Upload folder using huggingface_hub
Browse files- checkpoint-5000/config.json +130 -0
- checkpoint-5000/model.safetensors +3 -0
- checkpoint-5000/optimizer.pt +3 -0
- checkpoint-5000/preprocessor_config.json +9 -0
- checkpoint-5000/rng_state.pth +3 -0
- checkpoint-5000/scheduler.pt +3 -0
- checkpoint-5000/trainer_state.json +418 -0
- checkpoint-5000/training_args.bin +3 -0
- checkpoint-5500/config.json +130 -0
- checkpoint-5500/model.safetensors +3 -0
- checkpoint-5500/optimizer.pt +3 -0
- checkpoint-5500/preprocessor_config.json +9 -0
- checkpoint-5500/rng_state.pth +3 -0
- checkpoint-5500/scheduler.pt +3 -0
- checkpoint-5500/trainer_state.json +452 -0
- checkpoint-5500/training_args.bin +3 -0
- config.json +130 -0
- model.safetensors +3 -0
- preprocessor_config.json +9 -0
- runs/Jun08_06-23-52_d794ed0093e1/events.out.tfevents.1717827848.d794ed0093e1.34.0 +3 -0
- runs/Jun08_06-23-52_d794ed0093e1/events.out.tfevents.1717831961.d794ed0093e1.34.1 +3 -0
- runs/Jun08_07-33-35_d794ed0093e1/events.out.tfevents.1717832018.d794ed0093e1.34.2 +3 -0
- training_args.bin +3 -0
checkpoint-5000/config.json
ADDED
|
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base-960h",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 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 |
+
"Wav2Vec2ForSequenceClassification"
|
| 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": false,
|
| 18 |
+
"conv_dim": [
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512,
|
| 25 |
+
512
|
| 26 |
+
],
|
| 27 |
+
"conv_kernel": [
|
| 28 |
+
10,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
3,
|
| 33 |
+
2,
|
| 34 |
+
2
|
| 35 |
+
],
|
| 36 |
+
"conv_stride": [
|
| 37 |
+
5,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2,
|
| 43 |
+
2
|
| 44 |
+
],
|
| 45 |
+
"ctc_loss_reduction": "sum",
|
| 46 |
+
"ctc_zero_infinity": false,
|
| 47 |
+
"diversity_loss_weight": 0.1,
|
| 48 |
+
"do_stable_layer_norm": false,
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_dropout": 0.0,
|
| 52 |
+
"feat_extract_norm": "group",
|
| 53 |
+
"feat_proj_dropout": 0.1,
|
| 54 |
+
"feat_quantizer_dropout": 0.0,
|
| 55 |
+
"final_dropout": 0.1,
|
| 56 |
+
"gradient_checkpointing": false,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_dropout_prob": 0.1,
|
| 60 |
+
"hidden_size": 768,
|
| 61 |
+
"id2label": {
|
| 62 |
+
"0": "disgust",
|
| 63 |
+
"1": "angry",
|
| 64 |
+
"2": "neutral",
|
| 65 |
+
"3": "fear",
|
| 66 |
+
"4": "happy",
|
| 67 |
+
"5": "calm",
|
| 68 |
+
"6": "surprised",
|
| 69 |
+
"7": "sad"
|
| 70 |
+
},
|
| 71 |
+
"initializer_range": 0.02,
|
| 72 |
+
"intermediate_size": 3072,
|
| 73 |
+
"label2id": {
|
| 74 |
+
"angry": 1,
|
| 75 |
+
"calm": 5,
|
| 76 |
+
"disgust": 0,
|
| 77 |
+
"fear": 3,
|
| 78 |
+
"happy": 4,
|
| 79 |
+
"neutral": 2,
|
| 80 |
+
"sad": 7,
|
| 81 |
+
"surprised": 6
|
| 82 |
+
},
|
| 83 |
+
"layer_norm_eps": 1e-05,
|
| 84 |
+
"layerdrop": 0.1,
|
| 85 |
+
"mask_feature_length": 10,
|
| 86 |
+
"mask_feature_min_masks": 0,
|
| 87 |
+
"mask_feature_prob": 0.0,
|
| 88 |
+
"mask_time_length": 10,
|
| 89 |
+
"mask_time_min_masks": 2,
|
| 90 |
+
"mask_time_prob": 0.05,
|
| 91 |
+
"model_type": "wav2vec2",
|
| 92 |
+
"num_adapter_layers": 3,
|
| 93 |
+
"num_attention_heads": 12,
|
| 94 |
+
"num_codevector_groups": 2,
|
| 95 |
+
"num_codevectors_per_group": 320,
|
| 96 |
+
"num_conv_pos_embedding_groups": 16,
|
| 97 |
+
"num_conv_pos_embeddings": 128,
|
| 98 |
+
"num_feat_extract_layers": 7,
|
| 99 |
+
"num_hidden_layers": 12,
|
| 100 |
+
"num_negatives": 100,
|
| 101 |
+
"output_hidden_size": 768,
|
| 102 |
+
"pad_token_id": 0,
|
| 103 |
+
"proj_codevector_dim": 256,
|
| 104 |
+
"tdnn_dilation": [
|
| 105 |
+
1,
|
| 106 |
+
2,
|
| 107 |
+
3,
|
| 108 |
+
1,
|
| 109 |
+
1
|
| 110 |
+
],
|
| 111 |
+
"tdnn_dim": [
|
| 112 |
+
512,
|
| 113 |
+
512,
|
| 114 |
+
512,
|
| 115 |
+
512,
|
| 116 |
+
1500
|
| 117 |
+
],
|
| 118 |
+
"tdnn_kernel": [
|
| 119 |
+
5,
|
| 120 |
+
3,
|
| 121 |
+
3,
|
| 122 |
+
1,
|
| 123 |
+
1
|
| 124 |
+
],
|
| 125 |
+
"torch_dtype": "float32",
|
| 126 |
+
"transformers_version": "4.41.2",
|
| 127 |
+
"use_weighted_layer_sum": false,
|
| 128 |
+
"vocab_size": 32,
|
| 129 |
+
"xvector_output_dim": 512
|
| 130 |
+
}
|
checkpoint-5000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a667b4513e4cc22b6c0c24024a0fac3422e8bc89df2f9cdedf44d88b28b874b1
|
| 3 |
+
size 378308536
|
checkpoint-5000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd1d550c1fe6da739f86847053528e7e57d3dcdad7c8acf7f20f283224d2cf66
|
| 3 |
+
size 756744890
|
checkpoint-5000/preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": false,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
checkpoint-5000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e0702f033eaaf5995585b4a379e254dc99921d486cd670f73d61de6dcd34539
|
| 3 |
+
size 14244
|
checkpoint-5000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c50a05de136e80e7c2cf0aae061d12ff3b0d24f2c13805848848a7b3a4d036ee
|
| 3 |
+
size 1064
|
checkpoint-5000/trainer_state.json
ADDED
|
@@ -0,0 +1,418 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 35.2112676056338,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.20246478873239437,
|
| 14 |
+
"eval_loss": 1.9784369468688965,
|
| 15 |
+
"eval_runtime": 23.712,
|
| 16 |
+
"eval_samples_per_second": 23.954,
|
| 17 |
+
"eval_steps_per_second": 1.518,
|
| 18 |
+
"step": 142
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 2.0,
|
| 22 |
+
"eval_accuracy": 0.35563380281690143,
|
| 23 |
+
"eval_loss": 1.7192655801773071,
|
| 24 |
+
"eval_runtime": 23.7419,
|
| 25 |
+
"eval_samples_per_second": 23.924,
|
| 26 |
+
"eval_steps_per_second": 1.516,
|
| 27 |
+
"step": 284
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 3.0,
|
| 31 |
+
"eval_accuracy": 0.49295774647887325,
|
| 32 |
+
"eval_loss": 1.470097303390503,
|
| 33 |
+
"eval_runtime": 23.9735,
|
| 34 |
+
"eval_samples_per_second": 23.693,
|
| 35 |
+
"eval_steps_per_second": 1.502,
|
| 36 |
+
"step": 426
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 3.52112676056338,
|
| 40 |
+
"grad_norm": 13.082801818847656,
|
| 41 |
+
"learning_rate": 1.6478873239436623e-05,
|
| 42 |
+
"loss": 1.7619,
|
| 43 |
+
"step": 500
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 4.0,
|
| 47 |
+
"eval_accuracy": 0.6320422535211268,
|
| 48 |
+
"eval_loss": 1.233864188194275,
|
| 49 |
+
"eval_runtime": 23.6712,
|
| 50 |
+
"eval_samples_per_second": 23.995,
|
| 51 |
+
"eval_steps_per_second": 1.521,
|
| 52 |
+
"step": 568
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 5.0,
|
| 56 |
+
"eval_accuracy": 0.6091549295774648,
|
| 57 |
+
"eval_loss": 1.0945910215377808,
|
| 58 |
+
"eval_runtime": 23.687,
|
| 59 |
+
"eval_samples_per_second": 23.979,
|
| 60 |
+
"eval_steps_per_second": 1.52,
|
| 61 |
+
"step": 710
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 6.0,
|
| 65 |
+
"eval_accuracy": 0.7429577464788732,
|
| 66 |
+
"eval_loss": 0.8935854434967041,
|
| 67 |
+
"eval_runtime": 24.0091,
|
| 68 |
+
"eval_samples_per_second": 23.658,
|
| 69 |
+
"eval_steps_per_second": 1.499,
|
| 70 |
+
"step": 852
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 7.0,
|
| 74 |
+
"eval_accuracy": 0.7535211267605634,
|
| 75 |
+
"eval_loss": 0.8685159683227539,
|
| 76 |
+
"eval_runtime": 23.6401,
|
| 77 |
+
"eval_samples_per_second": 24.027,
|
| 78 |
+
"eval_steps_per_second": 1.523,
|
| 79 |
+
"step": 994
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 7.042253521126761,
|
| 83 |
+
"grad_norm": 7.731432914733887,
|
| 84 |
+
"learning_rate": 1.2957746478873242e-05,
|
| 85 |
+
"loss": 1.056,
|
| 86 |
+
"step": 1000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 8.0,
|
| 90 |
+
"eval_accuracy": 0.7816901408450704,
|
| 91 |
+
"eval_loss": 0.7374693155288696,
|
| 92 |
+
"eval_runtime": 23.6523,
|
| 93 |
+
"eval_samples_per_second": 24.015,
|
| 94 |
+
"eval_steps_per_second": 1.522,
|
| 95 |
+
"step": 1136
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 9.0,
|
| 99 |
+
"eval_accuracy": 0.7834507042253521,
|
| 100 |
+
"eval_loss": 0.6636029481887817,
|
| 101 |
+
"eval_runtime": 23.6027,
|
| 102 |
+
"eval_samples_per_second": 24.065,
|
| 103 |
+
"eval_steps_per_second": 1.525,
|
| 104 |
+
"step": 1278
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 10.0,
|
| 108 |
+
"eval_accuracy": 0.823943661971831,
|
| 109 |
+
"eval_loss": 0.5628218650817871,
|
| 110 |
+
"eval_runtime": 23.5285,
|
| 111 |
+
"eval_samples_per_second": 24.141,
|
| 112 |
+
"eval_steps_per_second": 1.53,
|
| 113 |
+
"step": 1420
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 10.56338028169014,
|
| 117 |
+
"grad_norm": 4.4197211265563965,
|
| 118 |
+
"learning_rate": 9.43661971830986e-06,
|
| 119 |
+
"loss": 0.6356,
|
| 120 |
+
"step": 1500
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 11.0,
|
| 124 |
+
"eval_accuracy": 0.8116197183098591,
|
| 125 |
+
"eval_loss": 0.5596714019775391,
|
| 126 |
+
"eval_runtime": 23.7086,
|
| 127 |
+
"eval_samples_per_second": 23.958,
|
| 128 |
+
"eval_steps_per_second": 1.518,
|
| 129 |
+
"step": 1562
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 12.0,
|
| 133 |
+
"eval_accuracy": 0.7764084507042254,
|
| 134 |
+
"eval_loss": 0.6482901573181152,
|
| 135 |
+
"eval_runtime": 23.7013,
|
| 136 |
+
"eval_samples_per_second": 23.965,
|
| 137 |
+
"eval_steps_per_second": 1.519,
|
| 138 |
+
"step": 1704
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 13.0,
|
| 142 |
+
"eval_accuracy": 0.823943661971831,
|
| 143 |
+
"eval_loss": 0.5288650989532471,
|
| 144 |
+
"eval_runtime": 23.642,
|
| 145 |
+
"eval_samples_per_second": 24.025,
|
| 146 |
+
"eval_steps_per_second": 1.523,
|
| 147 |
+
"step": 1846
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 14.0,
|
| 151 |
+
"eval_accuracy": 0.8397887323943662,
|
| 152 |
+
"eval_loss": 0.49578651785850525,
|
| 153 |
+
"eval_runtime": 23.6946,
|
| 154 |
+
"eval_samples_per_second": 23.972,
|
| 155 |
+
"eval_steps_per_second": 1.519,
|
| 156 |
+
"step": 1988
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 14.084507042253522,
|
| 160 |
+
"grad_norm": 9.863320350646973,
|
| 161 |
+
"learning_rate": 5.915492957746479e-06,
|
| 162 |
+
"loss": 0.434,
|
| 163 |
+
"step": 2000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 15.0,
|
| 167 |
+
"eval_accuracy": 0.8221830985915493,
|
| 168 |
+
"eval_loss": 0.5109295845031738,
|
| 169 |
+
"eval_runtime": 23.7207,
|
| 170 |
+
"eval_samples_per_second": 23.945,
|
| 171 |
+
"eval_steps_per_second": 1.518,
|
| 172 |
+
"step": 2130
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 16.0,
|
| 176 |
+
"eval_accuracy": 0.8257042253521126,
|
| 177 |
+
"eval_loss": 0.46026095747947693,
|
| 178 |
+
"eval_runtime": 23.6246,
|
| 179 |
+
"eval_samples_per_second": 24.043,
|
| 180 |
+
"eval_steps_per_second": 1.524,
|
| 181 |
+
"step": 2272
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 17.0,
|
| 185 |
+
"eval_accuracy": 0.8362676056338029,
|
| 186 |
+
"eval_loss": 0.48158177733421326,
|
| 187 |
+
"eval_runtime": 24.012,
|
| 188 |
+
"eval_samples_per_second": 23.655,
|
| 189 |
+
"eval_steps_per_second": 1.499,
|
| 190 |
+
"step": 2414
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 17.6056338028169,
|
| 194 |
+
"grad_norm": 13.80960464477539,
|
| 195 |
+
"learning_rate": 2.3943661971830984e-06,
|
| 196 |
+
"loss": 0.3225,
|
| 197 |
+
"step": 2500
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 18.0,
|
| 201 |
+
"eval_accuracy": 0.9049295774647887,
|
| 202 |
+
"eval_loss": 0.26005274057388306,
|
| 203 |
+
"eval_runtime": 23.6863,
|
| 204 |
+
"eval_samples_per_second": 23.98,
|
| 205 |
+
"eval_steps_per_second": 1.52,
|
| 206 |
+
"step": 2556
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 19.0,
|
| 210 |
+
"eval_accuracy": 0.8961267605633803,
|
| 211 |
+
"eval_loss": 0.24130958318710327,
|
| 212 |
+
"eval_runtime": 23.7497,
|
| 213 |
+
"eval_samples_per_second": 23.916,
|
| 214 |
+
"eval_steps_per_second": 1.516,
|
| 215 |
+
"step": 2698
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 20.0,
|
| 219 |
+
"eval_accuracy": 0.9014084507042254,
|
| 220 |
+
"eval_loss": 0.24745188653469086,
|
| 221 |
+
"eval_runtime": 23.7635,
|
| 222 |
+
"eval_samples_per_second": 23.902,
|
| 223 |
+
"eval_steps_per_second": 1.515,
|
| 224 |
+
"step": 2840
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 21.0,
|
| 228 |
+
"eval_accuracy": 0.9084507042253521,
|
| 229 |
+
"eval_loss": 0.19828735291957855,
|
| 230 |
+
"eval_runtime": 23.7046,
|
| 231 |
+
"eval_samples_per_second": 23.962,
|
| 232 |
+
"eval_steps_per_second": 1.519,
|
| 233 |
+
"step": 2982
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 21.12676056338028,
|
| 237 |
+
"grad_norm": 4.427192211151123,
|
| 238 |
+
"learning_rate": 9.43661971830986e-06,
|
| 239 |
+
"loss": 0.3404,
|
| 240 |
+
"step": 3000
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 22.0,
|
| 244 |
+
"eval_accuracy": 0.903169014084507,
|
| 245 |
+
"eval_loss": 0.21990956366062164,
|
| 246 |
+
"eval_runtime": 23.6099,
|
| 247 |
+
"eval_samples_per_second": 24.058,
|
| 248 |
+
"eval_steps_per_second": 1.525,
|
| 249 |
+
"step": 3124
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 23.0,
|
| 253 |
+
"eval_accuracy": 0.9049295774647887,
|
| 254 |
+
"eval_loss": 0.22014880180358887,
|
| 255 |
+
"eval_runtime": 23.6019,
|
| 256 |
+
"eval_samples_per_second": 24.066,
|
| 257 |
+
"eval_steps_per_second": 1.525,
|
| 258 |
+
"step": 3266
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 24.0,
|
| 262 |
+
"eval_accuracy": 0.897887323943662,
|
| 263 |
+
"eval_loss": 0.2570597231388092,
|
| 264 |
+
"eval_runtime": 23.7096,
|
| 265 |
+
"eval_samples_per_second": 23.957,
|
| 266 |
+
"eval_steps_per_second": 1.518,
|
| 267 |
+
"step": 3408
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 24.647887323943664,
|
| 271 |
+
"grad_norm": 16.97844886779785,
|
| 272 |
+
"learning_rate": 7.67605633802817e-06,
|
| 273 |
+
"loss": 0.265,
|
| 274 |
+
"step": 3500
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 25.0,
|
| 278 |
+
"eval_accuracy": 0.897887323943662,
|
| 279 |
+
"eval_loss": 0.24254758656024933,
|
| 280 |
+
"eval_runtime": 23.4896,
|
| 281 |
+
"eval_samples_per_second": 24.181,
|
| 282 |
+
"eval_steps_per_second": 1.533,
|
| 283 |
+
"step": 3550
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 26.0,
|
| 287 |
+
"eval_accuracy": 0.9119718309859155,
|
| 288 |
+
"eval_loss": 0.24259935319423676,
|
| 289 |
+
"eval_runtime": 23.5966,
|
| 290 |
+
"eval_samples_per_second": 24.071,
|
| 291 |
+
"eval_steps_per_second": 1.526,
|
| 292 |
+
"step": 3692
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 27.0,
|
| 296 |
+
"eval_accuracy": 0.8908450704225352,
|
| 297 |
+
"eval_loss": 0.278877854347229,
|
| 298 |
+
"eval_runtime": 23.6986,
|
| 299 |
+
"eval_samples_per_second": 23.968,
|
| 300 |
+
"eval_steps_per_second": 1.519,
|
| 301 |
+
"step": 3834
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 28.0,
|
| 305 |
+
"eval_accuracy": 0.897887323943662,
|
| 306 |
+
"eval_loss": 0.2662661075592041,
|
| 307 |
+
"eval_runtime": 23.6515,
|
| 308 |
+
"eval_samples_per_second": 24.015,
|
| 309 |
+
"eval_steps_per_second": 1.522,
|
| 310 |
+
"step": 3976
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 28.169014084507044,
|
| 314 |
+
"grad_norm": 10.978034973144531,
|
| 315 |
+
"learning_rate": 5.915492957746479e-06,
|
| 316 |
+
"loss": 0.2065,
|
| 317 |
+
"step": 4000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 29.0,
|
| 321 |
+
"eval_accuracy": 0.9049295774647887,
|
| 322 |
+
"eval_loss": 0.24609780311584473,
|
| 323 |
+
"eval_runtime": 23.6764,
|
| 324 |
+
"eval_samples_per_second": 23.99,
|
| 325 |
+
"eval_steps_per_second": 1.52,
|
| 326 |
+
"step": 4118
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 30.0,
|
| 330 |
+
"eval_accuracy": 0.9066901408450704,
|
| 331 |
+
"eval_loss": 0.21895237267017365,
|
| 332 |
+
"eval_runtime": 23.7437,
|
| 333 |
+
"eval_samples_per_second": 23.922,
|
| 334 |
+
"eval_steps_per_second": 1.516,
|
| 335 |
+
"step": 4260
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 31.0,
|
| 339 |
+
"eval_accuracy": 0.8926056338028169,
|
| 340 |
+
"eval_loss": 0.24163208901882172,
|
| 341 |
+
"eval_runtime": 23.5335,
|
| 342 |
+
"eval_samples_per_second": 24.136,
|
| 343 |
+
"eval_steps_per_second": 1.53,
|
| 344 |
+
"step": 4402
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 31.690140845070424,
|
| 348 |
+
"grad_norm": 0.09197159856557846,
|
| 349 |
+
"learning_rate": 4.154929577464789e-06,
|
| 350 |
+
"loss": 0.1812,
|
| 351 |
+
"step": 4500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 32.0,
|
| 355 |
+
"eval_accuracy": 0.8961267605633803,
|
| 356 |
+
"eval_loss": 0.2705635726451874,
|
| 357 |
+
"eval_runtime": 23.6938,
|
| 358 |
+
"eval_samples_per_second": 23.972,
|
| 359 |
+
"eval_steps_per_second": 1.519,
|
| 360 |
+
"step": 4544
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 33.0,
|
| 364 |
+
"eval_accuracy": 0.9049295774647887,
|
| 365 |
+
"eval_loss": 0.225112184882164,
|
| 366 |
+
"eval_runtime": 23.7202,
|
| 367 |
+
"eval_samples_per_second": 23.946,
|
| 368 |
+
"eval_steps_per_second": 1.518,
|
| 369 |
+
"step": 4686
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 34.0,
|
| 373 |
+
"eval_accuracy": 0.9084507042253521,
|
| 374 |
+
"eval_loss": 0.22959274053573608,
|
| 375 |
+
"eval_runtime": 23.6809,
|
| 376 |
+
"eval_samples_per_second": 23.986,
|
| 377 |
+
"eval_steps_per_second": 1.52,
|
| 378 |
+
"step": 4828
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 35.0,
|
| 382 |
+
"eval_accuracy": 0.9049295774647887,
|
| 383 |
+
"eval_loss": 0.22617577016353607,
|
| 384 |
+
"eval_runtime": 23.6903,
|
| 385 |
+
"eval_samples_per_second": 23.976,
|
| 386 |
+
"eval_steps_per_second": 1.52,
|
| 387 |
+
"step": 4970
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 35.2112676056338,
|
| 391 |
+
"grad_norm": 0.8059929013252258,
|
| 392 |
+
"learning_rate": 2.3943661971830984e-06,
|
| 393 |
+
"loss": 0.1701,
|
| 394 |
+
"step": 5000
|
| 395 |
+
}
|
| 396 |
+
],
|
| 397 |
+
"logging_steps": 500,
|
| 398 |
+
"max_steps": 5680,
|
| 399 |
+
"num_input_tokens_seen": 0,
|
| 400 |
+
"num_train_epochs": 40,
|
| 401 |
+
"save_steps": 500,
|
| 402 |
+
"stateful_callbacks": {
|
| 403 |
+
"TrainerControl": {
|
| 404 |
+
"args": {
|
| 405 |
+
"should_epoch_stop": false,
|
| 406 |
+
"should_evaluate": false,
|
| 407 |
+
"should_log": false,
|
| 408 |
+
"should_save": true,
|
| 409 |
+
"should_training_stop": false
|
| 410 |
+
},
|
| 411 |
+
"attributes": {}
|
| 412 |
+
}
|
| 413 |
+
},
|
| 414 |
+
"total_flos": 3.071488373035011e+18,
|
| 415 |
+
"train_batch_size": 16,
|
| 416 |
+
"trial_name": null,
|
| 417 |
+
"trial_params": null
|
| 418 |
+
}
|
checkpoint-5000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2a19c28b833bce99fad445d871a86144944aa23f823d74c7d052a88adbca6e3
|
| 3 |
+
size 5112
|
checkpoint-5500/config.json
ADDED
|
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base-960h",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 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 |
+
"Wav2Vec2ForSequenceClassification"
|
| 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": false,
|
| 18 |
+
"conv_dim": [
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512,
|
| 25 |
+
512
|
| 26 |
+
],
|
| 27 |
+
"conv_kernel": [
|
| 28 |
+
10,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
3,
|
| 33 |
+
2,
|
| 34 |
+
2
|
| 35 |
+
],
|
| 36 |
+
"conv_stride": [
|
| 37 |
+
5,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2,
|
| 43 |
+
2
|
| 44 |
+
],
|
| 45 |
+
"ctc_loss_reduction": "sum",
|
| 46 |
+
"ctc_zero_infinity": false,
|
| 47 |
+
"diversity_loss_weight": 0.1,
|
| 48 |
+
"do_stable_layer_norm": false,
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_dropout": 0.0,
|
| 52 |
+
"feat_extract_norm": "group",
|
| 53 |
+
"feat_proj_dropout": 0.1,
|
| 54 |
+
"feat_quantizer_dropout": 0.0,
|
| 55 |
+
"final_dropout": 0.1,
|
| 56 |
+
"gradient_checkpointing": false,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_dropout_prob": 0.1,
|
| 60 |
+
"hidden_size": 768,
|
| 61 |
+
"id2label": {
|
| 62 |
+
"0": "disgust",
|
| 63 |
+
"1": "angry",
|
| 64 |
+
"2": "neutral",
|
| 65 |
+
"3": "fear",
|
| 66 |
+
"4": "happy",
|
| 67 |
+
"5": "calm",
|
| 68 |
+
"6": "surprised",
|
| 69 |
+
"7": "sad"
|
| 70 |
+
},
|
| 71 |
+
"initializer_range": 0.02,
|
| 72 |
+
"intermediate_size": 3072,
|
| 73 |
+
"label2id": {
|
| 74 |
+
"angry": 1,
|
| 75 |
+
"calm": 5,
|
| 76 |
+
"disgust": 0,
|
| 77 |
+
"fear": 3,
|
| 78 |
+
"happy": 4,
|
| 79 |
+
"neutral": 2,
|
| 80 |
+
"sad": 7,
|
| 81 |
+
"surprised": 6
|
| 82 |
+
},
|
| 83 |
+
"layer_norm_eps": 1e-05,
|
| 84 |
+
"layerdrop": 0.1,
|
| 85 |
+
"mask_feature_length": 10,
|
| 86 |
+
"mask_feature_min_masks": 0,
|
| 87 |
+
"mask_feature_prob": 0.0,
|
| 88 |
+
"mask_time_length": 10,
|
| 89 |
+
"mask_time_min_masks": 2,
|
| 90 |
+
"mask_time_prob": 0.05,
|
| 91 |
+
"model_type": "wav2vec2",
|
| 92 |
+
"num_adapter_layers": 3,
|
| 93 |
+
"num_attention_heads": 12,
|
| 94 |
+
"num_codevector_groups": 2,
|
| 95 |
+
"num_codevectors_per_group": 320,
|
| 96 |
+
"num_conv_pos_embedding_groups": 16,
|
| 97 |
+
"num_conv_pos_embeddings": 128,
|
| 98 |
+
"num_feat_extract_layers": 7,
|
| 99 |
+
"num_hidden_layers": 12,
|
| 100 |
+
"num_negatives": 100,
|
| 101 |
+
"output_hidden_size": 768,
|
| 102 |
+
"pad_token_id": 0,
|
| 103 |
+
"proj_codevector_dim": 256,
|
| 104 |
+
"tdnn_dilation": [
|
| 105 |
+
1,
|
| 106 |
+
2,
|
| 107 |
+
3,
|
| 108 |
+
1,
|
| 109 |
+
1
|
| 110 |
+
],
|
| 111 |
+
"tdnn_dim": [
|
| 112 |
+
512,
|
| 113 |
+
512,
|
| 114 |
+
512,
|
| 115 |
+
512,
|
| 116 |
+
1500
|
| 117 |
+
],
|
| 118 |
+
"tdnn_kernel": [
|
| 119 |
+
5,
|
| 120 |
+
3,
|
| 121 |
+
3,
|
| 122 |
+
1,
|
| 123 |
+
1
|
| 124 |
+
],
|
| 125 |
+
"torch_dtype": "float32",
|
| 126 |
+
"transformers_version": "4.41.2",
|
| 127 |
+
"use_weighted_layer_sum": false,
|
| 128 |
+
"vocab_size": 32,
|
| 129 |
+
"xvector_output_dim": 512
|
| 130 |
+
}
|
checkpoint-5500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2a5b21731233895a5110f9c075304145504d53b8361202f435659fe77e4d6aa
|
| 3 |
+
size 378308536
|
checkpoint-5500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b68b239c251be76c1f2b7f05c3281561187d94b0feaf509fd96debb6b6e55b43
|
| 3 |
+
size 756744890
|
checkpoint-5500/preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": false,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
checkpoint-5500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ceb312bf2339a86f727278122a43dc2fcf976b0e6d402ae8708aa4d381e493ef
|
| 3 |
+
size 14244
|
checkpoint-5500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8b21f6ff1eb8266647c6296ec655ba53fd867a7045c97b8936d9ef2d1b90988
|
| 3 |
+
size 1064
|
checkpoint-5500/trainer_state.json
ADDED
|
@@ -0,0 +1,452 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 38.732394366197184,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"eval_accuracy": 0.20246478873239437,
|
| 14 |
+
"eval_loss": 1.9784369468688965,
|
| 15 |
+
"eval_runtime": 23.712,
|
| 16 |
+
"eval_samples_per_second": 23.954,
|
| 17 |
+
"eval_steps_per_second": 1.518,
|
| 18 |
+
"step": 142
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 2.0,
|
| 22 |
+
"eval_accuracy": 0.35563380281690143,
|
| 23 |
+
"eval_loss": 1.7192655801773071,
|
| 24 |
+
"eval_runtime": 23.7419,
|
| 25 |
+
"eval_samples_per_second": 23.924,
|
| 26 |
+
"eval_steps_per_second": 1.516,
|
| 27 |
+
"step": 284
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 3.0,
|
| 31 |
+
"eval_accuracy": 0.49295774647887325,
|
| 32 |
+
"eval_loss": 1.470097303390503,
|
| 33 |
+
"eval_runtime": 23.9735,
|
| 34 |
+
"eval_samples_per_second": 23.693,
|
| 35 |
+
"eval_steps_per_second": 1.502,
|
| 36 |
+
"step": 426
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 3.52112676056338,
|
| 40 |
+
"grad_norm": 13.082801818847656,
|
| 41 |
+
"learning_rate": 1.6478873239436623e-05,
|
| 42 |
+
"loss": 1.7619,
|
| 43 |
+
"step": 500
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 4.0,
|
| 47 |
+
"eval_accuracy": 0.6320422535211268,
|
| 48 |
+
"eval_loss": 1.233864188194275,
|
| 49 |
+
"eval_runtime": 23.6712,
|
| 50 |
+
"eval_samples_per_second": 23.995,
|
| 51 |
+
"eval_steps_per_second": 1.521,
|
| 52 |
+
"step": 568
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 5.0,
|
| 56 |
+
"eval_accuracy": 0.6091549295774648,
|
| 57 |
+
"eval_loss": 1.0945910215377808,
|
| 58 |
+
"eval_runtime": 23.687,
|
| 59 |
+
"eval_samples_per_second": 23.979,
|
| 60 |
+
"eval_steps_per_second": 1.52,
|
| 61 |
+
"step": 710
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 6.0,
|
| 65 |
+
"eval_accuracy": 0.7429577464788732,
|
| 66 |
+
"eval_loss": 0.8935854434967041,
|
| 67 |
+
"eval_runtime": 24.0091,
|
| 68 |
+
"eval_samples_per_second": 23.658,
|
| 69 |
+
"eval_steps_per_second": 1.499,
|
| 70 |
+
"step": 852
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 7.0,
|
| 74 |
+
"eval_accuracy": 0.7535211267605634,
|
| 75 |
+
"eval_loss": 0.8685159683227539,
|
| 76 |
+
"eval_runtime": 23.6401,
|
| 77 |
+
"eval_samples_per_second": 24.027,
|
| 78 |
+
"eval_steps_per_second": 1.523,
|
| 79 |
+
"step": 994
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 7.042253521126761,
|
| 83 |
+
"grad_norm": 7.731432914733887,
|
| 84 |
+
"learning_rate": 1.2957746478873242e-05,
|
| 85 |
+
"loss": 1.056,
|
| 86 |
+
"step": 1000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 8.0,
|
| 90 |
+
"eval_accuracy": 0.7816901408450704,
|
| 91 |
+
"eval_loss": 0.7374693155288696,
|
| 92 |
+
"eval_runtime": 23.6523,
|
| 93 |
+
"eval_samples_per_second": 24.015,
|
| 94 |
+
"eval_steps_per_second": 1.522,
|
| 95 |
+
"step": 1136
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 9.0,
|
| 99 |
+
"eval_accuracy": 0.7834507042253521,
|
| 100 |
+
"eval_loss": 0.6636029481887817,
|
| 101 |
+
"eval_runtime": 23.6027,
|
| 102 |
+
"eval_samples_per_second": 24.065,
|
| 103 |
+
"eval_steps_per_second": 1.525,
|
| 104 |
+
"step": 1278
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 10.0,
|
| 108 |
+
"eval_accuracy": 0.823943661971831,
|
| 109 |
+
"eval_loss": 0.5628218650817871,
|
| 110 |
+
"eval_runtime": 23.5285,
|
| 111 |
+
"eval_samples_per_second": 24.141,
|
| 112 |
+
"eval_steps_per_second": 1.53,
|
| 113 |
+
"step": 1420
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 10.56338028169014,
|
| 117 |
+
"grad_norm": 4.4197211265563965,
|
| 118 |
+
"learning_rate": 9.43661971830986e-06,
|
| 119 |
+
"loss": 0.6356,
|
| 120 |
+
"step": 1500
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 11.0,
|
| 124 |
+
"eval_accuracy": 0.8116197183098591,
|
| 125 |
+
"eval_loss": 0.5596714019775391,
|
| 126 |
+
"eval_runtime": 23.7086,
|
| 127 |
+
"eval_samples_per_second": 23.958,
|
| 128 |
+
"eval_steps_per_second": 1.518,
|
| 129 |
+
"step": 1562
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 12.0,
|
| 133 |
+
"eval_accuracy": 0.7764084507042254,
|
| 134 |
+
"eval_loss": 0.6482901573181152,
|
| 135 |
+
"eval_runtime": 23.7013,
|
| 136 |
+
"eval_samples_per_second": 23.965,
|
| 137 |
+
"eval_steps_per_second": 1.519,
|
| 138 |
+
"step": 1704
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 13.0,
|
| 142 |
+
"eval_accuracy": 0.823943661971831,
|
| 143 |
+
"eval_loss": 0.5288650989532471,
|
| 144 |
+
"eval_runtime": 23.642,
|
| 145 |
+
"eval_samples_per_second": 24.025,
|
| 146 |
+
"eval_steps_per_second": 1.523,
|
| 147 |
+
"step": 1846
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 14.0,
|
| 151 |
+
"eval_accuracy": 0.8397887323943662,
|
| 152 |
+
"eval_loss": 0.49578651785850525,
|
| 153 |
+
"eval_runtime": 23.6946,
|
| 154 |
+
"eval_samples_per_second": 23.972,
|
| 155 |
+
"eval_steps_per_second": 1.519,
|
| 156 |
+
"step": 1988
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 14.084507042253522,
|
| 160 |
+
"grad_norm": 9.863320350646973,
|
| 161 |
+
"learning_rate": 5.915492957746479e-06,
|
| 162 |
+
"loss": 0.434,
|
| 163 |
+
"step": 2000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 15.0,
|
| 167 |
+
"eval_accuracy": 0.8221830985915493,
|
| 168 |
+
"eval_loss": 0.5109295845031738,
|
| 169 |
+
"eval_runtime": 23.7207,
|
| 170 |
+
"eval_samples_per_second": 23.945,
|
| 171 |
+
"eval_steps_per_second": 1.518,
|
| 172 |
+
"step": 2130
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 16.0,
|
| 176 |
+
"eval_accuracy": 0.8257042253521126,
|
| 177 |
+
"eval_loss": 0.46026095747947693,
|
| 178 |
+
"eval_runtime": 23.6246,
|
| 179 |
+
"eval_samples_per_second": 24.043,
|
| 180 |
+
"eval_steps_per_second": 1.524,
|
| 181 |
+
"step": 2272
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 17.0,
|
| 185 |
+
"eval_accuracy": 0.8362676056338029,
|
| 186 |
+
"eval_loss": 0.48158177733421326,
|
| 187 |
+
"eval_runtime": 24.012,
|
| 188 |
+
"eval_samples_per_second": 23.655,
|
| 189 |
+
"eval_steps_per_second": 1.499,
|
| 190 |
+
"step": 2414
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 17.6056338028169,
|
| 194 |
+
"grad_norm": 13.80960464477539,
|
| 195 |
+
"learning_rate": 2.3943661971830984e-06,
|
| 196 |
+
"loss": 0.3225,
|
| 197 |
+
"step": 2500
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 18.0,
|
| 201 |
+
"eval_accuracy": 0.9049295774647887,
|
| 202 |
+
"eval_loss": 0.26005274057388306,
|
| 203 |
+
"eval_runtime": 23.6863,
|
| 204 |
+
"eval_samples_per_second": 23.98,
|
| 205 |
+
"eval_steps_per_second": 1.52,
|
| 206 |
+
"step": 2556
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 19.0,
|
| 210 |
+
"eval_accuracy": 0.8961267605633803,
|
| 211 |
+
"eval_loss": 0.24130958318710327,
|
| 212 |
+
"eval_runtime": 23.7497,
|
| 213 |
+
"eval_samples_per_second": 23.916,
|
| 214 |
+
"eval_steps_per_second": 1.516,
|
| 215 |
+
"step": 2698
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 20.0,
|
| 219 |
+
"eval_accuracy": 0.9014084507042254,
|
| 220 |
+
"eval_loss": 0.24745188653469086,
|
| 221 |
+
"eval_runtime": 23.7635,
|
| 222 |
+
"eval_samples_per_second": 23.902,
|
| 223 |
+
"eval_steps_per_second": 1.515,
|
| 224 |
+
"step": 2840
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 21.0,
|
| 228 |
+
"eval_accuracy": 0.9084507042253521,
|
| 229 |
+
"eval_loss": 0.19828735291957855,
|
| 230 |
+
"eval_runtime": 23.7046,
|
| 231 |
+
"eval_samples_per_second": 23.962,
|
| 232 |
+
"eval_steps_per_second": 1.519,
|
| 233 |
+
"step": 2982
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 21.12676056338028,
|
| 237 |
+
"grad_norm": 4.427192211151123,
|
| 238 |
+
"learning_rate": 9.43661971830986e-06,
|
| 239 |
+
"loss": 0.3404,
|
| 240 |
+
"step": 3000
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 22.0,
|
| 244 |
+
"eval_accuracy": 0.903169014084507,
|
| 245 |
+
"eval_loss": 0.21990956366062164,
|
| 246 |
+
"eval_runtime": 23.6099,
|
| 247 |
+
"eval_samples_per_second": 24.058,
|
| 248 |
+
"eval_steps_per_second": 1.525,
|
| 249 |
+
"step": 3124
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 23.0,
|
| 253 |
+
"eval_accuracy": 0.9049295774647887,
|
| 254 |
+
"eval_loss": 0.22014880180358887,
|
| 255 |
+
"eval_runtime": 23.6019,
|
| 256 |
+
"eval_samples_per_second": 24.066,
|
| 257 |
+
"eval_steps_per_second": 1.525,
|
| 258 |
+
"step": 3266
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 24.0,
|
| 262 |
+
"eval_accuracy": 0.897887323943662,
|
| 263 |
+
"eval_loss": 0.2570597231388092,
|
| 264 |
+
"eval_runtime": 23.7096,
|
| 265 |
+
"eval_samples_per_second": 23.957,
|
| 266 |
+
"eval_steps_per_second": 1.518,
|
| 267 |
+
"step": 3408
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 24.647887323943664,
|
| 271 |
+
"grad_norm": 16.97844886779785,
|
| 272 |
+
"learning_rate": 7.67605633802817e-06,
|
| 273 |
+
"loss": 0.265,
|
| 274 |
+
"step": 3500
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 25.0,
|
| 278 |
+
"eval_accuracy": 0.897887323943662,
|
| 279 |
+
"eval_loss": 0.24254758656024933,
|
| 280 |
+
"eval_runtime": 23.4896,
|
| 281 |
+
"eval_samples_per_second": 24.181,
|
| 282 |
+
"eval_steps_per_second": 1.533,
|
| 283 |
+
"step": 3550
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 26.0,
|
| 287 |
+
"eval_accuracy": 0.9119718309859155,
|
| 288 |
+
"eval_loss": 0.24259935319423676,
|
| 289 |
+
"eval_runtime": 23.5966,
|
| 290 |
+
"eval_samples_per_second": 24.071,
|
| 291 |
+
"eval_steps_per_second": 1.526,
|
| 292 |
+
"step": 3692
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 27.0,
|
| 296 |
+
"eval_accuracy": 0.8908450704225352,
|
| 297 |
+
"eval_loss": 0.278877854347229,
|
| 298 |
+
"eval_runtime": 23.6986,
|
| 299 |
+
"eval_samples_per_second": 23.968,
|
| 300 |
+
"eval_steps_per_second": 1.519,
|
| 301 |
+
"step": 3834
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 28.0,
|
| 305 |
+
"eval_accuracy": 0.897887323943662,
|
| 306 |
+
"eval_loss": 0.2662661075592041,
|
| 307 |
+
"eval_runtime": 23.6515,
|
| 308 |
+
"eval_samples_per_second": 24.015,
|
| 309 |
+
"eval_steps_per_second": 1.522,
|
| 310 |
+
"step": 3976
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 28.169014084507044,
|
| 314 |
+
"grad_norm": 10.978034973144531,
|
| 315 |
+
"learning_rate": 5.915492957746479e-06,
|
| 316 |
+
"loss": 0.2065,
|
| 317 |
+
"step": 4000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 29.0,
|
| 321 |
+
"eval_accuracy": 0.9049295774647887,
|
| 322 |
+
"eval_loss": 0.24609780311584473,
|
| 323 |
+
"eval_runtime": 23.6764,
|
| 324 |
+
"eval_samples_per_second": 23.99,
|
| 325 |
+
"eval_steps_per_second": 1.52,
|
| 326 |
+
"step": 4118
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 30.0,
|
| 330 |
+
"eval_accuracy": 0.9066901408450704,
|
| 331 |
+
"eval_loss": 0.21895237267017365,
|
| 332 |
+
"eval_runtime": 23.7437,
|
| 333 |
+
"eval_samples_per_second": 23.922,
|
| 334 |
+
"eval_steps_per_second": 1.516,
|
| 335 |
+
"step": 4260
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 31.0,
|
| 339 |
+
"eval_accuracy": 0.8926056338028169,
|
| 340 |
+
"eval_loss": 0.24163208901882172,
|
| 341 |
+
"eval_runtime": 23.5335,
|
| 342 |
+
"eval_samples_per_second": 24.136,
|
| 343 |
+
"eval_steps_per_second": 1.53,
|
| 344 |
+
"step": 4402
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 31.690140845070424,
|
| 348 |
+
"grad_norm": 0.09197159856557846,
|
| 349 |
+
"learning_rate": 4.154929577464789e-06,
|
| 350 |
+
"loss": 0.1812,
|
| 351 |
+
"step": 4500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 32.0,
|
| 355 |
+
"eval_accuracy": 0.8961267605633803,
|
| 356 |
+
"eval_loss": 0.2705635726451874,
|
| 357 |
+
"eval_runtime": 23.6938,
|
| 358 |
+
"eval_samples_per_second": 23.972,
|
| 359 |
+
"eval_steps_per_second": 1.519,
|
| 360 |
+
"step": 4544
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 33.0,
|
| 364 |
+
"eval_accuracy": 0.9049295774647887,
|
| 365 |
+
"eval_loss": 0.225112184882164,
|
| 366 |
+
"eval_runtime": 23.7202,
|
| 367 |
+
"eval_samples_per_second": 23.946,
|
| 368 |
+
"eval_steps_per_second": 1.518,
|
| 369 |
+
"step": 4686
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 34.0,
|
| 373 |
+
"eval_accuracy": 0.9084507042253521,
|
| 374 |
+
"eval_loss": 0.22959274053573608,
|
| 375 |
+
"eval_runtime": 23.6809,
|
| 376 |
+
"eval_samples_per_second": 23.986,
|
| 377 |
+
"eval_steps_per_second": 1.52,
|
| 378 |
+
"step": 4828
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 35.0,
|
| 382 |
+
"eval_accuracy": 0.9049295774647887,
|
| 383 |
+
"eval_loss": 0.22617577016353607,
|
| 384 |
+
"eval_runtime": 23.6903,
|
| 385 |
+
"eval_samples_per_second": 23.976,
|
| 386 |
+
"eval_steps_per_second": 1.52,
|
| 387 |
+
"step": 4970
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 35.2112676056338,
|
| 391 |
+
"grad_norm": 0.8059929013252258,
|
| 392 |
+
"learning_rate": 2.3943661971830984e-06,
|
| 393 |
+
"loss": 0.1701,
|
| 394 |
+
"step": 5000
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 36.0,
|
| 398 |
+
"eval_accuracy": 0.9014084507042254,
|
| 399 |
+
"eval_loss": 0.2291528433561325,
|
| 400 |
+
"eval_runtime": 23.6838,
|
| 401 |
+
"eval_samples_per_second": 23.983,
|
| 402 |
+
"eval_steps_per_second": 1.52,
|
| 403 |
+
"step": 5112
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 37.0,
|
| 407 |
+
"eval_accuracy": 0.9066901408450704,
|
| 408 |
+
"eval_loss": 0.24435779452323914,
|
| 409 |
+
"eval_runtime": 23.6984,
|
| 410 |
+
"eval_samples_per_second": 23.968,
|
| 411 |
+
"eval_steps_per_second": 1.519,
|
| 412 |
+
"step": 5254
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 38.0,
|
| 416 |
+
"eval_accuracy": 0.8943661971830986,
|
| 417 |
+
"eval_loss": 0.23858632147312164,
|
| 418 |
+
"eval_runtime": 24.1094,
|
| 419 |
+
"eval_samples_per_second": 23.559,
|
| 420 |
+
"eval_steps_per_second": 1.493,
|
| 421 |
+
"step": 5396
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 38.732394366197184,
|
| 425 |
+
"grad_norm": 1.8776308298110962,
|
| 426 |
+
"learning_rate": 6.338028169014085e-07,
|
| 427 |
+
"loss": 0.1465,
|
| 428 |
+
"step": 5500
|
| 429 |
+
}
|
| 430 |
+
],
|
| 431 |
+
"logging_steps": 500,
|
| 432 |
+
"max_steps": 5680,
|
| 433 |
+
"num_input_tokens_seen": 0,
|
| 434 |
+
"num_train_epochs": 40,
|
| 435 |
+
"save_steps": 500,
|
| 436 |
+
"stateful_callbacks": {
|
| 437 |
+
"TrainerControl": {
|
| 438 |
+
"args": {
|
| 439 |
+
"should_epoch_stop": false,
|
| 440 |
+
"should_evaluate": false,
|
| 441 |
+
"should_log": false,
|
| 442 |
+
"should_save": true,
|
| 443 |
+
"should_training_stop": false
|
| 444 |
+
},
|
| 445 |
+
"attributes": {}
|
| 446 |
+
}
|
| 447 |
+
},
|
| 448 |
+
"total_flos": 3.3775253992525896e+18,
|
| 449 |
+
"train_batch_size": 16,
|
| 450 |
+
"trial_name": null,
|
| 451 |
+
"trial_params": null
|
| 452 |
+
}
|
checkpoint-5500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2a19c28b833bce99fad445d871a86144944aa23f823d74c7d052a88adbca6e3
|
| 3 |
+
size 5112
|
config.json
ADDED
|
@@ -0,0 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base-960h",
|
| 3 |
+
"activation_dropout": 0.1,
|
| 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 |
+
"Wav2Vec2ForSequenceClassification"
|
| 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": false,
|
| 18 |
+
"conv_dim": [
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512,
|
| 25 |
+
512
|
| 26 |
+
],
|
| 27 |
+
"conv_kernel": [
|
| 28 |
+
10,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
3,
|
| 33 |
+
2,
|
| 34 |
+
2
|
| 35 |
+
],
|
| 36 |
+
"conv_stride": [
|
| 37 |
+
5,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2,
|
| 43 |
+
2
|
| 44 |
+
],
|
| 45 |
+
"ctc_loss_reduction": "sum",
|
| 46 |
+
"ctc_zero_infinity": false,
|
| 47 |
+
"diversity_loss_weight": 0.1,
|
| 48 |
+
"do_stable_layer_norm": false,
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_dropout": 0.0,
|
| 52 |
+
"feat_extract_norm": "group",
|
| 53 |
+
"feat_proj_dropout": 0.1,
|
| 54 |
+
"feat_quantizer_dropout": 0.0,
|
| 55 |
+
"final_dropout": 0.1,
|
| 56 |
+
"gradient_checkpointing": false,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_dropout_prob": 0.1,
|
| 60 |
+
"hidden_size": 768,
|
| 61 |
+
"id2label": {
|
| 62 |
+
"0": "disgust",
|
| 63 |
+
"1": "angry",
|
| 64 |
+
"2": "neutral",
|
| 65 |
+
"3": "fear",
|
| 66 |
+
"4": "happy",
|
| 67 |
+
"5": "calm",
|
| 68 |
+
"6": "surprised",
|
| 69 |
+
"7": "sad"
|
| 70 |
+
},
|
| 71 |
+
"initializer_range": 0.02,
|
| 72 |
+
"intermediate_size": 3072,
|
| 73 |
+
"label2id": {
|
| 74 |
+
"angry": 1,
|
| 75 |
+
"calm": 5,
|
| 76 |
+
"disgust": 0,
|
| 77 |
+
"fear": 3,
|
| 78 |
+
"happy": 4,
|
| 79 |
+
"neutral": 2,
|
| 80 |
+
"sad": 7,
|
| 81 |
+
"surprised": 6
|
| 82 |
+
},
|
| 83 |
+
"layer_norm_eps": 1e-05,
|
| 84 |
+
"layerdrop": 0.1,
|
| 85 |
+
"mask_feature_length": 10,
|
| 86 |
+
"mask_feature_min_masks": 0,
|
| 87 |
+
"mask_feature_prob": 0.0,
|
| 88 |
+
"mask_time_length": 10,
|
| 89 |
+
"mask_time_min_masks": 2,
|
| 90 |
+
"mask_time_prob": 0.05,
|
| 91 |
+
"model_type": "wav2vec2",
|
| 92 |
+
"num_adapter_layers": 3,
|
| 93 |
+
"num_attention_heads": 12,
|
| 94 |
+
"num_codevector_groups": 2,
|
| 95 |
+
"num_codevectors_per_group": 320,
|
| 96 |
+
"num_conv_pos_embedding_groups": 16,
|
| 97 |
+
"num_conv_pos_embeddings": 128,
|
| 98 |
+
"num_feat_extract_layers": 7,
|
| 99 |
+
"num_hidden_layers": 12,
|
| 100 |
+
"num_negatives": 100,
|
| 101 |
+
"output_hidden_size": 768,
|
| 102 |
+
"pad_token_id": 0,
|
| 103 |
+
"proj_codevector_dim": 256,
|
| 104 |
+
"tdnn_dilation": [
|
| 105 |
+
1,
|
| 106 |
+
2,
|
| 107 |
+
3,
|
| 108 |
+
1,
|
| 109 |
+
1
|
| 110 |
+
],
|
| 111 |
+
"tdnn_dim": [
|
| 112 |
+
512,
|
| 113 |
+
512,
|
| 114 |
+
512,
|
| 115 |
+
512,
|
| 116 |
+
1500
|
| 117 |
+
],
|
| 118 |
+
"tdnn_kernel": [
|
| 119 |
+
5,
|
| 120 |
+
3,
|
| 121 |
+
3,
|
| 122 |
+
1,
|
| 123 |
+
1
|
| 124 |
+
],
|
| 125 |
+
"torch_dtype": "float32",
|
| 126 |
+
"transformers_version": "4.41.2",
|
| 127 |
+
"use_weighted_layer_sum": false,
|
| 128 |
+
"vocab_size": 32,
|
| 129 |
+
"xvector_output_dim": 512
|
| 130 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5d09e251d6b6de447aff7b1b26f7e57a3753e0eabd0c97b04313cb82769a4377
|
| 3 |
+
size 378308536
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": false,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
runs/Jun08_06-23-52_d794ed0093e1/events.out.tfevents.1717827848.d794ed0093e1.34.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:583bfaf4ba213ff04b28e94c2c1d69eaef9265786b30c076406729c9c16f8e82
|
| 3 |
+
size 14463
|
runs/Jun08_06-23-52_d794ed0093e1/events.out.tfevents.1717831961.d794ed0093e1.34.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4108902e6b064ca9bc77ebd9429e36d9608b40bf4c6c92db5cd2ca70355855f8
|
| 3 |
+
size 6641
|
runs/Jun08_07-33-35_d794ed0093e1/events.out.tfevents.1717832018.d794ed0093e1.34.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad53a54928abb4c11c1dc7ca3c341287bb9124a5d587c13c3a5f4d5787b6984f
|
| 3 |
+
size 15630
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2a19c28b833bce99fad445d871a86144944aa23f823d74c7d052a88adbca6e3
|
| 3 |
+
size 5112
|