Upload 9 files
Browse files- config.json +82 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- preprocessor_config.json +11 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +401 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/w2v-bert-2.0",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"adapter_act": "relu",
|
| 5 |
+
"adapter_kernel_size": 3,
|
| 6 |
+
"adapter_stride": 2,
|
| 7 |
+
"add_adapter": true,
|
| 8 |
+
"apply_spec_augment": false,
|
| 9 |
+
"architectures": [
|
| 10 |
+
"Wav2Vec2BertForCTC"
|
| 11 |
+
],
|
| 12 |
+
"attention_dropout": 0.1,
|
| 13 |
+
"bos_token_id": 1,
|
| 14 |
+
"classifier_proj_size": 768,
|
| 15 |
+
"codevector_dim": 768,
|
| 16 |
+
"conformer_conv_dropout": 0.1,
|
| 17 |
+
"contrastive_logits_temperature": 0.1,
|
| 18 |
+
"conv_depthwise_kernel_size": 31,
|
| 19 |
+
"ctc_loss_reduction": "mean",
|
| 20 |
+
"ctc_zero_infinity": true,
|
| 21 |
+
"diversity_loss_weight": 0.1,
|
| 22 |
+
"eos_token_id": 2,
|
| 23 |
+
"feat_proj_dropout": 0.0,
|
| 24 |
+
"feat_quantizer_dropout": 0.0,
|
| 25 |
+
"feature_projection_input_dim": 160,
|
| 26 |
+
"final_dropout": 0.1,
|
| 27 |
+
"hidden_act": "swish",
|
| 28 |
+
"hidden_dropout": 0.1,
|
| 29 |
+
"hidden_size": 1024,
|
| 30 |
+
"initializer_range": 0.02,
|
| 31 |
+
"intermediate_size": 4096,
|
| 32 |
+
"layer_norm_eps": 1e-05,
|
| 33 |
+
"layerdrop": 0.1,
|
| 34 |
+
"left_max_position_embeddings": 64,
|
| 35 |
+
"mask_feature_length": 10,
|
| 36 |
+
"mask_feature_min_masks": 0,
|
| 37 |
+
"mask_feature_prob": 0.0,
|
| 38 |
+
"mask_time_length": 10,
|
| 39 |
+
"mask_time_min_masks": 2,
|
| 40 |
+
"mask_time_prob": 0.05,
|
| 41 |
+
"max_source_positions": 5000,
|
| 42 |
+
"model_type": "wav2vec2-bert",
|
| 43 |
+
"num_adapter_layers": 1,
|
| 44 |
+
"num_attention_heads": 16,
|
| 45 |
+
"num_codevector_groups": 2,
|
| 46 |
+
"num_codevectors_per_group": 320,
|
| 47 |
+
"num_hidden_layers": 24,
|
| 48 |
+
"num_negatives": 100,
|
| 49 |
+
"output_hidden_size": 1024,
|
| 50 |
+
"pad_token_id": 10,
|
| 51 |
+
"position_embeddings_type": "relative_key",
|
| 52 |
+
"proj_codevector_dim": 768,
|
| 53 |
+
"right_max_position_embeddings": 8,
|
| 54 |
+
"rotary_embedding_base": 10000,
|
| 55 |
+
"tdnn_dilation": [
|
| 56 |
+
1,
|
| 57 |
+
2,
|
| 58 |
+
3,
|
| 59 |
+
1,
|
| 60 |
+
1
|
| 61 |
+
],
|
| 62 |
+
"tdnn_dim": [
|
| 63 |
+
512,
|
| 64 |
+
512,
|
| 65 |
+
512,
|
| 66 |
+
512,
|
| 67 |
+
1500
|
| 68 |
+
],
|
| 69 |
+
"tdnn_kernel": [
|
| 70 |
+
5,
|
| 71 |
+
3,
|
| 72 |
+
3,
|
| 73 |
+
1,
|
| 74 |
+
1
|
| 75 |
+
],
|
| 76 |
+
"torch_dtype": "float32",
|
| 77 |
+
"transformers_version": "4.49.0",
|
| 78 |
+
"use_intermediate_ffn_before_adapter": false,
|
| 79 |
+
"use_weighted_layer_sum": false,
|
| 80 |
+
"vocab_size": 121,
|
| 81 |
+
"xvector_output_dim": 512
|
| 82 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fe3f69ea60011292e84ab93b103a62e1fce63b30f017aff547a4bdab44ad8597
|
| 3 |
+
size 2423314860
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27cfd8f3eb4f93d3708b58b75ac0aab026a11983d182ce8e71348c3029cc50c5
|
| 3 |
+
size 4847082090
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"feature_extractor_type": "SeamlessM4TFeatureExtractor",
|
| 3 |
+
"feature_size": 80,
|
| 4 |
+
"num_mel_bins": 80,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 1,
|
| 7 |
+
"processor_class": "Wav2Vec2BertProcessor",
|
| 8 |
+
"return_attention_mask": true,
|
| 9 |
+
"sampling_rate": 16000,
|
| 10 |
+
"stride": 2
|
| 11 |
+
}
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6776dffe13e7ea552ae69e8260f4c497e5ee6ac9a084534a89cacd6b0a0bdf13
|
| 3 |
+
size 14244
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:31ee2633ecfe9c976e676fd6a2873f9f50962a60410ec4ea8bca55113c97dd18
|
| 3 |
+
size 988
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee5fa6b4c0f6eef770da13676691943657574e184c6e6df83c6c1388b333cab2
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,401 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.2302779592132956,
|
| 3 |
+
"best_model_checkpoint": "Bengali_wav2vec2_bert_punct\\checkpoint-230000",
|
| 4 |
+
"epoch": 4.0,
|
| 5 |
+
"eval_steps": 10000,
|
| 6 |
+
"global_step": 232060,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.17236921485822632,
|
| 13 |
+
"grad_norm": 4.759167671203613,
|
| 14 |
+
"learning_rate": 4.79499913629297e-05,
|
| 15 |
+
"loss": 1.4067,
|
| 16 |
+
"step": 10000
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.17236921485822632,
|
| 20 |
+
"eval_WER": 0.40140793961825966,
|
| 21 |
+
"eval_loss": 0.3006719648838043,
|
| 22 |
+
"eval_runtime": 76.2909,
|
| 23 |
+
"eval_samples_per_second": 39.323,
|
| 24 |
+
"eval_steps_per_second": 4.915,
|
| 25 |
+
"step": 10000
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.34473842971645263,
|
| 29 |
+
"grad_norm": 3.7281529903411865,
|
| 30 |
+
"learning_rate": 4.5791371566764555e-05,
|
| 31 |
+
"loss": 1.0498,
|
| 32 |
+
"step": 20000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.34473842971645263,
|
| 36 |
+
"eval_WER": 0.35401698236446766,
|
| 37 |
+
"eval_loss": 0.24728824198246002,
|
| 38 |
+
"eval_runtime": 78.6228,
|
| 39 |
+
"eval_samples_per_second": 38.157,
|
| 40 |
+
"eval_steps_per_second": 4.77,
|
| 41 |
+
"step": 20000
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.517107644574679,
|
| 45 |
+
"grad_norm": 3.5825536251068115,
|
| 46 |
+
"learning_rate": 4.363253584384177e-05,
|
| 47 |
+
"loss": 0.9196,
|
| 48 |
+
"step": 30000
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 0.517107644574679,
|
| 52 |
+
"eval_WER": 0.3367080339647289,
|
| 53 |
+
"eval_loss": 0.22603251039981842,
|
| 54 |
+
"eval_runtime": 77.3702,
|
| 55 |
+
"eval_samples_per_second": 38.775,
|
| 56 |
+
"eval_steps_per_second": 4.847,
|
| 57 |
+
"step": 30000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.6894768594329053,
|
| 61 |
+
"grad_norm": 5.719975471496582,
|
| 62 |
+
"learning_rate": 4.147434790119192e-05,
|
| 63 |
+
"loss": 0.8347,
|
| 64 |
+
"step": 40000
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 0.6894768594329053,
|
| 68 |
+
"eval_WER": 0.32026997605051166,
|
| 69 |
+
"eval_loss": 0.1984129399061203,
|
| 70 |
+
"eval_runtime": 76.0573,
|
| 71 |
+
"eval_samples_per_second": 39.444,
|
| 72 |
+
"eval_steps_per_second": 4.93,
|
| 73 |
+
"step": 40000
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.8618460742911316,
|
| 77 |
+
"grad_norm": 12.707213401794434,
|
| 78 |
+
"learning_rate": 3.931594403178442e-05,
|
| 79 |
+
"loss": 0.7728,
|
| 80 |
+
"step": 50000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.8618460742911316,
|
| 84 |
+
"eval_WER": 0.30205385006168806,
|
| 85 |
+
"eval_loss": 0.18471553921699524,
|
| 86 |
+
"eval_runtime": 77.2641,
|
| 87 |
+
"eval_samples_per_second": 38.828,
|
| 88 |
+
"eval_steps_per_second": 4.853,
|
| 89 |
+
"step": 50000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.034215289149358,
|
| 93 |
+
"grad_norm": 2.484391689300537,
|
| 94 |
+
"learning_rate": 3.715732423561928e-05,
|
| 95 |
+
"loss": 0.7257,
|
| 96 |
+
"step": 60000
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.034215289149358,
|
| 100 |
+
"eval_WER": 0.29552217142027726,
|
| 101 |
+
"eval_loss": 0.17463411390781403,
|
| 102 |
+
"eval_runtime": 77.8642,
|
| 103 |
+
"eval_samples_per_second": 38.529,
|
| 104 |
+
"eval_steps_per_second": 4.816,
|
| 105 |
+
"step": 60000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.2065845040075842,
|
| 109 |
+
"grad_norm": 2.728205680847168,
|
| 110 |
+
"learning_rate": 3.4999136292969425e-05,
|
| 111 |
+
"loss": 0.6895,
|
| 112 |
+
"step": 70000
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.2065845040075842,
|
| 116 |
+
"eval_WER": 0.2906596995427825,
|
| 117 |
+
"eval_loss": 0.16489063203334808,
|
| 118 |
+
"eval_runtime": 85.1263,
|
| 119 |
+
"eval_samples_per_second": 35.242,
|
| 120 |
+
"eval_steps_per_second": 4.405,
|
| 121 |
+
"step": 70000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 1.3789537188658105,
|
| 125 |
+
"grad_norm": 3.63918399810791,
|
| 126 |
+
"learning_rate": 3.284030057004664e-05,
|
| 127 |
+
"loss": 0.6653,
|
| 128 |
+
"step": 80000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.3789537188658105,
|
| 132 |
+
"eval_WER": 0.28917192829668337,
|
| 133 |
+
"eval_loss": 0.16142985224723816,
|
| 134 |
+
"eval_runtime": 77.5124,
|
| 135 |
+
"eval_samples_per_second": 38.704,
|
| 136 |
+
"eval_steps_per_second": 4.838,
|
| 137 |
+
"step": 80000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.5513229337240368,
|
| 141 |
+
"grad_norm": 3.3453214168548584,
|
| 142 |
+
"learning_rate": 3.06816807738815e-05,
|
| 143 |
+
"loss": 0.6418,
|
| 144 |
+
"step": 90000
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 1.5513229337240368,
|
| 148 |
+
"eval_WER": 0.2765440162566224,
|
| 149 |
+
"eval_loss": 0.1598576009273529,
|
| 150 |
+
"eval_runtime": 77.4372,
|
| 151 |
+
"eval_samples_per_second": 38.741,
|
| 152 |
+
"eval_steps_per_second": 4.843,
|
| 153 |
+
"step": 90000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 1.7236921485822632,
|
| 157 |
+
"grad_norm": 3.1832072734832764,
|
| 158 |
+
"learning_rate": 2.8523492831231646e-05,
|
| 159 |
+
"loss": 0.6241,
|
| 160 |
+
"step": 100000
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 1.7236921485822632,
|
| 164 |
+
"eval_WER": 0.2766165904637492,
|
| 165 |
+
"eval_loss": 0.14717546105384827,
|
| 166 |
+
"eval_runtime": 77.4687,
|
| 167 |
+
"eval_samples_per_second": 38.725,
|
| 168 |
+
"eval_steps_per_second": 4.841,
|
| 169 |
+
"step": 100000
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 1.8960613634404895,
|
| 173 |
+
"grad_norm": 4.113869667053223,
|
| 174 |
+
"learning_rate": 2.6364873035066505e-05,
|
| 175 |
+
"loss": 0.6092,
|
| 176 |
+
"step": 110000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.8960613634404895,
|
| 180 |
+
"eval_WER": 0.26765367588359096,
|
| 181 |
+
"eval_loss": 0.1441964954137802,
|
| 182 |
+
"eval_runtime": 77.6139,
|
| 183 |
+
"eval_samples_per_second": 38.653,
|
| 184 |
+
"eval_steps_per_second": 4.832,
|
| 185 |
+
"step": 110000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.068430578298716,
|
| 189 |
+
"grad_norm": 1.85727858543396,
|
| 190 |
+
"learning_rate": 2.4206253238901367e-05,
|
| 191 |
+
"loss": 0.5805,
|
| 192 |
+
"step": 120000
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.068430578298716,
|
| 196 |
+
"eval_WER": 0.26322664924885697,
|
| 197 |
+
"eval_loss": 0.13760210573673248,
|
| 198 |
+
"eval_runtime": 77.692,
|
| 199 |
+
"eval_samples_per_second": 38.614,
|
| 200 |
+
"eval_steps_per_second": 4.827,
|
| 201 |
+
"step": 120000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 2.2407997931569423,
|
| 205 |
+
"grad_norm": 2.774275779724121,
|
| 206 |
+
"learning_rate": 2.204741751597858e-05,
|
| 207 |
+
"loss": 0.5611,
|
| 208 |
+
"step": 130000
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 2.2407997931569423,
|
| 212 |
+
"eval_WER": 0.2587633355105595,
|
| 213 |
+
"eval_loss": 0.1326003223657608,
|
| 214 |
+
"eval_runtime": 75.6029,
|
| 215 |
+
"eval_samples_per_second": 39.681,
|
| 216 |
+
"eval_steps_per_second": 4.96,
|
| 217 |
+
"step": 130000
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 2.4131690080151684,
|
| 221 |
+
"grad_norm": 1.8835299015045166,
|
| 222 |
+
"learning_rate": 1.9889013646571085e-05,
|
| 223 |
+
"loss": 0.5433,
|
| 224 |
+
"step": 140000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.4131690080151684,
|
| 228 |
+
"eval_WER": 0.2550983380506568,
|
| 229 |
+
"eval_loss": 0.12959806621074677,
|
| 230 |
+
"eval_runtime": 75.9556,
|
| 231 |
+
"eval_samples_per_second": 39.497,
|
| 232 |
+
"eval_steps_per_second": 4.937,
|
| 233 |
+
"step": 140000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 2.585538222873395,
|
| 237 |
+
"grad_norm": 2.7739031314849854,
|
| 238 |
+
"learning_rate": 1.773060977716359e-05,
|
| 239 |
+
"loss": 0.5315,
|
| 240 |
+
"step": 150000
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 2.585538222873395,
|
| 244 |
+
"eval_WER": 0.2523405181798389,
|
| 245 |
+
"eval_loss": 0.12333784997463226,
|
| 246 |
+
"eval_runtime": 75.6489,
|
| 247 |
+
"eval_samples_per_second": 39.657,
|
| 248 |
+
"eval_steps_per_second": 4.957,
|
| 249 |
+
"step": 150000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 2.757907437731621,
|
| 253 |
+
"grad_norm": 5.59724235534668,
|
| 254 |
+
"learning_rate": 1.5572421834513734e-05,
|
| 255 |
+
"loss": 0.5192,
|
| 256 |
+
"step": 160000
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 2.757907437731621,
|
| 260 |
+
"eval_WER": 0.2465708687132593,
|
| 261 |
+
"eval_loss": 0.11873666197061539,
|
| 262 |
+
"eval_runtime": 75.813,
|
| 263 |
+
"eval_samples_per_second": 39.571,
|
| 264 |
+
"eval_steps_per_second": 4.946,
|
| 265 |
+
"step": 160000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.9302766525898476,
|
| 269 |
+
"grad_norm": 1.7511285543441772,
|
| 270 |
+
"learning_rate": 1.3413802038348594e-05,
|
| 271 |
+
"loss": 0.508,
|
| 272 |
+
"step": 170000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.9302766525898476,
|
| 276 |
+
"eval_WER": 0.24355903911749763,
|
| 277 |
+
"eval_loss": 0.11491911858320236,
|
| 278 |
+
"eval_runtime": 76.2555,
|
| 279 |
+
"eval_samples_per_second": 39.341,
|
| 280 |
+
"eval_steps_per_second": 4.918,
|
| 281 |
+
"step": 170000
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.1026458674480737,
|
| 285 |
+
"grad_norm": 1.3240278959274292,
|
| 286 |
+
"learning_rate": 1.1255830022456384e-05,
|
| 287 |
+
"loss": 0.4885,
|
| 288 |
+
"step": 180000
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 3.1026458674480737,
|
| 292 |
+
"eval_WER": 0.24315988097830032,
|
| 293 |
+
"eval_loss": 0.11166874319314957,
|
| 294 |
+
"eval_runtime": 77.1138,
|
| 295 |
+
"eval_samples_per_second": 38.904,
|
| 296 |
+
"eval_steps_per_second": 4.863,
|
| 297 |
+
"step": 180000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.2750150823063002,
|
| 301 |
+
"grad_norm": 1.0281517505645752,
|
| 302 |
+
"learning_rate": 9.096994299533598e-06,
|
| 303 |
+
"loss": 0.4709,
|
| 304 |
+
"step": 190000
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.2750150823063002,
|
| 308 |
+
"eval_WER": 0.23887800275781987,
|
| 309 |
+
"eval_loss": 0.10879674553871155,
|
| 310 |
+
"eval_runtime": 77.1588,
|
| 311 |
+
"eval_samples_per_second": 38.881,
|
| 312 |
+
"eval_steps_per_second": 4.86,
|
| 313 |
+
"step": 190000
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 3.4473842971645263,
|
| 317 |
+
"grad_norm": 2.2434046268463135,
|
| 318 |
+
"learning_rate": 6.938590430126101e-06,
|
| 319 |
+
"loss": 0.4669,
|
| 320 |
+
"step": 200000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 3.4473842971645263,
|
| 324 |
+
"eval_WER": 0.23492270846940996,
|
| 325 |
+
"eval_loss": 0.10604960471391678,
|
| 326 |
+
"eval_runtime": 77.0841,
|
| 327 |
+
"eval_samples_per_second": 38.919,
|
| 328 |
+
"eval_steps_per_second": 4.865,
|
| 329 |
+
"step": 200000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 3.619753512022753,
|
| 333 |
+
"grad_norm": 0.7335111498832703,
|
| 334 |
+
"learning_rate": 4.780186560718605e-06,
|
| 335 |
+
"loss": 0.451,
|
| 336 |
+
"step": 210000
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 3.619753512022753,
|
| 340 |
+
"eval_WER": 0.23452355033021263,
|
| 341 |
+
"eval_loss": 0.10134341567754745,
|
| 342 |
+
"eval_runtime": 77.1143,
|
| 343 |
+
"eval_samples_per_second": 38.903,
|
| 344 |
+
"eval_steps_per_second": 4.863,
|
| 345 |
+
"step": 210000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 3.792122726880979,
|
| 349 |
+
"grad_norm": 1.7208735942840576,
|
| 350 |
+
"learning_rate": 2.6215667645534633e-06,
|
| 351 |
+
"loss": 0.445,
|
| 352 |
+
"step": 220000
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 3.792122726880979,
|
| 356 |
+
"eval_WER": 0.2329632048769867,
|
| 357 |
+
"eval_loss": 0.10342755168676376,
|
| 358 |
+
"eval_runtime": 77.7151,
|
| 359 |
+
"eval_samples_per_second": 38.603,
|
| 360 |
+
"eval_steps_per_second": 4.825,
|
| 361 |
+
"step": 220000
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 3.9644919417392055,
|
| 365 |
+
"grad_norm": 1.0953819751739502,
|
| 366 |
+
"learning_rate": 4.6316289514596653e-07,
|
| 367 |
+
"loss": 0.4379,
|
| 368 |
+
"step": 230000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 3.9644919417392055,
|
| 372 |
+
"eval_WER": 0.2302779592132956,
|
| 373 |
+
"eval_loss": 0.09944533556699753,
|
| 374 |
+
"eval_runtime": 78.1453,
|
| 375 |
+
"eval_samples_per_second": 38.39,
|
| 376 |
+
"eval_steps_per_second": 4.799,
|
| 377 |
+
"step": 230000
|
| 378 |
+
}
|
| 379 |
+
],
|
| 380 |
+
"logging_steps": 10000,
|
| 381 |
+
"max_steps": 232060,
|
| 382 |
+
"num_input_tokens_seen": 0,
|
| 383 |
+
"num_train_epochs": 4,
|
| 384 |
+
"save_steps": 10000,
|
| 385 |
+
"stateful_callbacks": {
|
| 386 |
+
"TrainerControl": {
|
| 387 |
+
"args": {
|
| 388 |
+
"should_epoch_stop": false,
|
| 389 |
+
"should_evaluate": false,
|
| 390 |
+
"should_log": false,
|
| 391 |
+
"should_save": true,
|
| 392 |
+
"should_training_stop": true
|
| 393 |
+
},
|
| 394 |
+
"attributes": {}
|
| 395 |
+
}
|
| 396 |
+
},
|
| 397 |
+
"total_flos": 4.7334394582539475e+20,
|
| 398 |
+
"train_batch_size": 8,
|
| 399 |
+
"trial_name": null,
|
| 400 |
+
"trial_params": null
|
| 401 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:26dffa834fdd7c5b630dd22f4c0b5efd73ad1ae918e2a4b5a2e02f9475aa5335
|
| 3 |
+
size 5304
|