End of training
Browse files- README.md +11 -11
- all_results.json +5 -5
- config.json +62 -22
- model.safetensors +2 -2
- preprocessor_config.json +2 -2
- runs/Apr14_05-42-06_5f7eb6924bd1/events.out.tfevents.1713073328.5f7eb6924bd1.34.0 +3 -0
- runs/Apr14_05-42-06_5f7eb6924bd1/events.out.tfevents.1713078581.5f7eb6924bd1.34.1 +3 -0
- train_results.json +5 -5
- trainer_state.json +165 -142
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
-
base_model:
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
metrics:
|
|
@@ -15,10 +15,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# ckpts
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- Accuracy: 0.
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
@@ -52,13 +52,13 @@ The following hyperparameters were used during training:
|
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 54 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.
|
| 58 |
-
| 0.
|
| 59 |
-
| 0.
|
| 60 |
-
| 0.
|
| 61 |
-
| 0.
|
| 62 |
|
| 63 |
|
| 64 |
### Framework versions
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
base_model: ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
metrics:
|
|
|
|
| 15 |
|
| 16 |
# ckpts
|
| 17 |
|
| 18 |
+
This model is a fine-tuned version of [ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition](https://huggingface.co/ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.0873
|
| 21 |
+
- Accuracy: 0.9758
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 54 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 55 |
+
| 0.2538 | 2.02 | 500 | 0.2764 | 0.9071 |
|
| 56 |
+
| 0.0805 | 4.04 | 1000 | 0.1722 | 0.9535 |
|
| 57 |
+
| 0.0505 | 6.06 | 1500 | 0.1142 | 0.9778 |
|
| 58 |
+
| 0.0466 | 8.08 | 2000 | 0.1177 | 0.9737 |
|
| 59 |
+
| 0.0269 | 10.1 | 2500 | 0.1183 | 0.9737 |
|
| 60 |
+
| 0.025 | 12.12 | 3000 | 0.0873 | 0.9758 |
|
| 61 |
+
| 0.0323 | 14.14 | 3500 | 0.0920 | 0.9798 |
|
| 62 |
|
| 63 |
|
| 64 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 14.97,
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 1979,
|
| 7 |
-
"train_samples_per_second":
|
| 8 |
-
"train_steps_per_second":
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 14.97,
|
| 3 |
+
"total_flos": 3.163398064220592e+18,
|
| 4 |
+
"train_loss": 0.12699281216954733,
|
| 5 |
+
"train_runtime": 5204.2292,
|
| 6 |
"train_samples": 1979,
|
| 7 |
+
"train_samples_per_second": 5.704,
|
| 8 |
+
"train_steps_per_second": 0.712
|
| 9 |
}
|
config.json
CHANGED
|
@@ -1,14 +1,20 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
-
"activation_dropout": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
"apply_spec_augment": true,
|
| 5 |
"architectures": [
|
| 6 |
-
"
|
| 7 |
],
|
| 8 |
"attention_dropout": 0.1,
|
| 9 |
"bos_token_id": 1,
|
| 10 |
"classifier_proj_size": 256,
|
| 11 |
-
"
|
|
|
|
|
|
|
| 12 |
"conv_dim": [
|
| 13 |
512,
|
| 14 |
512,
|
|
@@ -36,22 +42,21 @@
|
|
| 36 |
2,
|
| 37 |
2
|
| 38 |
],
|
| 39 |
-
"ctc_loss_reduction": "
|
| 40 |
-
"ctc_zero_infinity":
|
| 41 |
-
"
|
|
|
|
| 42 |
"eos_token_id": 2,
|
| 43 |
"feat_extract_activation": "gelu",
|
| 44 |
"feat_extract_dropout": 0.0,
|
| 45 |
-
"feat_extract_norm": "
|
| 46 |
-
"feat_proj_dropout": 0.
|
| 47 |
-
"
|
| 48 |
-
"final_dropout": 0.
|
| 49 |
"finetuning_task": "wav2vec2_clf",
|
| 50 |
-
"gradient_checkpointing": false,
|
| 51 |
"hidden_act": "gelu",
|
| 52 |
-
"hidden_dropout": 0.
|
| 53 |
-
"
|
| 54 |
-
"hidden_size": 768,
|
| 55 |
"id2label": {
|
| 56 |
"0": "01Neutral",
|
| 57 |
"1": "02Fearful",
|
|
@@ -60,7 +65,7 @@
|
|
| 60 |
"4": "05Angry"
|
| 61 |
},
|
| 62 |
"initializer_range": 0.02,
|
| 63 |
-
"intermediate_size":
|
| 64 |
"label2id": {
|
| 65 |
"01Neutral": 0,
|
| 66 |
"02Fearful": 1,
|
|
@@ -69,25 +74,60 @@
|
|
| 69 |
"05Angry": 4
|
| 70 |
},
|
| 71 |
"layer_norm_eps": 1e-05,
|
| 72 |
-
"layerdrop": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
"mask_feature_length": 10,
|
| 74 |
"mask_feature_min_masks": 0,
|
| 75 |
"mask_feature_prob": 0.0,
|
| 76 |
"mask_time_length": 10,
|
| 77 |
"mask_time_min_masks": 2,
|
|
|
|
|
|
|
| 78 |
"mask_time_prob": 0.05,
|
| 79 |
-
"
|
| 80 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
| 81 |
"num_conv_pos_embedding_groups": 16,
|
| 82 |
"num_conv_pos_embeddings": 128,
|
| 83 |
"num_feat_extract_layers": 7,
|
| 84 |
-
"num_hidden_layers":
|
|
|
|
|
|
|
| 85 |
"pad_token_id": 0,
|
| 86 |
"pooling_mode": "mean",
|
| 87 |
"problem_type": "single_label_classification",
|
| 88 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 89 |
"torch_dtype": "float32",
|
| 90 |
"transformers_version": "4.40.0.dev0",
|
| 91 |
"use_weighted_layer_sum": false,
|
| 92 |
-
"vocab_size":
|
|
|
|
| 93 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition",
|
| 3 |
+
"activation_dropout": 0.05,
|
| 4 |
+
"adapter_attn_dim": null,
|
| 5 |
+
"adapter_kernel_size": 3,
|
| 6 |
+
"adapter_stride": 2,
|
| 7 |
+
"add_adapter": false,
|
| 8 |
"apply_spec_augment": true,
|
| 9 |
"architectures": [
|
| 10 |
+
"Wav2Vec2ForSpeechClassification"
|
| 11 |
],
|
| 12 |
"attention_dropout": 0.1,
|
| 13 |
"bos_token_id": 1,
|
| 14 |
"classifier_proj_size": 256,
|
| 15 |
+
"codevector_dim": 256,
|
| 16 |
+
"contrastive_logits_temperature": 0.1,
|
| 17 |
+
"conv_bias": true,
|
| 18 |
"conv_dim": [
|
| 19 |
512,
|
| 20 |
512,
|
|
|
|
| 42 |
2,
|
| 43 |
2
|
| 44 |
],
|
| 45 |
+
"ctc_loss_reduction": "mean",
|
| 46 |
+
"ctc_zero_infinity": true,
|
| 47 |
+
"diversity_loss_weight": 0.1,
|
| 48 |
+
"do_stable_layer_norm": true,
|
| 49 |
"eos_token_id": 2,
|
| 50 |
"feat_extract_activation": "gelu",
|
| 51 |
"feat_extract_dropout": 0.0,
|
| 52 |
+
"feat_extract_norm": "layer",
|
| 53 |
+
"feat_proj_dropout": 0.05,
|
| 54 |
+
"feat_quantizer_dropout": 0.0,
|
| 55 |
+
"final_dropout": 0.0,
|
| 56 |
"finetuning_task": "wav2vec2_clf",
|
|
|
|
| 57 |
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.05,
|
| 59 |
+
"hidden_size": 1024,
|
|
|
|
| 60 |
"id2label": {
|
| 61 |
"0": "01Neutral",
|
| 62 |
"1": "02Fearful",
|
|
|
|
| 65 |
"4": "05Angry"
|
| 66 |
},
|
| 67 |
"initializer_range": 0.02,
|
| 68 |
+
"intermediate_size": 4096,
|
| 69 |
"label2id": {
|
| 70 |
"01Neutral": 0,
|
| 71 |
"02Fearful": 1,
|
|
|
|
| 74 |
"05Angry": 4
|
| 75 |
},
|
| 76 |
"layer_norm_eps": 1e-05,
|
| 77 |
+
"layerdrop": 0.05,
|
| 78 |
+
"mask_channel_length": 10,
|
| 79 |
+
"mask_channel_min_space": 1,
|
| 80 |
+
"mask_channel_other": 0.0,
|
| 81 |
+
"mask_channel_prob": 0.0,
|
| 82 |
+
"mask_channel_selection": "static",
|
| 83 |
"mask_feature_length": 10,
|
| 84 |
"mask_feature_min_masks": 0,
|
| 85 |
"mask_feature_prob": 0.0,
|
| 86 |
"mask_time_length": 10,
|
| 87 |
"mask_time_min_masks": 2,
|
| 88 |
+
"mask_time_min_space": 1,
|
| 89 |
+
"mask_time_other": 0.0,
|
| 90 |
"mask_time_prob": 0.05,
|
| 91 |
+
"mask_time_selection": "static",
|
| 92 |
+
"model_type": "wav2vec2",
|
| 93 |
+
"num_adapter_layers": 3,
|
| 94 |
+
"num_attention_heads": 16,
|
| 95 |
+
"num_codevector_groups": 2,
|
| 96 |
+
"num_codevectors_per_group": 320,
|
| 97 |
"num_conv_pos_embedding_groups": 16,
|
| 98 |
"num_conv_pos_embeddings": 128,
|
| 99 |
"num_feat_extract_layers": 7,
|
| 100 |
+
"num_hidden_layers": 24,
|
| 101 |
+
"num_negatives": 100,
|
| 102 |
+
"output_hidden_size": 1024,
|
| 103 |
"pad_token_id": 0,
|
| 104 |
"pooling_mode": "mean",
|
| 105 |
"problem_type": "single_label_classification",
|
| 106 |
+
"proj_codevector_dim": 256,
|
| 107 |
+
"tdnn_dilation": [
|
| 108 |
+
1,
|
| 109 |
+
2,
|
| 110 |
+
3,
|
| 111 |
+
1,
|
| 112 |
+
1
|
| 113 |
+
],
|
| 114 |
+
"tdnn_dim": [
|
| 115 |
+
512,
|
| 116 |
+
512,
|
| 117 |
+
512,
|
| 118 |
+
512,
|
| 119 |
+
1500
|
| 120 |
+
],
|
| 121 |
+
"tdnn_kernel": [
|
| 122 |
+
5,
|
| 123 |
+
3,
|
| 124 |
+
3,
|
| 125 |
+
1,
|
| 126 |
+
1
|
| 127 |
+
],
|
| 128 |
"torch_dtype": "float32",
|
| 129 |
"transformers_version": "4.40.0.dev0",
|
| 130 |
"use_weighted_layer_sum": false,
|
| 131 |
+
"vocab_size": 33,
|
| 132 |
+
"xvector_output_dim": 512
|
| 133 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36ed8619730ada58c91dc801f90d035693a43655ce5856d2dc3968d3fc2e881d
|
| 3 |
+
size 1266026604
|
preprocessor_config.json
CHANGED
|
@@ -3,7 +3,7 @@
|
|
| 3 |
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
"feature_size": 1,
|
| 5 |
"padding_side": "right",
|
| 6 |
-
"padding_value": 0,
|
| 7 |
-
"return_attention_mask":
|
| 8 |
"sampling_rate": 16000
|
| 9 |
}
|
|
|
|
| 3 |
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
"feature_size": 1,
|
| 5 |
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": true,
|
| 8 |
"sampling_rate": 16000
|
| 9 |
}
|
runs/Apr14_05-42-06_5f7eb6924bd1/events.out.tfevents.1713073328.5f7eb6924bd1.34.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12e39d666ac5bc0cddf9e4ebc91ed7cc4772c19b168f88dda7d0c4f6bcc4e681
|
| 3 |
+
size 17107
|
runs/Apr14_05-42-06_5f7eb6924bd1/events.out.tfevents.1713078581.5f7eb6924bd1.34.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:790cf68ce3de91187bf42a1f079402bfdf1d532a431e773fcad00ae461bc91e5
|
| 3 |
+
size 411
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 14.97,
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 1979,
|
| 7 |
-
"train_samples_per_second":
|
| 8 |
-
"train_steps_per_second":
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 14.97,
|
| 3 |
+
"total_flos": 3.163398064220592e+18,
|
| 4 |
+
"train_loss": 0.12699281216954733,
|
| 5 |
+
"train_runtime": 5204.2292,
|
| 6 |
"train_samples": 1979,
|
| 7 |
+
"train_samples_per_second": 5.704,
|
| 8 |
+
"train_steps_per_second": 0.712
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -1,320 +1,343 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "/kaggle/working/ckpts/checkpoint-
|
| 4 |
-
"epoch": 14.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.4,
|
| 13 |
-
"grad_norm":
|
| 14 |
"learning_rate": 9.730094466936572e-06,
|
| 15 |
-
"loss": 1.
|
| 16 |
"step": 100
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.81,
|
| 20 |
-
"grad_norm":
|
| 21 |
"learning_rate": 9.460188933873145e-06,
|
| 22 |
-
"loss":
|
| 23 |
"step": 200
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 1.21,
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate": 9.
|
| 29 |
-
"loss": 0.
|
| 30 |
"step": 300
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 1.62,
|
| 34 |
-
"grad_norm":
|
| 35 |
-
"learning_rate": 8.
|
| 36 |
-
"loss": 0.
|
| 37 |
"step": 400
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 2.02,
|
| 41 |
-
"grad_norm":
|
| 42 |
-
"learning_rate": 8.
|
| 43 |
-
"loss": 0.
|
| 44 |
"step": 500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 2.02,
|
| 48 |
-
"eval_accuracy": 0.
|
| 49 |
-
"eval_loss": 0.
|
| 50 |
-
"eval_runtime":
|
| 51 |
-
"eval_samples_per_second":
|
| 52 |
-
"eval_steps_per_second":
|
| 53 |
"step": 500
|
| 54 |
},
|
| 55 |
{
|
| 56 |
"epoch": 2.42,
|
| 57 |
-
"grad_norm":
|
| 58 |
"learning_rate": 8.385964912280704e-06,
|
| 59 |
-
"loss": 0.
|
| 60 |
"step": 600
|
| 61 |
},
|
| 62 |
{
|
| 63 |
"epoch": 2.83,
|
| 64 |
-
"grad_norm":
|
| 65 |
-
"learning_rate": 8.
|
| 66 |
-
"loss": 0.
|
| 67 |
"step": 700
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 3.23,
|
| 71 |
-
"grad_norm":
|
| 72 |
-
"learning_rate": 7.
|
| 73 |
-
"loss": 0.
|
| 74 |
"step": 800
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 3.64,
|
| 78 |
-
"grad_norm":
|
| 79 |
-
"learning_rate": 7.
|
| 80 |
-
"loss": 0.
|
| 81 |
"step": 900
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 4.04,
|
| 85 |
-
"grad_norm": 0.
|
| 86 |
-
"learning_rate": 7.
|
| 87 |
-
"loss": 0.
|
| 88 |
"step": 1000
|
| 89 |
},
|
| 90 |
{
|
| 91 |
"epoch": 4.04,
|
| 92 |
-
"eval_accuracy": 0.
|
| 93 |
-
"eval_loss": 0.
|
| 94 |
-
"eval_runtime":
|
| 95 |
-
"eval_samples_per_second":
|
| 96 |
-
"eval_steps_per_second":
|
| 97 |
"step": 1000
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 4.44,
|
| 101 |
-
"grad_norm":
|
| 102 |
-
"learning_rate": 7.
|
| 103 |
-
"loss": 0.
|
| 104 |
"step": 1100
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 4.85,
|
| 108 |
-
"grad_norm":
|
| 109 |
-
"learning_rate": 6.
|
| 110 |
-
"loss": 0.
|
| 111 |
"step": 1200
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 5.25,
|
| 115 |
-
"grad_norm":
|
| 116 |
-
"learning_rate": 6.
|
| 117 |
-
"loss": 0.
|
| 118 |
"step": 1300
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 5.66,
|
| 122 |
-
"grad_norm":
|
| 123 |
-
"learning_rate": 6.
|
| 124 |
-
"loss": 0.
|
| 125 |
"step": 1400
|
| 126 |
},
|
| 127 |
{
|
| 128 |
"epoch": 6.06,
|
| 129 |
-
"grad_norm":
|
| 130 |
-
"learning_rate": 5.
|
| 131 |
-
"loss": 0.
|
| 132 |
"step": 1500
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 6.06,
|
| 136 |
-
"eval_accuracy": 0.
|
| 137 |
-
"eval_loss": 0.
|
| 138 |
-
"eval_runtime":
|
| 139 |
-
"eval_samples_per_second":
|
| 140 |
-
"eval_steps_per_second":
|
| 141 |
"step": 1500
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 6.46,
|
| 145 |
-
"grad_norm":
|
| 146 |
-
"learning_rate": 5.
|
| 147 |
-
"loss": 0.
|
| 148 |
"step": 1600
|
| 149 |
},
|
| 150 |
{
|
| 151 |
"epoch": 6.87,
|
| 152 |
-
"grad_norm": 0.
|
| 153 |
-
"learning_rate": 5.
|
| 154 |
-
"loss": 0.
|
| 155 |
"step": 1700
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 7.27,
|
| 159 |
-
"grad_norm": 2.
|
| 160 |
-
"learning_rate": 5.
|
| 161 |
-
"loss": 0.
|
| 162 |
"step": 1800
|
| 163 |
},
|
| 164 |
{
|
| 165 |
"epoch": 7.68,
|
| 166 |
-
"grad_norm": 0.
|
| 167 |
-
"learning_rate": 4.
|
| 168 |
-
"loss": 0.
|
| 169 |
"step": 1900
|
| 170 |
},
|
| 171 |
{
|
| 172 |
"epoch": 8.08,
|
| 173 |
-
"grad_norm":
|
| 174 |
-
"learning_rate": 4.
|
| 175 |
-
"loss": 0.
|
| 176 |
"step": 2000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 8.08,
|
| 180 |
-
"eval_accuracy": 0.
|
| 181 |
-
"eval_loss": 0.
|
| 182 |
-
"eval_runtime":
|
| 183 |
-
"eval_samples_per_second":
|
| 184 |
-
"eval_steps_per_second":
|
| 185 |
"step": 2000
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 8.48,
|
| 189 |
-
"grad_norm":
|
| 190 |
-
"learning_rate": 4.
|
| 191 |
-
"loss": 0.
|
| 192 |
"step": 2100
|
| 193 |
},
|
| 194 |
{
|
| 195 |
"epoch": 8.89,
|
| 196 |
-
"grad_norm":
|
| 197 |
-
"learning_rate": 4.
|
| 198 |
-
"loss": 0.
|
| 199 |
"step": 2200
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"epoch": 9.29,
|
| 203 |
-
"grad_norm":
|
| 204 |
-
"learning_rate": 3.
|
| 205 |
-
"loss": 0.
|
| 206 |
"step": 2300
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 9.7,
|
| 210 |
-
"grad_norm": 0.
|
| 211 |
-
"learning_rate": 3.
|
| 212 |
-
"loss": 0.
|
| 213 |
"step": 2400
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 10.1,
|
| 217 |
-
"grad_norm": 0.
|
| 218 |
-
"learning_rate": 3.
|
| 219 |
-
"loss": 0.
|
| 220 |
"step": 2500
|
| 221 |
},
|
| 222 |
{
|
| 223 |
"epoch": 10.1,
|
| 224 |
-
"eval_accuracy": 0.
|
| 225 |
-
"eval_loss": 0.
|
| 226 |
-
"eval_runtime":
|
| 227 |
-
"eval_samples_per_second":
|
| 228 |
-
"eval_steps_per_second":
|
| 229 |
"step": 2500
|
| 230 |
},
|
| 231 |
{
|
| 232 |
"epoch": 10.51,
|
| 233 |
-
"grad_norm": 0.
|
| 234 |
-
"learning_rate": 2.
|
| 235 |
-
"loss": 0.
|
| 236 |
"step": 2600
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 10.91,
|
| 240 |
-
"grad_norm": 0.
|
| 241 |
-
"learning_rate": 2.
|
| 242 |
-
"loss": 0.
|
| 243 |
"step": 2700
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 11.31,
|
| 247 |
-
"grad_norm":
|
| 248 |
-
"learning_rate": 2.
|
| 249 |
-
"loss": 0.
|
| 250 |
"step": 2800
|
| 251 |
},
|
| 252 |
{
|
| 253 |
"epoch": 11.72,
|
| 254 |
-
"grad_norm":
|
| 255 |
-
"learning_rate": 2.
|
| 256 |
-
"loss": 0.
|
| 257 |
"step": 2900
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 12.12,
|
| 261 |
-
"grad_norm": 0.
|
| 262 |
-
"learning_rate": 1.
|
| 263 |
-
"loss": 0.
|
| 264 |
"step": 3000
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 12.12,
|
| 268 |
-
"eval_accuracy": 0.
|
| 269 |
-
"eval_loss": 0.
|
| 270 |
-
"eval_runtime":
|
| 271 |
-
"eval_samples_per_second":
|
| 272 |
-
"eval_steps_per_second":
|
| 273 |
"step": 3000
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 12.53,
|
| 277 |
-
"grad_norm": 0.
|
| 278 |
-
"learning_rate": 1.
|
| 279 |
-
"loss": 0.
|
| 280 |
"step": 3100
|
| 281 |
},
|
| 282 |
{
|
| 283 |
"epoch": 12.93,
|
| 284 |
-
"grad_norm":
|
| 285 |
-
"learning_rate": 1.
|
| 286 |
-
"loss": 0.
|
| 287 |
"step": 3200
|
| 288 |
},
|
| 289 |
{
|
| 290 |
"epoch": 13.33,
|
| 291 |
-
"grad_norm": 0.
|
| 292 |
-
"learning_rate": 1.
|
| 293 |
-
"loss": 0.
|
| 294 |
"step": 3300
|
| 295 |
},
|
| 296 |
{
|
| 297 |
"epoch": 13.74,
|
| 298 |
-
"grad_norm": 0.
|
| 299 |
-
"learning_rate": 8.
|
| 300 |
-
"loss": 0.
|
| 301 |
"step": 3400
|
| 302 |
},
|
| 303 |
{
|
| 304 |
"epoch": 14.14,
|
| 305 |
-
"grad_norm": 0.
|
| 306 |
-
"learning_rate": 5.
|
| 307 |
-
"loss": 0.
|
| 308 |
"step": 3500
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 14.14,
|
| 312 |
-
"eval_accuracy": 0.
|
| 313 |
-
"eval_loss": 0.
|
| 314 |
-
"eval_runtime":
|
| 315 |
-
"eval_samples_per_second":
|
| 316 |
-
"eval_steps_per_second":
|
| 317 |
"step": 3500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 318 |
}
|
| 319 |
],
|
| 320 |
"logging_steps": 100,
|
|
@@ -322,7 +345,7 @@
|
|
| 322 |
"num_input_tokens_seen": 0,
|
| 323 |
"num_train_epochs": 15,
|
| 324 |
"save_steps": 500,
|
| 325 |
-
"total_flos":
|
| 326 |
"train_batch_size": 4,
|
| 327 |
"trial_name": null,
|
| 328 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.08732648193836212,
|
| 3 |
+
"best_model_checkpoint": "/kaggle/working/ckpts/checkpoint-3000",
|
| 4 |
+
"epoch": 14.969696969696969,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 3705,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
"epoch": 0.4,
|
| 13 |
+
"grad_norm": 4.305359363555908,
|
| 14 |
"learning_rate": 9.730094466936572e-06,
|
| 15 |
+
"loss": 1.2225,
|
| 16 |
"step": 100
|
| 17 |
},
|
| 18 |
{
|
| 19 |
"epoch": 0.81,
|
| 20 |
+
"grad_norm": 4.125752925872803,
|
| 21 |
"learning_rate": 9.460188933873145e-06,
|
| 22 |
+
"loss": 0.6537,
|
| 23 |
"step": 200
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 1.21,
|
| 27 |
+
"grad_norm": 1.4285308122634888,
|
| 28 |
+
"learning_rate": 9.190283400809717e-06,
|
| 29 |
+
"loss": 0.3636,
|
| 30 |
"step": 300
|
| 31 |
},
|
| 32 |
{
|
| 33 |
"epoch": 1.62,
|
| 34 |
+
"grad_norm": 0.3685835301876068,
|
| 35 |
+
"learning_rate": 8.92037786774629e-06,
|
| 36 |
+
"loss": 0.2794,
|
| 37 |
"step": 400
|
| 38 |
},
|
| 39 |
{
|
| 40 |
"epoch": 2.02,
|
| 41 |
+
"grad_norm": 9.573077201843262,
|
| 42 |
+
"learning_rate": 8.653171390013496e-06,
|
| 43 |
+
"loss": 0.2538,
|
| 44 |
"step": 500
|
| 45 |
},
|
| 46 |
{
|
| 47 |
"epoch": 2.02,
|
| 48 |
+
"eval_accuracy": 0.9070706963539124,
|
| 49 |
+
"eval_loss": 0.2763901352882385,
|
| 50 |
+
"eval_runtime": 44.7688,
|
| 51 |
+
"eval_samples_per_second": 11.057,
|
| 52 |
+
"eval_steps_per_second": 2.77,
|
| 53 |
"step": 500
|
| 54 |
},
|
| 55 |
{
|
| 56 |
"epoch": 2.42,
|
| 57 |
+
"grad_norm": 23.30544090270996,
|
| 58 |
"learning_rate": 8.385964912280704e-06,
|
| 59 |
+
"loss": 0.1608,
|
| 60 |
"step": 600
|
| 61 |
},
|
| 62 |
{
|
| 63 |
"epoch": 2.83,
|
| 64 |
+
"grad_norm": 0.28133082389831543,
|
| 65 |
+
"learning_rate": 8.116059379217275e-06,
|
| 66 |
+
"loss": 0.149,
|
| 67 |
"step": 700
|
| 68 |
},
|
| 69 |
{
|
| 70 |
"epoch": 3.23,
|
| 71 |
+
"grad_norm": 20.600244522094727,
|
| 72 |
+
"learning_rate": 7.846153846153847e-06,
|
| 73 |
+
"loss": 0.1573,
|
| 74 |
"step": 800
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 3.64,
|
| 78 |
+
"grad_norm": 5.6504645347595215,
|
| 79 |
+
"learning_rate": 7.576248313090419e-06,
|
| 80 |
+
"loss": 0.1273,
|
| 81 |
"step": 900
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 4.04,
|
| 85 |
+
"grad_norm": 0.06874903291463852,
|
| 86 |
+
"learning_rate": 7.306342780026991e-06,
|
| 87 |
+
"loss": 0.0805,
|
| 88 |
"step": 1000
|
| 89 |
},
|
| 90 |
{
|
| 91 |
"epoch": 4.04,
|
| 92 |
+
"eval_accuracy": 0.9535353779792786,
|
| 93 |
+
"eval_loss": 0.1721607893705368,
|
| 94 |
+
"eval_runtime": 45.1822,
|
| 95 |
+
"eval_samples_per_second": 10.956,
|
| 96 |
+
"eval_steps_per_second": 2.744,
|
| 97 |
"step": 1000
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 4.44,
|
| 101 |
+
"grad_norm": 2.5783190727233887,
|
| 102 |
+
"learning_rate": 7.036437246963563e-06,
|
| 103 |
+
"loss": 0.124,
|
| 104 |
"step": 1100
|
| 105 |
},
|
| 106 |
{
|
| 107 |
"epoch": 4.85,
|
| 108 |
+
"grad_norm": 0.08274731785058975,
|
| 109 |
+
"learning_rate": 6.766531713900135e-06,
|
| 110 |
+
"loss": 0.0919,
|
| 111 |
"step": 1200
|
| 112 |
},
|
| 113 |
{
|
| 114 |
"epoch": 5.25,
|
| 115 |
+
"grad_norm": 0.7308838367462158,
|
| 116 |
+
"learning_rate": 6.496626180836708e-06,
|
| 117 |
+
"loss": 0.0882,
|
| 118 |
"step": 1300
|
| 119 |
},
|
| 120 |
{
|
| 121 |
"epoch": 5.66,
|
| 122 |
+
"grad_norm": 0.09978947788476944,
|
| 123 |
+
"learning_rate": 6.22672064777328e-06,
|
| 124 |
+
"loss": 0.0536,
|
| 125 |
"step": 1400
|
| 126 |
},
|
| 127 |
{
|
| 128 |
"epoch": 6.06,
|
| 129 |
+
"grad_norm": 0.02113749459385872,
|
| 130 |
+
"learning_rate": 5.956815114709852e-06,
|
| 131 |
+
"loss": 0.0505,
|
| 132 |
"step": 1500
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 6.06,
|
| 136 |
+
"eval_accuracy": 0.9777777791023254,
|
| 137 |
+
"eval_loss": 0.11419288069009781,
|
| 138 |
+
"eval_runtime": 45.0408,
|
| 139 |
+
"eval_samples_per_second": 10.99,
|
| 140 |
+
"eval_steps_per_second": 2.753,
|
| 141 |
"step": 1500
|
| 142 |
},
|
| 143 |
{
|
| 144 |
"epoch": 6.46,
|
| 145 |
+
"grad_norm": 0.04003220424056053,
|
| 146 |
+
"learning_rate": 5.686909581646424e-06,
|
| 147 |
+
"loss": 0.0283,
|
| 148 |
"step": 1600
|
| 149 |
},
|
| 150 |
{
|
| 151 |
"epoch": 6.87,
|
| 152 |
+
"grad_norm": 0.01778920739889145,
|
| 153 |
+
"learning_rate": 5.417004048582997e-06,
|
| 154 |
+
"loss": 0.0862,
|
| 155 |
"step": 1700
|
| 156 |
},
|
| 157 |
{
|
| 158 |
"epoch": 7.27,
|
| 159 |
+
"grad_norm": 2.5195560455322266,
|
| 160 |
+
"learning_rate": 5.147098515519568e-06,
|
| 161 |
+
"loss": 0.0623,
|
| 162 |
"step": 1800
|
| 163 |
},
|
| 164 |
{
|
| 165 |
"epoch": 7.68,
|
| 166 |
+
"grad_norm": 0.016020264476537704,
|
| 167 |
+
"learning_rate": 4.877192982456141e-06,
|
| 168 |
+
"loss": 0.0622,
|
| 169 |
"step": 1900
|
| 170 |
},
|
| 171 |
{
|
| 172 |
"epoch": 8.08,
|
| 173 |
+
"grad_norm": 0.1588612198829651,
|
| 174 |
+
"learning_rate": 4.607287449392713e-06,
|
| 175 |
+
"loss": 0.0466,
|
| 176 |
"step": 2000
|
| 177 |
},
|
| 178 |
{
|
| 179 |
"epoch": 8.08,
|
| 180 |
+
"eval_accuracy": 0.973737359046936,
|
| 181 |
+
"eval_loss": 0.11773889511823654,
|
| 182 |
+
"eval_runtime": 44.6026,
|
| 183 |
+
"eval_samples_per_second": 11.098,
|
| 184 |
+
"eval_steps_per_second": 2.78,
|
| 185 |
"step": 2000
|
| 186 |
},
|
| 187 |
{
|
| 188 |
"epoch": 8.48,
|
| 189 |
+
"grad_norm": 0.017447825521230698,
|
| 190 |
+
"learning_rate": 4.337381916329285e-06,
|
| 191 |
+
"loss": 0.0372,
|
| 192 |
"step": 2100
|
| 193 |
},
|
| 194 |
{
|
| 195 |
"epoch": 8.89,
|
| 196 |
+
"grad_norm": 0.07465946674346924,
|
| 197 |
+
"learning_rate": 4.067476383265857e-06,
|
| 198 |
+
"loss": 0.0338,
|
| 199 |
"step": 2200
|
| 200 |
},
|
| 201 |
{
|
| 202 |
"epoch": 9.29,
|
| 203 |
+
"grad_norm": 1.1038620471954346,
|
| 204 |
+
"learning_rate": 3.800269905533064e-06,
|
| 205 |
+
"loss": 0.0458,
|
| 206 |
"step": 2300
|
| 207 |
},
|
| 208 |
{
|
| 209 |
"epoch": 9.7,
|
| 210 |
+
"grad_norm": 0.033347614109516144,
|
| 211 |
+
"learning_rate": 3.5303643724696358e-06,
|
| 212 |
+
"loss": 0.0491,
|
| 213 |
"step": 2400
|
| 214 |
},
|
| 215 |
{
|
| 216 |
"epoch": 10.1,
|
| 217 |
+
"grad_norm": 0.010389924980700016,
|
| 218 |
+
"learning_rate": 3.260458839406208e-06,
|
| 219 |
+
"loss": 0.0269,
|
| 220 |
"step": 2500
|
| 221 |
},
|
| 222 |
{
|
| 223 |
"epoch": 10.1,
|
| 224 |
+
"eval_accuracy": 0.973737359046936,
|
| 225 |
+
"eval_loss": 0.11825177818536758,
|
| 226 |
+
"eval_runtime": 44.732,
|
| 227 |
+
"eval_samples_per_second": 11.066,
|
| 228 |
+
"eval_steps_per_second": 2.772,
|
| 229 |
"step": 2500
|
| 230 |
},
|
| 231 |
{
|
| 232 |
"epoch": 10.51,
|
| 233 |
+
"grad_norm": 0.014653917402029037,
|
| 234 |
+
"learning_rate": 2.9905533063427807e-06,
|
| 235 |
+
"loss": 0.0517,
|
| 236 |
"step": 2600
|
| 237 |
},
|
| 238 |
{
|
| 239 |
"epoch": 10.91,
|
| 240 |
+
"grad_norm": 0.012147756293416023,
|
| 241 |
+
"learning_rate": 2.7206477732793525e-06,
|
| 242 |
+
"loss": 0.0424,
|
| 243 |
"step": 2700
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"epoch": 11.31,
|
| 247 |
+
"grad_norm": 0.018718041479587555,
|
| 248 |
+
"learning_rate": 2.4507422402159244e-06,
|
| 249 |
+
"loss": 0.0216,
|
| 250 |
"step": 2800
|
| 251 |
},
|
| 252 |
{
|
| 253 |
"epoch": 11.72,
|
| 254 |
+
"grad_norm": 0.012194407172501087,
|
| 255 |
+
"learning_rate": 2.180836707152497e-06,
|
| 256 |
+
"loss": 0.0219,
|
| 257 |
"step": 2900
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 12.12,
|
| 261 |
+
"grad_norm": 0.06089940294623375,
|
| 262 |
+
"learning_rate": 1.910931174089069e-06,
|
| 263 |
+
"loss": 0.025,
|
| 264 |
"step": 3000
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 12.12,
|
| 268 |
+
"eval_accuracy": 0.9757575988769531,
|
| 269 |
+
"eval_loss": 0.08732648193836212,
|
| 270 |
+
"eval_runtime": 44.5561,
|
| 271 |
+
"eval_samples_per_second": 11.11,
|
| 272 |
+
"eval_steps_per_second": 2.783,
|
| 273 |
"step": 3000
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 12.53,
|
| 277 |
+
"grad_norm": 0.07689320296049118,
|
| 278 |
+
"learning_rate": 1.6410256410256412e-06,
|
| 279 |
+
"loss": 0.0398,
|
| 280 |
"step": 3100
|
| 281 |
},
|
| 282 |
{
|
| 283 |
"epoch": 12.93,
|
| 284 |
+
"grad_norm": 0.35102277994155884,
|
| 285 |
+
"learning_rate": 1.3711201079622132e-06,
|
| 286 |
+
"loss": 0.0289,
|
| 287 |
"step": 3200
|
| 288 |
},
|
| 289 |
{
|
| 290 |
"epoch": 13.33,
|
| 291 |
+
"grad_norm": 0.007283875718712807,
|
| 292 |
+
"learning_rate": 1.1012145748987855e-06,
|
| 293 |
+
"loss": 0.0169,
|
| 294 |
"step": 3300
|
| 295 |
},
|
| 296 |
{
|
| 297 |
"epoch": 13.74,
|
| 298 |
+
"grad_norm": 0.00747378496453166,
|
| 299 |
+
"learning_rate": 8.313090418353577e-07,
|
| 300 |
+
"loss": 0.0391,
|
| 301 |
"step": 3400
|
| 302 |
},
|
| 303 |
{
|
| 304 |
"epoch": 14.14,
|
| 305 |
+
"grad_norm": 0.0248583871871233,
|
| 306 |
+
"learning_rate": 5.614035087719299e-07,
|
| 307 |
+
"loss": 0.0323,
|
| 308 |
"step": 3500
|
| 309 |
},
|
| 310 |
{
|
| 311 |
"epoch": 14.14,
|
| 312 |
+
"eval_accuracy": 0.9797979593276978,
|
| 313 |
+
"eval_loss": 0.09203575551509857,
|
| 314 |
+
"eval_runtime": 46.8049,
|
| 315 |
+
"eval_samples_per_second": 10.576,
|
| 316 |
+
"eval_steps_per_second": 2.649,
|
| 317 |
"step": 3500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 14.55,
|
| 321 |
+
"grad_norm": 0.1301543265581131,
|
| 322 |
+
"learning_rate": 2.9149797570850203e-07,
|
| 323 |
+
"loss": 0.0252,
|
| 324 |
+
"step": 3600
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 14.95,
|
| 328 |
+
"grad_norm": 0.09905605018138885,
|
| 329 |
+
"learning_rate": 2.1592442645074224e-08,
|
| 330 |
+
"loss": 0.0146,
|
| 331 |
+
"step": 3700
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 14.97,
|
| 335 |
+
"step": 3705,
|
| 336 |
+
"total_flos": 3.163398064220592e+18,
|
| 337 |
+
"train_loss": 0.12699281216954733,
|
| 338 |
+
"train_runtime": 5204.2292,
|
| 339 |
+
"train_samples_per_second": 5.704,
|
| 340 |
+
"train_steps_per_second": 0.712
|
| 341 |
}
|
| 342 |
],
|
| 343 |
"logging_steps": 100,
|
|
|
|
| 345 |
"num_input_tokens_seen": 0,
|
| 346 |
"num_train_epochs": 15,
|
| 347 |
"save_steps": 500,
|
| 348 |
+
"total_flos": 3.163398064220592e+18,
|
| 349 |
"train_batch_size": 4,
|
| 350 |
"trial_name": null,
|
| 351 |
"trial_params": null
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4984
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac6c83dae5ee0fc426b289b9abef79c6451dd3a371db69497611b3ba3258e6ec
|
| 3 |
size 4984
|