-05,
|
| 2164 |
+
"loss": 0.0875,
|
| 2165 |
+
"step": 3080
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7271015942114242,
|
| 2169 |
+
"grad_norm": 1.2975929975509644,
|
| 2170 |
+
"learning_rate": 8.1831019063309e-05,
|
| 2171 |
+
"loss": 0.1023,
|
| 2172 |
+
"step": 3090
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7294546738043415,
|
| 2176 |
+
"grad_norm": 0.5758102536201477,
|
| 2177 |
+
"learning_rate": 8.112497058131324e-05,
|
| 2178 |
+
"loss": 0.1019,
|
| 2179 |
+
"step": 3100
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7318077533972587,
|
| 2183 |
+
"grad_norm": 0.553327202796936,
|
| 2184 |
+
"learning_rate": 8.041892209931749e-05,
|
| 2185 |
+
"loss": 0.1128,
|
| 2186 |
+
"step": 3110
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7341608329901759,
|
| 2190 |
+
"grad_norm": 0.5465438961982727,
|
| 2191 |
+
"learning_rate": 7.971287361732173e-05,
|
| 2192 |
+
"loss": 0.1237,
|
| 2193 |
+
"step": 3120
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7365139125830932,
|
| 2197 |
+
"grad_norm": 0.46917715668678284,
|
| 2198 |
+
"learning_rate": 7.900682513532596e-05,
|
| 2199 |
+
"loss": 0.0964,
|
| 2200 |
+
"step": 3130
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7388669921760104,
|
| 2204 |
+
"grad_norm": 0.8454899787902832,
|
| 2205 |
+
"learning_rate": 7.83007766533302e-05,
|
| 2206 |
+
"loss": 0.0901,
|
| 2207 |
+
"step": 3140
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7412200717689276,
|
| 2211 |
+
"grad_norm": 0.8698781728744507,
|
| 2212 |
+
"learning_rate": 7.759472817133443e-05,
|
| 2213 |
+
"loss": 0.1119,
|
| 2214 |
+
"step": 3150
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7435731513618448,
|
| 2218 |
+
"grad_norm": 1.7399003505706787,
|
| 2219 |
+
"learning_rate": 7.688867968933867e-05,
|
| 2220 |
+
"loss": 0.1122,
|
| 2221 |
+
"step": 3160
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7459262309547621,
|
| 2225 |
+
"grad_norm": 0.4506986141204834,
|
| 2226 |
+
"learning_rate": 7.61826312073429e-05,
|
| 2227 |
+
"loss": 0.09,
|
| 2228 |
+
"step": 3170
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7482793105476793,
|
| 2232 |
+
"grad_norm": 0.7856936454772949,
|
| 2233 |
+
"learning_rate": 7.547658272534714e-05,
|
| 2234 |
+
"loss": 0.1194,
|
| 2235 |
+
"step": 3180
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7506323901405965,
|
| 2239 |
+
"grad_norm": 0.5778619050979614,
|
| 2240 |
+
"learning_rate": 7.477053424335137e-05,
|
| 2241 |
+
"loss": 0.1167,
|
| 2242 |
+
"step": 3190
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7529854697335138,
|
| 2246 |
+
"grad_norm": 0.4940952658653259,
|
| 2247 |
+
"learning_rate": 7.40644857613556e-05,
|
| 2248 |
+
"loss": 0.0813,
|
| 2249 |
+
"step": 3200
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.755338549326431,
|
| 2253 |
+
"grad_norm": 1.1496696472167969,
|
| 2254 |
+
"learning_rate": 7.335843727935984e-05,
|
| 2255 |
+
"loss": 0.0966,
|
| 2256 |
+
"step": 3210
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7576916289193482,
|
| 2260 |
+
"grad_norm": 0.4551859498023987,
|
| 2261 |
+
"learning_rate": 7.265238879736408e-05,
|
| 2262 |
+
"loss": 0.0956,
|
| 2263 |
+
"step": 3220
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7600447085122655,
|
| 2267 |
+
"grad_norm": 0.5476594567298889,
|
| 2268 |
+
"learning_rate": 7.194634031536831e-05,
|
| 2269 |
+
"loss": 0.1225,
|
| 2270 |
+
"step": 3230
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7623977881051827,
|
| 2274 |
+
"grad_norm": 0.4413054287433624,
|
| 2275 |
+
"learning_rate": 7.124029183337255e-05,
|
| 2276 |
+
"loss": 0.0996,
|
| 2277 |
+
"step": 3240
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.7647508676980999,
|
| 2281 |
+
"grad_norm": 0.6522489190101624,
|
| 2282 |
+
"learning_rate": 7.053424335137678e-05,
|
| 2283 |
+
"loss": 0.0946,
|
| 2284 |
+
"step": 3250
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7671039472910172,
|
| 2288 |
+
"grad_norm": 0.4750779867172241,
|
| 2289 |
+
"learning_rate": 6.982819486938102e-05,
|
| 2290 |
+
"loss": 0.0788,
|
| 2291 |
+
"step": 3260
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.7694570268839344,
|
| 2295 |
+
"grad_norm": 0.336505264043808,
|
| 2296 |
+
"learning_rate": 6.912214638738527e-05,
|
| 2297 |
+
"loss": 0.0841,
|
| 2298 |
+
"step": 3270
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.7718101064768516,
|
| 2302 |
+
"grad_norm": 1.4274874925613403,
|
| 2303 |
+
"learning_rate": 6.84160979053895e-05,
|
| 2304 |
+
"loss": 0.0991,
|
| 2305 |
+
"step": 3280
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7741631860697689,
|
| 2309 |
+
"grad_norm": 0.6464115977287292,
|
| 2310 |
+
"learning_rate": 6.771004942339374e-05,
|
| 2311 |
+
"loss": 0.1172,
|
| 2312 |
+
"step": 3290
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.7765162656626861,
|
| 2316 |
+
"grad_norm": 0.35535725951194763,
|
| 2317 |
+
"learning_rate": 6.700400094139797e-05,
|
| 2318 |
+
"loss": 0.09,
|
| 2319 |
+
"step": 3300
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.7788693452556033,
|
| 2323 |
+
"grad_norm": 0.22626227140426636,
|
| 2324 |
+
"learning_rate": 6.62979524594022e-05,
|
| 2325 |
+
"loss": 0.089,
|
| 2326 |
+
"step": 3310
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.7812224248485204,
|
| 2330 |
+
"grad_norm": 0.6091925501823425,
|
| 2331 |
+
"learning_rate": 6.559190397740644e-05,
|
| 2332 |
+
"loss": 0.0851,
|
| 2333 |
+
"step": 3320
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.7835755044414378,
|
| 2337 |
+
"grad_norm": 2.3381729125976562,
|
| 2338 |
+
"learning_rate": 6.488585549541068e-05,
|
| 2339 |
+
"loss": 0.1001,
|
| 2340 |
+
"step": 3330
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.785928584034355,
|
| 2344 |
+
"grad_norm": 0.41597291827201843,
|
| 2345 |
+
"learning_rate": 6.417980701341491e-05,
|
| 2346 |
+
"loss": 0.0985,
|
| 2347 |
+
"step": 3340
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.7882816636272721,
|
| 2351 |
+
"grad_norm": 0.6187950372695923,
|
| 2352 |
+
"learning_rate": 6.347375853141915e-05,
|
| 2353 |
+
"loss": 0.0877,
|
| 2354 |
+
"step": 3350
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.7906347432201895,
|
| 2358 |
+
"grad_norm": 0.4807620942592621,
|
| 2359 |
+
"learning_rate": 6.276771004942338e-05,
|
| 2360 |
+
"loss": 0.1074,
|
| 2361 |
+
"step": 3360
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.7929878228131066,
|
| 2365 |
+
"grad_norm": 0.2998965382575989,
|
| 2366 |
+
"learning_rate": 6.206166156742763e-05,
|
| 2367 |
+
"loss": 0.1044,
|
| 2368 |
+
"step": 3370
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7953409024060238,
|
| 2372 |
+
"grad_norm": 0.5904129147529602,
|
| 2373 |
+
"learning_rate": 6.135561308543187e-05,
|
| 2374 |
+
"loss": 0.1073,
|
| 2375 |
+
"step": 3380
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7976939819989411,
|
| 2379 |
+
"grad_norm": 0.6356788277626038,
|
| 2380 |
+
"learning_rate": 6.06495646034361e-05,
|
| 2381 |
+
"loss": 0.1121,
|
| 2382 |
+
"step": 3390
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.8000470615918583,
|
| 2386 |
+
"grad_norm": 0.9147433638572693,
|
| 2387 |
+
"learning_rate": 5.994351612144034e-05,
|
| 2388 |
+
"loss": 0.1103,
|
| 2389 |
+
"step": 3400
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.8024001411847755,
|
| 2393 |
+
"grad_norm": 0.8032605051994324,
|
| 2394 |
+
"learning_rate": 5.923746763944457e-05,
|
| 2395 |
+
"loss": 0.0965,
|
| 2396 |
+
"step": 3410
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.8047532207776928,
|
| 2400 |
+
"grad_norm": 0.7935906052589417,
|
| 2401 |
+
"learning_rate": 5.853141915744881e-05,
|
| 2402 |
+
"loss": 0.1007,
|
| 2403 |
+
"step": 3420
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.80710630037061,
|
| 2407 |
+
"grad_norm": 0.4112412631511688,
|
| 2408 |
+
"learning_rate": 5.782537067545304e-05,
|
| 2409 |
+
"loss": 0.0882,
|
| 2410 |
+
"step": 3430
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.8094593799635272,
|
| 2414 |
+
"grad_norm": 0.8190514445304871,
|
| 2415 |
+
"learning_rate": 5.7119322193457284e-05,
|
| 2416 |
+
"loss": 0.1019,
|
| 2417 |
+
"step": 3440
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.8118124595564445,
|
| 2421 |
+
"grad_norm": 0.6029698848724365,
|
| 2422 |
+
"learning_rate": 5.641327371146152e-05,
|
| 2423 |
+
"loss": 0.1053,
|
| 2424 |
+
"step": 3450
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.8141655391493617,
|
| 2428 |
+
"grad_norm": 0.43347781896591187,
|
| 2429 |
+
"learning_rate": 5.5707225229465755e-05,
|
| 2430 |
+
"loss": 0.1044,
|
| 2431 |
+
"step": 3460
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.8165186187422789,
|
| 2435 |
+
"grad_norm": 1.5235440731048584,
|
| 2436 |
+
"learning_rate": 5.500117674746999e-05,
|
| 2437 |
+
"loss": 0.0982,
|
| 2438 |
+
"step": 3470
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.8188716983351962,
|
| 2442 |
+
"grad_norm": 0.5716174244880676,
|
| 2443 |
+
"learning_rate": 5.4295128265474225e-05,
|
| 2444 |
+
"loss": 0.1078,
|
| 2445 |
+
"step": 3480
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.8212247779281134,
|
| 2449 |
+
"grad_norm": 1.5008090734481812,
|
| 2450 |
+
"learning_rate": 5.358907978347847e-05,
|
| 2451 |
+
"loss": 0.0915,
|
| 2452 |
+
"step": 3490
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.8235778575210306,
|
| 2456 |
+
"grad_norm": 0.49782001972198486,
|
| 2457 |
+
"learning_rate": 5.28830313014827e-05,
|
| 2458 |
+
"loss": 0.0892,
|
| 2459 |
+
"step": 3500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8259309371139478,
|
| 2463 |
+
"grad_norm": 0.4466950297355652,
|
| 2464 |
+
"learning_rate": 5.217698281948694e-05,
|
| 2465 |
+
"loss": 0.0905,
|
| 2466 |
+
"step": 3510
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.8282840167068651,
|
| 2470 |
+
"grad_norm": 0.5504721403121948,
|
| 2471 |
+
"learning_rate": 5.147093433749117e-05,
|
| 2472 |
+
"loss": 0.1128,
|
| 2473 |
+
"step": 3520
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8306370962997823,
|
| 2477 |
+
"grad_norm": 0.4870951473712921,
|
| 2478 |
+
"learning_rate": 5.076488585549541e-05,
|
| 2479 |
+
"loss": 0.0876,
|
| 2480 |
+
"step": 3530
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8329901758926995,
|
| 2484 |
+
"grad_norm": 0.6789172887802124,
|
| 2485 |
+
"learning_rate": 5.005883737349965e-05,
|
| 2486 |
+
"loss": 0.1004,
|
| 2487 |
+
"step": 3540
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.8353432554856168,
|
| 2491 |
+
"grad_norm": 0.5021870136260986,
|
| 2492 |
+
"learning_rate": 4.935278889150388e-05,
|
| 2493 |
+
"loss": 0.0862,
|
| 2494 |
+
"step": 3550
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.837696335078534,
|
| 2498 |
+
"grad_norm": 0.5829181671142578,
|
| 2499 |
+
"learning_rate": 4.864674040950811e-05,
|
| 2500 |
+
"loss": 0.0994,
|
| 2501 |
+
"step": 3560
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8400494146714512,
|
| 2505 |
+
"grad_norm": 1.029181957244873,
|
| 2506 |
+
"learning_rate": 4.794069192751235e-05,
|
| 2507 |
+
"loss": 0.1144,
|
| 2508 |
+
"step": 3570
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8424024942643685,
|
| 2512 |
+
"grad_norm": 0.6730376482009888,
|
| 2513 |
+
"learning_rate": 4.723464344551658e-05,
|
| 2514 |
+
"loss": 0.106,
|
| 2515 |
+
"step": 3580
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8447555738572857,
|
| 2519 |
+
"grad_norm": 0.6129499673843384,
|
| 2520 |
+
"learning_rate": 4.652859496352082e-05,
|
| 2521 |
+
"loss": 0.0896,
|
| 2522 |
+
"step": 3590
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.8471086534502029,
|
| 2526 |
+
"grad_norm": 0.422830194234848,
|
| 2527 |
+
"learning_rate": 4.582254648152506e-05,
|
| 2528 |
+
"loss": 0.0846,
|
| 2529 |
+
"step": 3600
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.8494617330431202,
|
| 2533 |
+
"grad_norm": 0.5306664109230042,
|
| 2534 |
+
"learning_rate": 4.5116497999529296e-05,
|
| 2535 |
+
"loss": 0.1059,
|
| 2536 |
+
"step": 3610
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.8518148126360374,
|
| 2540 |
+
"grad_norm": 0.6436883807182312,
|
| 2541 |
+
"learning_rate": 4.441044951753353e-05,
|
| 2542 |
+
"loss": 0.1132,
|
| 2543 |
+
"step": 3620
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.8541678922289546,
|
| 2547 |
+
"grad_norm": 0.4121890962123871,
|
| 2548 |
+
"learning_rate": 4.3704401035537766e-05,
|
| 2549 |
+
"loss": 0.0864,
|
| 2550 |
+
"step": 3630
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8565209718218719,
|
| 2554 |
+
"grad_norm": 0.42521169781684875,
|
| 2555 |
+
"learning_rate": 4.2998352553542e-05,
|
| 2556 |
+
"loss": 0.1011,
|
| 2557 |
+
"step": 3640
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.8588740514147891,
|
| 2561 |
+
"grad_norm": 0.49623095989227295,
|
| 2562 |
+
"learning_rate": 4.229230407154624e-05,
|
| 2563 |
+
"loss": 0.1093,
|
| 2564 |
+
"step": 3650
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8612271310077063,
|
| 2568 |
+
"grad_norm": 0.5516742467880249,
|
| 2569 |
+
"learning_rate": 4.158625558955048e-05,
|
| 2570 |
+
"loss": 0.1009,
|
| 2571 |
+
"step": 3660
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.8635802106006236,
|
| 2575 |
+
"grad_norm": 0.37128451466560364,
|
| 2576 |
+
"learning_rate": 4.0880207107554713e-05,
|
| 2577 |
+
"loss": 0.0717,
|
| 2578 |
+
"step": 3670
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.8659332901935408,
|
| 2582 |
+
"grad_norm": 0.3802624046802521,
|
| 2583 |
+
"learning_rate": 4.017415862555895e-05,
|
| 2584 |
+
"loss": 0.0891,
|
| 2585 |
+
"step": 3680
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.868286369786458,
|
| 2589 |
+
"grad_norm": 0.35558944940567017,
|
| 2590 |
+
"learning_rate": 3.9468110143563184e-05,
|
| 2591 |
+
"loss": 0.0863,
|
| 2592 |
+
"step": 3690
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8706394493793752,
|
| 2596 |
+
"grad_norm": 0.2548139989376068,
|
| 2597 |
+
"learning_rate": 3.8762061661567426e-05,
|
| 2598 |
+
"loss": 0.0946,
|
| 2599 |
+
"step": 3700
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8729925289722925,
|
| 2603 |
+
"grad_norm": 0.3489900827407837,
|
| 2604 |
+
"learning_rate": 3.805601317957166e-05,
|
| 2605 |
+
"loss": 0.0794,
|
| 2606 |
+
"step": 3710
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.8753456085652097,
|
| 2610 |
+
"grad_norm": 0.7514833807945251,
|
| 2611 |
+
"learning_rate": 3.7349964697575896e-05,
|
| 2612 |
+
"loss": 0.1026,
|
| 2613 |
+
"step": 3720
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.8776986881581269,
|
| 2617 |
+
"grad_norm": 0.28846803307533264,
|
| 2618 |
+
"learning_rate": 3.664391621558013e-05,
|
| 2619 |
+
"loss": 0.107,
|
| 2620 |
+
"step": 3730
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8800517677510442,
|
| 2624 |
+
"grad_norm": 0.3054257333278656,
|
| 2625 |
+
"learning_rate": 3.5937867733584366e-05,
|
| 2626 |
+
"loss": 0.0839,
|
| 2627 |
+
"step": 3740
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.8824048473439614,
|
| 2631 |
+
"grad_norm": 0.487393856048584,
|
| 2632 |
+
"learning_rate": 3.523181925158861e-05,
|
| 2633 |
+
"loss": 0.099,
|
| 2634 |
+
"step": 3750
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8847579269368786,
|
| 2638 |
+
"grad_norm": 0.7874276041984558,
|
| 2639 |
+
"learning_rate": 3.4525770769592843e-05,
|
| 2640 |
+
"loss": 0.0873,
|
| 2641 |
+
"step": 3760
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8871110065297959,
|
| 2645 |
+
"grad_norm": 0.3583498001098633,
|
| 2646 |
+
"learning_rate": 3.381972228759708e-05,
|
| 2647 |
+
"loss": 0.0854,
|
| 2648 |
+
"step": 3770
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.8894640861227131,
|
| 2652 |
+
"grad_norm": 0.5606823563575745,
|
| 2653 |
+
"learning_rate": 3.3113673805601314e-05,
|
| 2654 |
+
"loss": 0.1106,
|
| 2655 |
+
"step": 3780
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8918171657156303,
|
| 2659 |
+
"grad_norm": 0.48208296298980713,
|
| 2660 |
+
"learning_rate": 3.240762532360555e-05,
|
| 2661 |
+
"loss": 0.1138,
|
| 2662 |
+
"step": 3790
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8941702453085476,
|
| 2666 |
+
"grad_norm": 1.026995301246643,
|
| 2667 |
+
"learning_rate": 3.170157684160979e-05,
|
| 2668 |
+
"loss": 0.0877,
|
| 2669 |
+
"step": 3800
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8965233249014648,
|
| 2673 |
+
"grad_norm": 0.7940952777862549,
|
| 2674 |
+
"learning_rate": 3.0995528359614026e-05,
|
| 2675 |
+
"loss": 0.069,
|
| 2676 |
+
"step": 3810
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.898876404494382,
|
| 2680 |
+
"grad_norm": 0.7711090445518494,
|
| 2681 |
+
"learning_rate": 3.028947987761826e-05,
|
| 2682 |
+
"loss": 0.0884,
|
| 2683 |
+
"step": 3820
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.9012294840872993,
|
| 2687 |
+
"grad_norm": 0.6985650062561035,
|
| 2688 |
+
"learning_rate": 2.9583431395622496e-05,
|
| 2689 |
+
"loss": 0.0929,
|
| 2690 |
+
"step": 3830
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.9035825636802165,
|
| 2694 |
+
"grad_norm": 0.5291894674301147,
|
| 2695 |
+
"learning_rate": 2.8877382913626735e-05,
|
| 2696 |
+
"loss": 0.1166,
|
| 2697 |
+
"step": 3840
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.9059356432731337,
|
| 2701 |
+
"grad_norm": 0.3929837644100189,
|
| 2702 |
+
"learning_rate": 2.817133443163097e-05,
|
| 2703 |
+
"loss": 0.0904,
|
| 2704 |
+
"step": 3850
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.9082887228660509,
|
| 2708 |
+
"grad_norm": 0.492017537355423,
|
| 2709 |
+
"learning_rate": 2.746528594963521e-05,
|
| 2710 |
+
"loss": 0.0986,
|
| 2711 |
+
"step": 3860
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.9106418024589682,
|
| 2715 |
+
"grad_norm": 0.5756918787956238,
|
| 2716 |
+
"learning_rate": 2.6759237467639444e-05,
|
| 2717 |
+
"loss": 0.0897,
|
| 2718 |
+
"step": 3870
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.9129948820518854,
|
| 2722 |
+
"grad_norm": 0.5781024098396301,
|
| 2723 |
+
"learning_rate": 2.605318898564368e-05,
|
| 2724 |
+
"loss": 0.0799,
|
| 2725 |
+
"step": 3880
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.9153479616448026,
|
| 2729 |
+
"grad_norm": 0.28270334005355835,
|
| 2730 |
+
"learning_rate": 2.5347140503647918e-05,
|
| 2731 |
+
"loss": 0.0889,
|
| 2732 |
+
"step": 3890
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.9177010412377199,
|
| 2736 |
+
"grad_norm": 0.5788043737411499,
|
| 2737 |
+
"learning_rate": 2.464109202165215e-05,
|
| 2738 |
+
"loss": 0.0788,
|
| 2739 |
+
"step": 3900
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.9200541208306371,
|
| 2743 |
+
"grad_norm": 0.563836932182312,
|
| 2744 |
+
"learning_rate": 2.3935043539656384e-05,
|
| 2745 |
+
"loss": 0.0914,
|
| 2746 |
+
"step": 3910
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.9224072004235543,
|
| 2750 |
+
"grad_norm": 0.4077290892601013,
|
| 2751 |
+
"learning_rate": 2.3228995057660623e-05,
|
| 2752 |
+
"loss": 0.1057,
|
| 2753 |
+
"step": 3920
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.9247602800164716,
|
| 2757 |
+
"grad_norm": 0.6209468841552734,
|
| 2758 |
+
"learning_rate": 2.2522946575664858e-05,
|
| 2759 |
+
"loss": 0.0812,
|
| 2760 |
+
"step": 3930
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.9271133596093888,
|
| 2764 |
+
"grad_norm": 0.542506754398346,
|
| 2765 |
+
"learning_rate": 2.1816898093669097e-05,
|
| 2766 |
+
"loss": 0.0951,
|
| 2767 |
+
"step": 3940
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.929466439202306,
|
| 2771 |
+
"grad_norm": 0.5754973292350769,
|
| 2772 |
+
"learning_rate": 2.1110849611673332e-05,
|
| 2773 |
+
"loss": 0.1023,
|
| 2774 |
+
"step": 3950
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.9318195187952233,
|
| 2778 |
+
"grad_norm": 0.3798030912876129,
|
| 2779 |
+
"learning_rate": 2.0404801129677567e-05,
|
| 2780 |
+
"loss": 0.0997,
|
| 2781 |
+
"step": 3960
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.9341725983881405,
|
| 2785 |
+
"grad_norm": 0.6593634486198425,
|
| 2786 |
+
"learning_rate": 1.9698752647681806e-05,
|
| 2787 |
+
"loss": 0.1056,
|
| 2788 |
+
"step": 3970
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9365256779810577,
|
| 2792 |
+
"grad_norm": 0.46481505036354065,
|
| 2793 |
+
"learning_rate": 1.899270416568604e-05,
|
| 2794 |
+
"loss": 0.0795,
|
| 2795 |
+
"step": 3980
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.938878757573975,
|
| 2799 |
+
"grad_norm": 0.5140686631202698,
|
| 2800 |
+
"learning_rate": 1.828665568369028e-05,
|
| 2801 |
+
"loss": 0.0809,
|
| 2802 |
+
"step": 3990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9412318371668922,
|
| 2806 |
+
"grad_norm": 0.8201892375946045,
|
| 2807 |
+
"learning_rate": 1.7580607201694515e-05,
|
| 2808 |
+
"loss": 0.0851,
|
| 2809 |
+
"step": 4000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.9435849167598094,
|
| 2813 |
+
"grad_norm": 0.3848946690559387,
|
| 2814 |
+
"learning_rate": 1.687455871969875e-05,
|
| 2815 |
+
"loss": 0.0789,
|
| 2816 |
+
"step": 4010
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9459379963527267,
|
| 2820 |
+
"grad_norm": 0.3362932503223419,
|
| 2821 |
+
"learning_rate": 1.6168510237702988e-05,
|
| 2822 |
+
"loss": 0.0841,
|
| 2823 |
+
"step": 4020
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9482910759456439,
|
| 2827 |
+
"grad_norm": 0.400037556886673,
|
| 2828 |
+
"learning_rate": 1.5462461755707223e-05,
|
| 2829 |
+
"loss": 0.1004,
|
| 2830 |
+
"step": 4030
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.9506441555385611,
|
| 2834 |
+
"grad_norm": 0.6505069136619568,
|
| 2835 |
+
"learning_rate": 1.475641327371146e-05,
|
| 2836 |
+
"loss": 0.0977,
|
| 2837 |
+
"step": 4040
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.9529972351314783,
|
| 2841 |
+
"grad_norm": 0.710784375667572,
|
| 2842 |
+
"learning_rate": 1.4050364791715697e-05,
|
| 2843 |
+
"loss": 0.085,
|
| 2844 |
+
"step": 4050
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9553503147243956,
|
| 2848 |
+
"grad_norm": 0.4263714849948883,
|
| 2849 |
+
"learning_rate": 1.3344316309719934e-05,
|
| 2850 |
+
"loss": 0.1044,
|
| 2851 |
+
"step": 4060
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.9577033943173128,
|
| 2855 |
+
"grad_norm": 0.42400240898132324,
|
| 2856 |
+
"learning_rate": 1.2638267827724171e-05,
|
| 2857 |
+
"loss": 0.1113,
|
| 2858 |
+
"step": 4070
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.96005647391023,
|
| 2862 |
+
"grad_norm": 0.2722209393978119,
|
| 2863 |
+
"learning_rate": 1.1932219345728404e-05,
|
| 2864 |
+
"loss": 0.0792,
|
| 2865 |
+
"step": 4080
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.9624095535031473,
|
| 2869 |
+
"grad_norm": 0.9779515862464905,
|
| 2870 |
+
"learning_rate": 1.1226170863732641e-05,
|
| 2871 |
+
"loss": 0.1044,
|
| 2872 |
+
"step": 4090
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.9647626330960645,
|
| 2876 |
+
"grad_norm": 1.028387188911438,
|
| 2877 |
+
"learning_rate": 1.0520122381736878e-05,
|
| 2878 |
+
"loss": 0.1043,
|
| 2879 |
+
"step": 4100
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.9671157126889817,
|
| 2883 |
+
"grad_norm": 0.5009176135063171,
|
| 2884 |
+
"learning_rate": 9.814073899741115e-06,
|
| 2885 |
+
"loss": 0.087,
|
| 2886 |
+
"step": 4110
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.969468792281899,
|
| 2890 |
+
"grad_norm": 0.33020302653312683,
|
| 2891 |
+
"learning_rate": 9.10802541774535e-06,
|
| 2892 |
+
"loss": 0.092,
|
| 2893 |
+
"step": 4120
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.9718218718748162,
|
| 2897 |
+
"grad_norm": 0.4314991533756256,
|
| 2898 |
+
"learning_rate": 8.401976935749587e-06,
|
| 2899 |
+
"loss": 0.075,
|
| 2900 |
+
"step": 4130
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9741749514677334,
|
| 2904 |
+
"grad_norm": 0.6121822595596313,
|
| 2905 |
+
"learning_rate": 7.695928453753824e-06,
|
| 2906 |
+
"loss": 0.0892,
|
| 2907 |
+
"step": 4140
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.9765280310606507,
|
| 2911 |
+
"grad_norm": 0.3374115824699402,
|
| 2912 |
+
"learning_rate": 6.989879971758061e-06,
|
| 2913 |
+
"loss": 0.0755,
|
| 2914 |
+
"step": 4150
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.9788811106535679,
|
| 2918 |
+
"grad_norm": 0.5865825414657593,
|
| 2919 |
+
"learning_rate": 6.283831489762297e-06,
|
| 2920 |
+
"loss": 0.0957,
|
| 2921 |
+
"step": 4160
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.9812341902464851,
|
| 2925 |
+
"grad_norm": 0.2131696194410324,
|
| 2926 |
+
"learning_rate": 5.577783007766533e-06,
|
| 2927 |
+
"loss": 0.0849,
|
| 2928 |
+
"step": 4170
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.9835872698394024,
|
| 2932 |
+
"grad_norm": 1.3489303588867188,
|
| 2933 |
+
"learning_rate": 4.871734525770769e-06,
|
| 2934 |
+
"loss": 0.098,
|
| 2935 |
+
"step": 4180
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9859403494323196,
|
| 2939 |
+
"grad_norm": 0.15470068156719208,
|
| 2940 |
+
"learning_rate": 4.1656860437750056e-06,
|
| 2941 |
+
"loss": 0.0847,
|
| 2942 |
+
"step": 4190
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9882934290252368,
|
| 2946 |
+
"grad_norm": 0.8059414625167847,
|
| 2947 |
+
"learning_rate": 3.459637561779242e-06,
|
| 2948 |
+
"loss": 0.1094,
|
| 2949 |
+
"step": 4200
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.990646508618154,
|
| 2953 |
+
"grad_norm": 0.6808902621269226,
|
| 2954 |
+
"learning_rate": 2.753589079783478e-06,
|
| 2955 |
+
"loss": 0.0867,
|
| 2956 |
+
"step": 4210
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9929995882110713,
|
| 2960 |
+
"grad_norm": 0.29802441596984863,
|
| 2961 |
+
"learning_rate": 2.0475405977877145e-06,
|
| 2962 |
+
"loss": 0.0866,
|
| 2963 |
+
"step": 4220
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.9953526678039885,
|
| 2967 |
+
"grad_norm": 0.5227815508842468,
|
| 2968 |
+
"learning_rate": 1.341492115791951e-06,
|
| 2969 |
+
"loss": 0.0758,
|
| 2970 |
+
"step": 4230
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9977057473969057,
|
| 2974 |
+
"grad_norm": 0.25185248255729675,
|
| 2975 |
+
"learning_rate": 6.354436337961872e-07,
|
| 2976 |
+
"loss": 0.0906,
|
| 2977 |
+
"step": 4240
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9998235190305312,
|
| 2981 |
+
"step": 4249,
|
| 2982 |
+
"total_flos": 1.037665224400896e+16,
|
| 2983 |
+
"train_loss": 0.1519955188758403,
|
| 2984 |
+
"train_runtime": 5739.131,
|
| 2985 |
+
"train_samples_per_second": 11.848,
|
| 2986 |
+
"train_steps_per_second": 0.74
|
| 2987 |
+
}
|
| 2988 |
+
],
|
| 2989 |
+
"logging_steps": 10,
|
| 2990 |
+
"max_steps": 4249,
|
| 2991 |
+
"num_input_tokens_seen": 0,
|
| 2992 |
+
"num_train_epochs": 1,
|
| 2993 |
+
"save_steps": 500,
|
| 2994 |
+
"total_flos": 1.037665224400896e+16,
|
| 2995 |
+
"train_batch_size": 2,
|
| 2996 |
+
"trial_name": null,
|
| 2997 |
+
"trial_params": null
|
| 2998 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33d5b83a931432fb668a7936764cbe9eea86b9284ac9f7972603743596f1a111
|
| 3 |
+
size 5176
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|