add model
Browse files- checkpoint-10000/config.json +88 -0
- checkpoint-10000/optimizer.pt +3 -0
- checkpoint-10000/preprocessor_config.json +9 -0
- checkpoint-10000/pytorch_model.bin +3 -0
- checkpoint-10000/rng_state.pth +3 -0
- checkpoint-10000/scaler.pt +3 -0
- checkpoint-10000/scheduler.pt +3 -0
- checkpoint-10000/trainer_state.json +316 -0
- checkpoint-10000/training_args.bin +3 -0
- checkpoint-10500/config.json +88 -0
- checkpoint-10500/optimizer.pt +3 -0
- checkpoint-10500/preprocessor_config.json +9 -0
- checkpoint-10500/pytorch_model.bin +3 -0
- checkpoint-10500/rng_state.pth +3 -0
- checkpoint-10500/scaler.pt +3 -0
- checkpoint-10500/scheduler.pt +3 -0
- checkpoint-10500/trainer_state.json +331 -0
- checkpoint-10500/training_args.bin +3 -0
- config.json +88 -0
- preprocessor_config.json +9 -0
- pytorch_model.bin +3 -0
- training_args.bin +3 -0
checkpoint-10000/config.json
ADDED
|
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"apply_spec_augment": true,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Wav2Vec2ForCTC"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.1,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"classifier_proj_size": 256,
|
| 11 |
+
"codevector_dim": 256,
|
| 12 |
+
"contrastive_logits_temperature": 0.1,
|
| 13 |
+
"conv_bias": false,
|
| 14 |
+
"conv_dim": [
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512,
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512
|
| 22 |
+
],
|
| 23 |
+
"conv_kernel": [
|
| 24 |
+
10,
|
| 25 |
+
3,
|
| 26 |
+
3,
|
| 27 |
+
3,
|
| 28 |
+
3,
|
| 29 |
+
2,
|
| 30 |
+
2
|
| 31 |
+
],
|
| 32 |
+
"conv_stride": [
|
| 33 |
+
5,
|
| 34 |
+
2,
|
| 35 |
+
2,
|
| 36 |
+
2,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2
|
| 40 |
+
],
|
| 41 |
+
"ctc_loss_reduction": "mean",
|
| 42 |
+
"ctc_zero_infinity": false,
|
| 43 |
+
"diversity_loss_weight": 0.1,
|
| 44 |
+
"do_stable_layer_norm": false,
|
| 45 |
+
"eos_token_id": 2,
|
| 46 |
+
"feat_extract_activation": "gelu",
|
| 47 |
+
"feat_extract_norm": "group",
|
| 48 |
+
"feat_proj_dropout": 0.1,
|
| 49 |
+
"feat_quantizer_dropout": 0.0,
|
| 50 |
+
"final_dropout": 0.0,
|
| 51 |
+
"freeze_feat_extract_train": true,
|
| 52 |
+
"hidden_act": "gelu",
|
| 53 |
+
"hidden_dropout": 0.1,
|
| 54 |
+
"hidden_size": 768,
|
| 55 |
+
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 3072,
|
| 57 |
+
"layer_norm_eps": 1e-05,
|
| 58 |
+
"layerdrop": 0.0,
|
| 59 |
+
"mask_channel_length": 10,
|
| 60 |
+
"mask_channel_min_space": 1,
|
| 61 |
+
"mask_channel_other": 0.0,
|
| 62 |
+
"mask_channel_prob": 0.0,
|
| 63 |
+
"mask_channel_selection": "static",
|
| 64 |
+
"mask_feature_length": 10,
|
| 65 |
+
"mask_feature_prob": 0.0,
|
| 66 |
+
"mask_time_length": 5,
|
| 67 |
+
"mask_time_min_space": 1,
|
| 68 |
+
"mask_time_other": 0.0,
|
| 69 |
+
"mask_time_prob": 0.05,
|
| 70 |
+
"mask_time_selection": "static",
|
| 71 |
+
"model_type": "wav2vec2",
|
| 72 |
+
"no_mask_channel_overlap": false,
|
| 73 |
+
"no_mask_time_overlap": false,
|
| 74 |
+
"num_attention_heads": 12,
|
| 75 |
+
"num_codevector_groups": 2,
|
| 76 |
+
"num_codevectors_per_group": 320,
|
| 77 |
+
"num_conv_pos_embedding_groups": 16,
|
| 78 |
+
"num_conv_pos_embeddings": 128,
|
| 79 |
+
"num_feat_extract_layers": 7,
|
| 80 |
+
"num_hidden_layers": 12,
|
| 81 |
+
"num_negatives": 100,
|
| 82 |
+
"pad_token_id": 29,
|
| 83 |
+
"proj_codevector_dim": 256,
|
| 84 |
+
"torch_dtype": "float32",
|
| 85 |
+
"transformers_version": "4.11.3",
|
| 86 |
+
"use_weighted_layer_sum": false,
|
| 87 |
+
"vocab_size": 32
|
| 88 |
+
}
|
checkpoint-10000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a0f2e04a7f4017dab4cd20f07f5d5de610962c103ce0008ba25d0d51c2bc7819
|
| 3 |
+
size 721685265
|
checkpoint-10000/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-10000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a6f6554b9888f24e4870f3516b7e9a519c0b9e81213db58cfec3af5c06c91e4
|
| 3 |
+
size 377670039
|
checkpoint-10000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e4e30751787b6fb66ae49da730726c7c3fee10163175dc2a373144256f736f4
|
| 3 |
+
size 17563
|
checkpoint-10000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:771cf8609fdaaed2ab63948ad61b6f20669806667bb75ab8d1a0a1f02c855d8f
|
| 3 |
+
size 559
|
checkpoint-10000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9c95942339e0b222e33ff2982743910a05bb8b9bf2c1b7856cc26cbb05674d4
|
| 3 |
+
size 623
|
checkpoint-10000/trainer_state.json
ADDED
|
@@ -0,0 +1,316 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 27.47252747252747,
|
| 5 |
+
"global_step": 10000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 1.37,
|
| 12 |
+
"learning_rate": 4.99e-05,
|
| 13 |
+
"loss": 3.5823,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.37,
|
| 18 |
+
"eval_loss": 3.32504940032959,
|
| 19 |
+
"eval_runtime": 45.0187,
|
| 20 |
+
"eval_samples_per_second": 14.394,
|
| 21 |
+
"eval_steps_per_second": 0.466,
|
| 22 |
+
"eval_wer": 1.0,
|
| 23 |
+
"step": 500
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 2.75,
|
| 27 |
+
"learning_rate": 9.970000000000001e-05,
|
| 28 |
+
"loss": 1.6148,
|
| 29 |
+
"step": 1000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 2.75,
|
| 33 |
+
"eval_loss": 0.9707893133163452,
|
| 34 |
+
"eval_runtime": 41.6379,
|
| 35 |
+
"eval_samples_per_second": 15.563,
|
| 36 |
+
"eval_steps_per_second": 0.504,
|
| 37 |
+
"eval_wer": 0.5453111956219274,
|
| 38 |
+
"step": 1000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 4.12,
|
| 42 |
+
"learning_rate": 9.50100806451613e-05,
|
| 43 |
+
"loss": 0.7901,
|
| 44 |
+
"step": 1500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 4.12,
|
| 48 |
+
"eval_loss": 0.7987369298934937,
|
| 49 |
+
"eval_runtime": 42.0128,
|
| 50 |
+
"eval_samples_per_second": 15.424,
|
| 51 |
+
"eval_steps_per_second": 0.5,
|
| 52 |
+
"eval_wer": 0.4292737222892125,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 5.49,
|
| 57 |
+
"learning_rate": 8.996975806451613e-05,
|
| 58 |
+
"loss": 0.5529,
|
| 59 |
+
"step": 2000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 5.49,
|
| 63 |
+
"eval_loss": 0.803723156452179,
|
| 64 |
+
"eval_runtime": 41.9975,
|
| 65 |
+
"eval_samples_per_second": 15.43,
|
| 66 |
+
"eval_steps_per_second": 0.5,
|
| 67 |
+
"eval_wer": 0.4209257026249884,
|
| 68 |
+
"step": 2000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 6.87,
|
| 72 |
+
"learning_rate": 8.492943548387097e-05,
|
| 73 |
+
"loss": 0.4404,
|
| 74 |
+
"step": 2500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 6.87,
|
| 78 |
+
"eval_loss": 0.7943486571311951,
|
| 79 |
+
"eval_runtime": 42.0152,
|
| 80 |
+
"eval_samples_per_second": 15.423,
|
| 81 |
+
"eval_steps_per_second": 0.5,
|
| 82 |
+
"eval_wer": 0.45524533902235415,
|
| 83 |
+
"step": 2500
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 8.24,
|
| 87 |
+
"learning_rate": 7.988911290322581e-05,
|
| 88 |
+
"loss": 0.3528,
|
| 89 |
+
"step": 3000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 8.24,
|
| 93 |
+
"eval_loss": 0.8733369708061218,
|
| 94 |
+
"eval_runtime": 42.4523,
|
| 95 |
+
"eval_samples_per_second": 15.264,
|
| 96 |
+
"eval_steps_per_second": 0.495,
|
| 97 |
+
"eval_wer": 0.37658844263055374,
|
| 98 |
+
"step": 3000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 9.62,
|
| 102 |
+
"learning_rate": 7.484879032258065e-05,
|
| 103 |
+
"loss": 0.3008,
|
| 104 |
+
"step": 3500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 9.62,
|
| 108 |
+
"eval_loss": 0.914967954158783,
|
| 109 |
+
"eval_runtime": 42.0647,
|
| 110 |
+
"eval_samples_per_second": 15.405,
|
| 111 |
+
"eval_steps_per_second": 0.499,
|
| 112 |
+
"eval_wer": 0.3725999443465356,
|
| 113 |
+
"step": 3500
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 10.99,
|
| 117 |
+
"learning_rate": 6.98084677419355e-05,
|
| 118 |
+
"loss": 0.2625,
|
| 119 |
+
"step": 4000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 10.99,
|
| 123 |
+
"eval_loss": 0.9753163456916809,
|
| 124 |
+
"eval_runtime": 41.804,
|
| 125 |
+
"eval_samples_per_second": 15.501,
|
| 126 |
+
"eval_steps_per_second": 0.502,
|
| 127 |
+
"eval_wer": 0.3690752249327521,
|
| 128 |
+
"step": 4000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 12.36,
|
| 132 |
+
"learning_rate": 6.476814516129032e-05,
|
| 133 |
+
"loss": 0.2153,
|
| 134 |
+
"step": 4500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 12.36,
|
| 138 |
+
"eval_loss": 0.9292902946472168,
|
| 139 |
+
"eval_runtime": 53.6443,
|
| 140 |
+
"eval_samples_per_second": 12.08,
|
| 141 |
+
"eval_steps_per_second": 0.391,
|
| 142 |
+
"eval_wer": 0.35284296447453856,
|
| 143 |
+
"step": 4500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 13.74,
|
| 147 |
+
"learning_rate": 5.9727822580645166e-05,
|
| 148 |
+
"loss": 0.1926,
|
| 149 |
+
"step": 5000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 13.74,
|
| 153 |
+
"eval_loss": 0.9629846215248108,
|
| 154 |
+
"eval_runtime": 43.3856,
|
| 155 |
+
"eval_samples_per_second": 14.936,
|
| 156 |
+
"eval_steps_per_second": 0.484,
|
| 157 |
+
"eval_wer": 0.3697245153510806,
|
| 158 |
+
"step": 5000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 15.11,
|
| 162 |
+
"learning_rate": 5.46875e-05,
|
| 163 |
+
"loss": 0.1634,
|
| 164 |
+
"step": 5500
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 15.11,
|
| 168 |
+
"eval_loss": 0.9719225764274597,
|
| 169 |
+
"eval_runtime": 43.1527,
|
| 170 |
+
"eval_samples_per_second": 15.016,
|
| 171 |
+
"eval_steps_per_second": 0.487,
|
| 172 |
+
"eval_wer": 0.37473332714961505,
|
| 173 |
+
"step": 5500
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 16.48,
|
| 177 |
+
"learning_rate": 4.964717741935484e-05,
|
| 178 |
+
"loss": 0.1422,
|
| 179 |
+
"step": 6000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 16.48,
|
| 183 |
+
"eval_loss": 1.0191898345947266,
|
| 184 |
+
"eval_runtime": 43.4487,
|
| 185 |
+
"eval_samples_per_second": 14.914,
|
| 186 |
+
"eval_steps_per_second": 0.483,
|
| 187 |
+
"eval_wer": 0.38085520823671276,
|
| 188 |
+
"step": 6000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 17.86,
|
| 192 |
+
"learning_rate": 4.460685483870968e-05,
|
| 193 |
+
"loss": 0.1253,
|
| 194 |
+
"step": 6500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 17.86,
|
| 198 |
+
"eval_loss": 1.0450010299682617,
|
| 199 |
+
"eval_runtime": 42.8089,
|
| 200 |
+
"eval_samples_per_second": 15.137,
|
| 201 |
+
"eval_steps_per_second": 0.491,
|
| 202 |
+
"eval_wer": 0.37686670995269456,
|
| 203 |
+
"step": 6500
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 19.23,
|
| 207 |
+
"learning_rate": 3.956653225806452e-05,
|
| 208 |
+
"loss": 0.1068,
|
| 209 |
+
"step": 7000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 19.23,
|
| 213 |
+
"eval_loss": 1.0906848907470703,
|
| 214 |
+
"eval_runtime": 43.3307,
|
| 215 |
+
"eval_samples_per_second": 14.955,
|
| 216 |
+
"eval_steps_per_second": 0.485,
|
| 217 |
+
"eval_wer": 0.359243112883777,
|
| 218 |
+
"step": 7000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 20.6,
|
| 222 |
+
"learning_rate": 3.4526209677419356e-05,
|
| 223 |
+
"loss": 0.095,
|
| 224 |
+
"step": 7500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 20.6,
|
| 228 |
+
"eval_loss": 1.0443437099456787,
|
| 229 |
+
"eval_runtime": 53.9592,
|
| 230 |
+
"eval_samples_per_second": 12.009,
|
| 231 |
+
"eval_steps_per_second": 0.389,
|
| 232 |
+
"eval_wer": 0.35794453204711996,
|
| 233 |
+
"step": 7500
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 21.98,
|
| 237 |
+
"learning_rate": 2.9495967741935488e-05,
|
| 238 |
+
"loss": 0.0814,
|
| 239 |
+
"step": 8000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 21.98,
|
| 243 |
+
"eval_loss": 1.140635371208191,
|
| 244 |
+
"eval_runtime": 43.1083,
|
| 245 |
+
"eval_samples_per_second": 15.032,
|
| 246 |
+
"eval_steps_per_second": 0.487,
|
| 247 |
+
"eval_wer": 0.36258232074946667,
|
| 248 |
+
"step": 8000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 23.35,
|
| 252 |
+
"learning_rate": 2.4455645161290326e-05,
|
| 253 |
+
"loss": 0.069,
|
| 254 |
+
"step": 8500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 23.35,
|
| 258 |
+
"eval_loss": 1.1894210577011108,
|
| 259 |
+
"eval_runtime": 42.7802,
|
| 260 |
+
"eval_samples_per_second": 15.147,
|
| 261 |
+
"eval_steps_per_second": 0.491,
|
| 262 |
+
"eval_wer": 0.3690752249327521,
|
| 263 |
+
"step": 8500
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 24.73,
|
| 267 |
+
"learning_rate": 1.942540322580645e-05,
|
| 268 |
+
"loss": 0.0623,
|
| 269 |
+
"step": 9000
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 24.73,
|
| 273 |
+
"eval_loss": 1.2040750980377197,
|
| 274 |
+
"eval_runtime": 43.2859,
|
| 275 |
+
"eval_samples_per_second": 14.97,
|
| 276 |
+
"eval_steps_per_second": 0.485,
|
| 277 |
+
"eval_wer": 0.3687042018365643,
|
| 278 |
+
"step": 9000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 26.1,
|
| 282 |
+
"learning_rate": 1.4385080645161292e-05,
|
| 283 |
+
"loss": 0.0554,
|
| 284 |
+
"step": 9500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 26.1,
|
| 288 |
+
"eval_loss": 1.2136509418487549,
|
| 289 |
+
"eval_runtime": 42.5198,
|
| 290 |
+
"eval_samples_per_second": 15.24,
|
| 291 |
+
"eval_steps_per_second": 0.494,
|
| 292 |
+
"eval_wer": 0.35720248585474446,
|
| 293 |
+
"step": 9500
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 27.47,
|
| 297 |
+
"learning_rate": 9.344758064516129e-06,
|
| 298 |
+
"loss": 0.0475,
|
| 299 |
+
"step": 10000
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 27.47,
|
| 303 |
+
"eval_loss": 1.2355363368988037,
|
| 304 |
+
"eval_runtime": 43.577,
|
| 305 |
+
"eval_samples_per_second": 14.87,
|
| 306 |
+
"eval_steps_per_second": 0.482,
|
| 307 |
+
"eval_wer": 0.3565531954364159,
|
| 308 |
+
"step": 10000
|
| 309 |
+
}
|
| 310 |
+
],
|
| 311 |
+
"max_steps": 10920,
|
| 312 |
+
"num_train_epochs": 30,
|
| 313 |
+
"total_flos": 1.4947492785279646e+19,
|
| 314 |
+
"trial_name": null,
|
| 315 |
+
"trial_params": null
|
| 316 |
+
}
|
checkpoint-10000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d998aad278c0bb911fd9ded0353bc04ca5056945240341f90e545e377ba43f2
|
| 3 |
+
size 2863
|
checkpoint-10500/config.json
ADDED
|
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"apply_spec_augment": true,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Wav2Vec2ForCTC"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.1,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"classifier_proj_size": 256,
|
| 11 |
+
"codevector_dim": 256,
|
| 12 |
+
"contrastive_logits_temperature": 0.1,
|
| 13 |
+
"conv_bias": false,
|
| 14 |
+
"conv_dim": [
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512,
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512
|
| 22 |
+
],
|
| 23 |
+
"conv_kernel": [
|
| 24 |
+
10,
|
| 25 |
+
3,
|
| 26 |
+
3,
|
| 27 |
+
3,
|
| 28 |
+
3,
|
| 29 |
+
2,
|
| 30 |
+
2
|
| 31 |
+
],
|
| 32 |
+
"conv_stride": [
|
| 33 |
+
5,
|
| 34 |
+
2,
|
| 35 |
+
2,
|
| 36 |
+
2,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2
|
| 40 |
+
],
|
| 41 |
+
"ctc_loss_reduction": "mean",
|
| 42 |
+
"ctc_zero_infinity": false,
|
| 43 |
+
"diversity_loss_weight": 0.1,
|
| 44 |
+
"do_stable_layer_norm": false,
|
| 45 |
+
"eos_token_id": 2,
|
| 46 |
+
"feat_extract_activation": "gelu",
|
| 47 |
+
"feat_extract_norm": "group",
|
| 48 |
+
"feat_proj_dropout": 0.1,
|
| 49 |
+
"feat_quantizer_dropout": 0.0,
|
| 50 |
+
"final_dropout": 0.0,
|
| 51 |
+
"freeze_feat_extract_train": true,
|
| 52 |
+
"hidden_act": "gelu",
|
| 53 |
+
"hidden_dropout": 0.1,
|
| 54 |
+
"hidden_size": 768,
|
| 55 |
+
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 3072,
|
| 57 |
+
"layer_norm_eps": 1e-05,
|
| 58 |
+
"layerdrop": 0.0,
|
| 59 |
+
"mask_channel_length": 10,
|
| 60 |
+
"mask_channel_min_space": 1,
|
| 61 |
+
"mask_channel_other": 0.0,
|
| 62 |
+
"mask_channel_prob": 0.0,
|
| 63 |
+
"mask_channel_selection": "static",
|
| 64 |
+
"mask_feature_length": 10,
|
| 65 |
+
"mask_feature_prob": 0.0,
|
| 66 |
+
"mask_time_length": 5,
|
| 67 |
+
"mask_time_min_space": 1,
|
| 68 |
+
"mask_time_other": 0.0,
|
| 69 |
+
"mask_time_prob": 0.05,
|
| 70 |
+
"mask_time_selection": "static",
|
| 71 |
+
"model_type": "wav2vec2",
|
| 72 |
+
"no_mask_channel_overlap": false,
|
| 73 |
+
"no_mask_time_overlap": false,
|
| 74 |
+
"num_attention_heads": 12,
|
| 75 |
+
"num_codevector_groups": 2,
|
| 76 |
+
"num_codevectors_per_group": 320,
|
| 77 |
+
"num_conv_pos_embedding_groups": 16,
|
| 78 |
+
"num_conv_pos_embeddings": 128,
|
| 79 |
+
"num_feat_extract_layers": 7,
|
| 80 |
+
"num_hidden_layers": 12,
|
| 81 |
+
"num_negatives": 100,
|
| 82 |
+
"pad_token_id": 29,
|
| 83 |
+
"proj_codevector_dim": 256,
|
| 84 |
+
"torch_dtype": "float32",
|
| 85 |
+
"transformers_version": "4.11.3",
|
| 86 |
+
"use_weighted_layer_sum": false,
|
| 87 |
+
"vocab_size": 32
|
| 88 |
+
}
|
checkpoint-10500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ea4f76fcc9f6d9ea40150bb4e01d304d0772c7fba4c1aa7fea14b1725df1414e
|
| 3 |
+
size 721685265
|
checkpoint-10500/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-10500/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d42de1bc77469fe30684b2e5c05fcafab3828fa74d7190534b6fbea66f10844
|
| 3 |
+
size 377670039
|
checkpoint-10500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8c18378be313c39a465357862e4f74412d52436f287b6d03fa899b9d973d72e2
|
| 3 |
+
size 17563
|
checkpoint-10500/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:36accd578a3753adf01ee13fa036622b2454b71534c2490afa37778ef7236892
|
| 3 |
+
size 559
|
checkpoint-10500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7c2140cd816daecbe941131ce3e89734069c9b4974308433698907fb82469575
|
| 3 |
+
size 623
|
checkpoint-10500/trainer_state.json
ADDED
|
@@ -0,0 +1,331 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 28.846153846153847,
|
| 5 |
+
"global_step": 10500,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 1.37,
|
| 12 |
+
"learning_rate": 4.99e-05,
|
| 13 |
+
"loss": 3.5823,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.37,
|
| 18 |
+
"eval_loss": 3.32504940032959,
|
| 19 |
+
"eval_runtime": 45.0187,
|
| 20 |
+
"eval_samples_per_second": 14.394,
|
| 21 |
+
"eval_steps_per_second": 0.466,
|
| 22 |
+
"eval_wer": 1.0,
|
| 23 |
+
"step": 500
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 2.75,
|
| 27 |
+
"learning_rate": 9.970000000000001e-05,
|
| 28 |
+
"loss": 1.6148,
|
| 29 |
+
"step": 1000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 2.75,
|
| 33 |
+
"eval_loss": 0.9707893133163452,
|
| 34 |
+
"eval_runtime": 41.6379,
|
| 35 |
+
"eval_samples_per_second": 15.563,
|
| 36 |
+
"eval_steps_per_second": 0.504,
|
| 37 |
+
"eval_wer": 0.5453111956219274,
|
| 38 |
+
"step": 1000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 4.12,
|
| 42 |
+
"learning_rate": 9.50100806451613e-05,
|
| 43 |
+
"loss": 0.7901,
|
| 44 |
+
"step": 1500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 4.12,
|
| 48 |
+
"eval_loss": 0.7987369298934937,
|
| 49 |
+
"eval_runtime": 42.0128,
|
| 50 |
+
"eval_samples_per_second": 15.424,
|
| 51 |
+
"eval_steps_per_second": 0.5,
|
| 52 |
+
"eval_wer": 0.4292737222892125,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 5.49,
|
| 57 |
+
"learning_rate": 8.996975806451613e-05,
|
| 58 |
+
"loss": 0.5529,
|
| 59 |
+
"step": 2000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 5.49,
|
| 63 |
+
"eval_loss": 0.803723156452179,
|
| 64 |
+
"eval_runtime": 41.9975,
|
| 65 |
+
"eval_samples_per_second": 15.43,
|
| 66 |
+
"eval_steps_per_second": 0.5,
|
| 67 |
+
"eval_wer": 0.4209257026249884,
|
| 68 |
+
"step": 2000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 6.87,
|
| 72 |
+
"learning_rate": 8.492943548387097e-05,
|
| 73 |
+
"loss": 0.4404,
|
| 74 |
+
"step": 2500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 6.87,
|
| 78 |
+
"eval_loss": 0.7943486571311951,
|
| 79 |
+
"eval_runtime": 42.0152,
|
| 80 |
+
"eval_samples_per_second": 15.423,
|
| 81 |
+
"eval_steps_per_second": 0.5,
|
| 82 |
+
"eval_wer": 0.45524533902235415,
|
| 83 |
+
"step": 2500
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 8.24,
|
| 87 |
+
"learning_rate": 7.988911290322581e-05,
|
| 88 |
+
"loss": 0.3528,
|
| 89 |
+
"step": 3000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 8.24,
|
| 93 |
+
"eval_loss": 0.8733369708061218,
|
| 94 |
+
"eval_runtime": 42.4523,
|
| 95 |
+
"eval_samples_per_second": 15.264,
|
| 96 |
+
"eval_steps_per_second": 0.495,
|
| 97 |
+
"eval_wer": 0.37658844263055374,
|
| 98 |
+
"step": 3000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 9.62,
|
| 102 |
+
"learning_rate": 7.484879032258065e-05,
|
| 103 |
+
"loss": 0.3008,
|
| 104 |
+
"step": 3500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 9.62,
|
| 108 |
+
"eval_loss": 0.914967954158783,
|
| 109 |
+
"eval_runtime": 42.0647,
|
| 110 |
+
"eval_samples_per_second": 15.405,
|
| 111 |
+
"eval_steps_per_second": 0.499,
|
| 112 |
+
"eval_wer": 0.3725999443465356,
|
| 113 |
+
"step": 3500
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 10.99,
|
| 117 |
+
"learning_rate": 6.98084677419355e-05,
|
| 118 |
+
"loss": 0.2625,
|
| 119 |
+
"step": 4000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 10.99,
|
| 123 |
+
"eval_loss": 0.9753163456916809,
|
| 124 |
+
"eval_runtime": 41.804,
|
| 125 |
+
"eval_samples_per_second": 15.501,
|
| 126 |
+
"eval_steps_per_second": 0.502,
|
| 127 |
+
"eval_wer": 0.3690752249327521,
|
| 128 |
+
"step": 4000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 12.36,
|
| 132 |
+
"learning_rate": 6.476814516129032e-05,
|
| 133 |
+
"loss": 0.2153,
|
| 134 |
+
"step": 4500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 12.36,
|
| 138 |
+
"eval_loss": 0.9292902946472168,
|
| 139 |
+
"eval_runtime": 53.6443,
|
| 140 |
+
"eval_samples_per_second": 12.08,
|
| 141 |
+
"eval_steps_per_second": 0.391,
|
| 142 |
+
"eval_wer": 0.35284296447453856,
|
| 143 |
+
"step": 4500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 13.74,
|
| 147 |
+
"learning_rate": 5.9727822580645166e-05,
|
| 148 |
+
"loss": 0.1926,
|
| 149 |
+
"step": 5000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 13.74,
|
| 153 |
+
"eval_loss": 0.9629846215248108,
|
| 154 |
+
"eval_runtime": 43.3856,
|
| 155 |
+
"eval_samples_per_second": 14.936,
|
| 156 |
+
"eval_steps_per_second": 0.484,
|
| 157 |
+
"eval_wer": 0.3697245153510806,
|
| 158 |
+
"step": 5000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 15.11,
|
| 162 |
+
"learning_rate": 5.46875e-05,
|
| 163 |
+
"loss": 0.1634,
|
| 164 |
+
"step": 5500
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 15.11,
|
| 168 |
+
"eval_loss": 0.9719225764274597,
|
| 169 |
+
"eval_runtime": 43.1527,
|
| 170 |
+
"eval_samples_per_second": 15.016,
|
| 171 |
+
"eval_steps_per_second": 0.487,
|
| 172 |
+
"eval_wer": 0.37473332714961505,
|
| 173 |
+
"step": 5500
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 16.48,
|
| 177 |
+
"learning_rate": 4.964717741935484e-05,
|
| 178 |
+
"loss": 0.1422,
|
| 179 |
+
"step": 6000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 16.48,
|
| 183 |
+
"eval_loss": 1.0191898345947266,
|
| 184 |
+
"eval_runtime": 43.4487,
|
| 185 |
+
"eval_samples_per_second": 14.914,
|
| 186 |
+
"eval_steps_per_second": 0.483,
|
| 187 |
+
"eval_wer": 0.38085520823671276,
|
| 188 |
+
"step": 6000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 17.86,
|
| 192 |
+
"learning_rate": 4.460685483870968e-05,
|
| 193 |
+
"loss": 0.1253,
|
| 194 |
+
"step": 6500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 17.86,
|
| 198 |
+
"eval_loss": 1.0450010299682617,
|
| 199 |
+
"eval_runtime": 42.8089,
|
| 200 |
+
"eval_samples_per_second": 15.137,
|
| 201 |
+
"eval_steps_per_second": 0.491,
|
| 202 |
+
"eval_wer": 0.37686670995269456,
|
| 203 |
+
"step": 6500
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 19.23,
|
| 207 |
+
"learning_rate": 3.956653225806452e-05,
|
| 208 |
+
"loss": 0.1068,
|
| 209 |
+
"step": 7000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 19.23,
|
| 213 |
+
"eval_loss": 1.0906848907470703,
|
| 214 |
+
"eval_runtime": 43.3307,
|
| 215 |
+
"eval_samples_per_second": 14.955,
|
| 216 |
+
"eval_steps_per_second": 0.485,
|
| 217 |
+
"eval_wer": 0.359243112883777,
|
| 218 |
+
"step": 7000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 20.6,
|
| 222 |
+
"learning_rate": 3.4526209677419356e-05,
|
| 223 |
+
"loss": 0.095,
|
| 224 |
+
"step": 7500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 20.6,
|
| 228 |
+
"eval_loss": 1.0443437099456787,
|
| 229 |
+
"eval_runtime": 53.9592,
|
| 230 |
+
"eval_samples_per_second": 12.009,
|
| 231 |
+
"eval_steps_per_second": 0.389,
|
| 232 |
+
"eval_wer": 0.35794453204711996,
|
| 233 |
+
"step": 7500
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 21.98,
|
| 237 |
+
"learning_rate": 2.9495967741935488e-05,
|
| 238 |
+
"loss": 0.0814,
|
| 239 |
+
"step": 8000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 21.98,
|
| 243 |
+
"eval_loss": 1.140635371208191,
|
| 244 |
+
"eval_runtime": 43.1083,
|
| 245 |
+
"eval_samples_per_second": 15.032,
|
| 246 |
+
"eval_steps_per_second": 0.487,
|
| 247 |
+
"eval_wer": 0.36258232074946667,
|
| 248 |
+
"step": 8000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 23.35,
|
| 252 |
+
"learning_rate": 2.4455645161290326e-05,
|
| 253 |
+
"loss": 0.069,
|
| 254 |
+
"step": 8500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 23.35,
|
| 258 |
+
"eval_loss": 1.1894210577011108,
|
| 259 |
+
"eval_runtime": 42.7802,
|
| 260 |
+
"eval_samples_per_second": 15.147,
|
| 261 |
+
"eval_steps_per_second": 0.491,
|
| 262 |
+
"eval_wer": 0.3690752249327521,
|
| 263 |
+
"step": 8500
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 24.73,
|
| 267 |
+
"learning_rate": 1.942540322580645e-05,
|
| 268 |
+
"loss": 0.0623,
|
| 269 |
+
"step": 9000
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 24.73,
|
| 273 |
+
"eval_loss": 1.2040750980377197,
|
| 274 |
+
"eval_runtime": 43.2859,
|
| 275 |
+
"eval_samples_per_second": 14.97,
|
| 276 |
+
"eval_steps_per_second": 0.485,
|
| 277 |
+
"eval_wer": 0.3687042018365643,
|
| 278 |
+
"step": 9000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 26.1,
|
| 282 |
+
"learning_rate": 1.4385080645161292e-05,
|
| 283 |
+
"loss": 0.0554,
|
| 284 |
+
"step": 9500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 26.1,
|
| 288 |
+
"eval_loss": 1.2136509418487549,
|
| 289 |
+
"eval_runtime": 42.5198,
|
| 290 |
+
"eval_samples_per_second": 15.24,
|
| 291 |
+
"eval_steps_per_second": 0.494,
|
| 292 |
+
"eval_wer": 0.35720248585474446,
|
| 293 |
+
"step": 9500
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 27.47,
|
| 297 |
+
"learning_rate": 9.344758064516129e-06,
|
| 298 |
+
"loss": 0.0475,
|
| 299 |
+
"step": 10000
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 27.47,
|
| 303 |
+
"eval_loss": 1.2355363368988037,
|
| 304 |
+
"eval_runtime": 43.577,
|
| 305 |
+
"eval_samples_per_second": 14.87,
|
| 306 |
+
"eval_steps_per_second": 0.482,
|
| 307 |
+
"eval_wer": 0.3565531954364159,
|
| 308 |
+
"step": 10000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 28.85,
|
| 312 |
+
"learning_rate": 4.304435483870968e-06,
|
| 313 |
+
"loss": 0.042,
|
| 314 |
+
"step": 10500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 28.85,
|
| 318 |
+
"eval_loss": 1.2368338108062744,
|
| 319 |
+
"eval_runtime": 43.2519,
|
| 320 |
+
"eval_samples_per_second": 14.982,
|
| 321 |
+
"eval_steps_per_second": 0.486,
|
| 322 |
+
"eval_wer": 0.35553288192189964,
|
| 323 |
+
"step": 10500
|
| 324 |
+
}
|
| 325 |
+
],
|
| 326 |
+
"max_steps": 10920,
|
| 327 |
+
"num_train_epochs": 30,
|
| 328 |
+
"total_flos": 1.5691152295536976e+19,
|
| 329 |
+
"trial_name": null,
|
| 330 |
+
"trial_params": null
|
| 331 |
+
}
|
checkpoint-10500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d998aad278c0bb911fd9ded0353bc04ca5056945240341f90e545e377ba43f2
|
| 3 |
+
size 2863
|
config.json
ADDED
|
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"apply_spec_augment": true,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Wav2Vec2ForCTC"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.1,
|
| 9 |
+
"bos_token_id": 1,
|
| 10 |
+
"classifier_proj_size": 256,
|
| 11 |
+
"codevector_dim": 256,
|
| 12 |
+
"contrastive_logits_temperature": 0.1,
|
| 13 |
+
"conv_bias": false,
|
| 14 |
+
"conv_dim": [
|
| 15 |
+
512,
|
| 16 |
+
512,
|
| 17 |
+
512,
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512
|
| 22 |
+
],
|
| 23 |
+
"conv_kernel": [
|
| 24 |
+
10,
|
| 25 |
+
3,
|
| 26 |
+
3,
|
| 27 |
+
3,
|
| 28 |
+
3,
|
| 29 |
+
2,
|
| 30 |
+
2
|
| 31 |
+
],
|
| 32 |
+
"conv_stride": [
|
| 33 |
+
5,
|
| 34 |
+
2,
|
| 35 |
+
2,
|
| 36 |
+
2,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2
|
| 40 |
+
],
|
| 41 |
+
"ctc_loss_reduction": "mean",
|
| 42 |
+
"ctc_zero_infinity": false,
|
| 43 |
+
"diversity_loss_weight": 0.1,
|
| 44 |
+
"do_stable_layer_norm": false,
|
| 45 |
+
"eos_token_id": 2,
|
| 46 |
+
"feat_extract_activation": "gelu",
|
| 47 |
+
"feat_extract_norm": "group",
|
| 48 |
+
"feat_proj_dropout": 0.1,
|
| 49 |
+
"feat_quantizer_dropout": 0.0,
|
| 50 |
+
"final_dropout": 0.0,
|
| 51 |
+
"freeze_feat_extract_train": true,
|
| 52 |
+
"hidden_act": "gelu",
|
| 53 |
+
"hidden_dropout": 0.1,
|
| 54 |
+
"hidden_size": 768,
|
| 55 |
+
"initializer_range": 0.02,
|
| 56 |
+
"intermediate_size": 3072,
|
| 57 |
+
"layer_norm_eps": 1e-05,
|
| 58 |
+
"layerdrop": 0.0,
|
| 59 |
+
"mask_channel_length": 10,
|
| 60 |
+
"mask_channel_min_space": 1,
|
| 61 |
+
"mask_channel_other": 0.0,
|
| 62 |
+
"mask_channel_prob": 0.0,
|
| 63 |
+
"mask_channel_selection": "static",
|
| 64 |
+
"mask_feature_length": 10,
|
| 65 |
+
"mask_feature_prob": 0.0,
|
| 66 |
+
"mask_time_length": 5,
|
| 67 |
+
"mask_time_min_space": 1,
|
| 68 |
+
"mask_time_other": 0.0,
|
| 69 |
+
"mask_time_prob": 0.05,
|
| 70 |
+
"mask_time_selection": "static",
|
| 71 |
+
"model_type": "wav2vec2",
|
| 72 |
+
"no_mask_channel_overlap": false,
|
| 73 |
+
"no_mask_time_overlap": false,
|
| 74 |
+
"num_attention_heads": 12,
|
| 75 |
+
"num_codevector_groups": 2,
|
| 76 |
+
"num_codevectors_per_group": 320,
|
| 77 |
+
"num_conv_pos_embedding_groups": 16,
|
| 78 |
+
"num_conv_pos_embeddings": 128,
|
| 79 |
+
"num_feat_extract_layers": 7,
|
| 80 |
+
"num_hidden_layers": 12,
|
| 81 |
+
"num_negatives": 100,
|
| 82 |
+
"pad_token_id": 29,
|
| 83 |
+
"proj_codevector_dim": 256,
|
| 84 |
+
"torch_dtype": "float32",
|
| 85 |
+
"transformers_version": "4.11.3",
|
| 86 |
+
"use_weighted_layer_sum": false,
|
| 87 |
+
"vocab_size": 32
|
| 88 |
+
}
|
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 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7f4e1617defbcf85b67d55294f8177b259e3daf55269592cf44d09962844275
|
| 3 |
+
size 377670039
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d998aad278c0bb911fd9ded0353bc04ca5056945240341f90e545e377ba43f2
|
| 3 |
+
size 2863
|