Upload 9 files
Browse files- config.json +117 -0
- optimizer.pt +3 -0
- preprocessor_config.json +10 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1621 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,117 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/wav2vec2-base",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForCTC"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.1,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 256,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": false,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "mean",
|
| 45 |
+
"ctc_zero_infinity": false,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": false,
|
| 48 |
+
"eos_token_id": 2,
|
| 49 |
+
"feat_extract_activation": "gelu",
|
| 50 |
+
"feat_extract_norm": "group",
|
| 51 |
+
"feat_proj_dropout": 0.1,
|
| 52 |
+
"feat_quantizer_dropout": 0.0,
|
| 53 |
+
"final_dropout": 0.0,
|
| 54 |
+
"freeze_feat_extract_train": true,
|
| 55 |
+
"hidden_act": "gelu",
|
| 56 |
+
"hidden_dropout": 0.1,
|
| 57 |
+
"hidden_size": 768,
|
| 58 |
+
"initializer_range": 0.02,
|
| 59 |
+
"intermediate_size": 3072,
|
| 60 |
+
"layer_norm_eps": 1e-05,
|
| 61 |
+
"layerdrop": 0.0,
|
| 62 |
+
"mask_channel_length": 10,
|
| 63 |
+
"mask_channel_min_space": 1,
|
| 64 |
+
"mask_channel_other": 0.0,
|
| 65 |
+
"mask_channel_prob": 0.0,
|
| 66 |
+
"mask_channel_selection": "static",
|
| 67 |
+
"mask_feature_length": 10,
|
| 68 |
+
"mask_feature_min_masks": 0,
|
| 69 |
+
"mask_feature_prob": 0.0,
|
| 70 |
+
"mask_time_length": 10,
|
| 71 |
+
"mask_time_min_masks": 2,
|
| 72 |
+
"mask_time_min_space": 1,
|
| 73 |
+
"mask_time_other": 0.0,
|
| 74 |
+
"mask_time_prob": 0.05,
|
| 75 |
+
"mask_time_selection": "static",
|
| 76 |
+
"model_type": "wav2vec2",
|
| 77 |
+
"no_mask_channel_overlap": false,
|
| 78 |
+
"no_mask_time_overlap": false,
|
| 79 |
+
"num_adapter_layers": 3,
|
| 80 |
+
"num_attention_heads": 12,
|
| 81 |
+
"num_codevector_groups": 2,
|
| 82 |
+
"num_codevectors_per_group": 320,
|
| 83 |
+
"num_conv_pos_embedding_groups": 16,
|
| 84 |
+
"num_conv_pos_embeddings": 128,
|
| 85 |
+
"num_feat_extract_layers": 7,
|
| 86 |
+
"num_hidden_layers": 12,
|
| 87 |
+
"num_negatives": 100,
|
| 88 |
+
"output_hidden_size": 768,
|
| 89 |
+
"pad_token_id": 28,
|
| 90 |
+
"proj_codevector_dim": 256,
|
| 91 |
+
"tdnn_dilation": [
|
| 92 |
+
1,
|
| 93 |
+
2,
|
| 94 |
+
3,
|
| 95 |
+
1,
|
| 96 |
+
1
|
| 97 |
+
],
|
| 98 |
+
"tdnn_dim": [
|
| 99 |
+
512,
|
| 100 |
+
512,
|
| 101 |
+
512,
|
| 102 |
+
512,
|
| 103 |
+
1500
|
| 104 |
+
],
|
| 105 |
+
"tdnn_kernel": [
|
| 106 |
+
5,
|
| 107 |
+
3,
|
| 108 |
+
3,
|
| 109 |
+
1,
|
| 110 |
+
1
|
| 111 |
+
],
|
| 112 |
+
"torch_dtype": "float32",
|
| 113 |
+
"transformers_version": "4.17.0",
|
| 114 |
+
"use_weighted_layer_sum": false,
|
| 115 |
+
"vocab_size": 32,
|
| 116 |
+
"xvector_output_dim": 512
|
| 117 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c256927a27da68b67120eeac851a76d0db02662b13ad4196ecf4f84c05f4cfab
|
| 3 |
+
size 29822559
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"processor_class": "Wav2Vec2Processor",
|
| 8 |
+
"return_attention_mask": false,
|
| 9 |
+
"sampling_rate": 16000
|
| 10 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:56f3248ca1580ac5c75f8a0c5c5e616d46d62fdc46e7dae73cacf92d5a7b26e9
|
| 3 |
+
size 29980831
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7284e7f19df473c3a179ec4662bc6ad2aa5b3f879ac682f6858983a95d109e02
|
| 3 |
+
size 14639
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e0bf33d9dd70901e017176dd7f4d53da792e70c44225c094c290b98be218cca
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ba6395ec13a8a096493a777284e644acdccdbe49def73c60b9ff333880e3ea8
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1621 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 29.98878923766816,
|
| 5 |
+
"global_step": 53500,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.28,
|
| 12 |
+
"learning_rate": 4.92e-05,
|
| 13 |
+
"loss": 3.6567,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.28,
|
| 18 |
+
"eval_loss": 2.2588322162628174,
|
| 19 |
+
"eval_runtime": 172.2499,
|
| 20 |
+
"eval_samples_per_second": 15.21,
|
| 21 |
+
"eval_steps_per_second": 1.904,
|
| 22 |
+
"eval_wer": 1.006447808886184,
|
| 23 |
+
"step": 500
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.56,
|
| 27 |
+
"learning_rate": 9.92e-05,
|
| 28 |
+
"loss": 0.5607,
|
| 29 |
+
"step": 1000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.56,
|
| 33 |
+
"eval_loss": 0.9839454293251038,
|
| 34 |
+
"eval_runtime": 172.8549,
|
| 35 |
+
"eval_samples_per_second": 15.157,
|
| 36 |
+
"eval_steps_per_second": 1.898,
|
| 37 |
+
"eval_wer": 0.829142574558734,
|
| 38 |
+
"step": 1000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.84,
|
| 42 |
+
"learning_rate": 9.906321401370907e-05,
|
| 43 |
+
"loss": 0.2159,
|
| 44 |
+
"step": 1500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.84,
|
| 48 |
+
"eval_loss": 0.720267653465271,
|
| 49 |
+
"eval_runtime": 172.1686,
|
| 50 |
+
"eval_samples_per_second": 15.218,
|
| 51 |
+
"eval_steps_per_second": 1.905,
|
| 52 |
+
"eval_wer": 0.8293137553256239,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 1.12,
|
| 57 |
+
"learning_rate": 9.811119573495812e-05,
|
| 58 |
+
"loss": 0.167,
|
| 59 |
+
"step": 2000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.12,
|
| 63 |
+
"eval_loss": 0.2308063805103302,
|
| 64 |
+
"eval_runtime": 171.8565,
|
| 65 |
+
"eval_samples_per_second": 15.245,
|
| 66 |
+
"eval_steps_per_second": 1.909,
|
| 67 |
+
"eval_wer": 0.8286670724284845,
|
| 68 |
+
"step": 2000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.4,
|
| 72 |
+
"learning_rate": 9.715917745620717e-05,
|
| 73 |
+
"loss": 0.1381,
|
| 74 |
+
"step": 2500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.4,
|
| 78 |
+
"eval_loss": 0.1806459128856659,
|
| 79 |
+
"eval_runtime": 173.2886,
|
| 80 |
+
"eval_samples_per_second": 15.119,
|
| 81 |
+
"eval_steps_per_second": 1.893,
|
| 82 |
+
"eval_wer": 0.8606017954960439,
|
| 83 |
+
"step": 2500
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.68,
|
| 87 |
+
"learning_rate": 9.62071591774562e-05,
|
| 88 |
+
"loss": 0.1313,
|
| 89 |
+
"step": 3000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.68,
|
| 93 |
+
"eval_loss": 0.20655015110969543,
|
| 94 |
+
"eval_runtime": 173.1734,
|
| 95 |
+
"eval_samples_per_second": 15.129,
|
| 96 |
+
"eval_steps_per_second": 1.894,
|
| 97 |
+
"eval_wer": 0.8578819233110164,
|
| 98 |
+
"step": 3000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.96,
|
| 102 |
+
"learning_rate": 9.525514089870527e-05,
|
| 103 |
+
"loss": 0.125,
|
| 104 |
+
"step": 3500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.96,
|
| 108 |
+
"eval_loss": 0.17935717105865479,
|
| 109 |
+
"eval_runtime": 172.7498,
|
| 110 |
+
"eval_samples_per_second": 15.166,
|
| 111 |
+
"eval_steps_per_second": 1.899,
|
| 112 |
+
"eval_wer": 0.8602213937918441,
|
| 113 |
+
"step": 3500
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 2.24,
|
| 117 |
+
"learning_rate": 9.43031226199543e-05,
|
| 118 |
+
"loss": 0.1044,
|
| 119 |
+
"step": 4000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.24,
|
| 123 |
+
"eval_loss": 0.1682695746421814,
|
| 124 |
+
"eval_runtime": 172.1701,
|
| 125 |
+
"eval_samples_per_second": 15.218,
|
| 126 |
+
"eval_steps_per_second": 1.905,
|
| 127 |
+
"eval_wer": 0.8638352099817407,
|
| 128 |
+
"step": 4000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 2.52,
|
| 132 |
+
"learning_rate": 9.335110434120336e-05,
|
| 133 |
+
"loss": 0.1017,
|
| 134 |
+
"step": 4500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 2.52,
|
| 138 |
+
"eval_loss": 0.16941602528095245,
|
| 139 |
+
"eval_runtime": 173.2001,
|
| 140 |
+
"eval_samples_per_second": 15.127,
|
| 141 |
+
"eval_steps_per_second": 1.894,
|
| 142 |
+
"eval_wer": 0.8537735849056604,
|
| 143 |
+
"step": 4500
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 2.8,
|
| 147 |
+
"learning_rate": 9.23990860624524e-05,
|
| 148 |
+
"loss": 0.0991,
|
| 149 |
+
"step": 5000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 2.8,
|
| 153 |
+
"eval_loss": 0.31116464734077454,
|
| 154 |
+
"eval_runtime": 174.2757,
|
| 155 |
+
"eval_samples_per_second": 15.034,
|
| 156 |
+
"eval_steps_per_second": 1.882,
|
| 157 |
+
"eval_wer": 0.8972344796104686,
|
| 158 |
+
"step": 5000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 3.08,
|
| 162 |
+
"learning_rate": 9.144706778370146e-05,
|
| 163 |
+
"loss": 0.0944,
|
| 164 |
+
"step": 5500
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 3.08,
|
| 168 |
+
"eval_loss": 0.15802101790905,
|
| 169 |
+
"eval_runtime": 173.1104,
|
| 170 |
+
"eval_samples_per_second": 15.135,
|
| 171 |
+
"eval_steps_per_second": 1.895,
|
| 172 |
+
"eval_wer": 0.8589850882531954,
|
| 173 |
+
"step": 5500
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 3.36,
|
| 177 |
+
"learning_rate": 9.04950495049505e-05,
|
| 178 |
+
"loss": 0.0839,
|
| 179 |
+
"step": 6000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.36,
|
| 183 |
+
"eval_loss": 0.14672406017780304,
|
| 184 |
+
"eval_runtime": 172.4692,
|
| 185 |
+
"eval_samples_per_second": 15.191,
|
| 186 |
+
"eval_steps_per_second": 1.902,
|
| 187 |
+
"eval_wer": 0.8793746195982958,
|
| 188 |
+
"step": 6000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.64,
|
| 192 |
+
"learning_rate": 8.954303122619954e-05,
|
| 193 |
+
"loss": 0.0847,
|
| 194 |
+
"step": 6500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.64,
|
| 198 |
+
"eval_loss": 0.1362634152173996,
|
| 199 |
+
"eval_runtime": 172.397,
|
| 200 |
+
"eval_samples_per_second": 15.197,
|
| 201 |
+
"eval_steps_per_second": 1.903,
|
| 202 |
+
"eval_wer": 0.8613055386488131,
|
| 203 |
+
"step": 6500
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 3.92,
|
| 207 |
+
"learning_rate": 8.85910129474486e-05,
|
| 208 |
+
"loss": 0.0839,
|
| 209 |
+
"step": 7000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 3.92,
|
| 213 |
+
"eval_loss": 0.13779252767562866,
|
| 214 |
+
"eval_runtime": 173.1232,
|
| 215 |
+
"eval_samples_per_second": 15.134,
|
| 216 |
+
"eval_steps_per_second": 1.895,
|
| 217 |
+
"eval_wer": 0.8749809799147901,
|
| 218 |
+
"step": 7000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 4.2,
|
| 222 |
+
"learning_rate": 8.763899466869764e-05,
|
| 223 |
+
"loss": 0.0756,
|
| 224 |
+
"step": 7500
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.2,
|
| 228 |
+
"eval_loss": 0.1471327692270279,
|
| 229 |
+
"eval_runtime": 173.0172,
|
| 230 |
+
"eval_samples_per_second": 15.143,
|
| 231 |
+
"eval_steps_per_second": 1.896,
|
| 232 |
+
"eval_wer": 0.857330340839927,
|
| 233 |
+
"step": 7500
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 4.48,
|
| 237 |
+
"learning_rate": 8.66869763899467e-05,
|
| 238 |
+
"loss": 0.0726,
|
| 239 |
+
"step": 8000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.48,
|
| 243 |
+
"eval_loss": 0.14172284305095673,
|
| 244 |
+
"eval_runtime": 175.2167,
|
| 245 |
+
"eval_samples_per_second": 14.953,
|
| 246 |
+
"eval_steps_per_second": 1.872,
|
| 247 |
+
"eval_wer": 0.8881428788800974,
|
| 248 |
+
"step": 8000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 4.76,
|
| 252 |
+
"learning_rate": 8.573495811119574e-05,
|
| 253 |
+
"loss": 0.0733,
|
| 254 |
+
"step": 8500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 4.76,
|
| 258 |
+
"eval_loss": 0.14988631010055542,
|
| 259 |
+
"eval_runtime": 174.4611,
|
| 260 |
+
"eval_samples_per_second": 15.018,
|
| 261 |
+
"eval_steps_per_second": 1.88,
|
| 262 |
+
"eval_wer": 0.8689896530736457,
|
| 263 |
+
"step": 8500
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 5.04,
|
| 267 |
+
"learning_rate": 8.478293983244478e-05,
|
| 268 |
+
"loss": 0.0711,
|
| 269 |
+
"step": 9000
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 5.04,
|
| 273 |
+
"eval_loss": 0.2800207734107971,
|
| 274 |
+
"eval_runtime": 173.1977,
|
| 275 |
+
"eval_samples_per_second": 15.127,
|
| 276 |
+
"eval_steps_per_second": 1.894,
|
| 277 |
+
"eval_wer": 0.8613245587340231,
|
| 278 |
+
"step": 9000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.33,
|
| 282 |
+
"learning_rate": 8.383092155369383e-05,
|
| 283 |
+
"loss": 0.0661,
|
| 284 |
+
"step": 9500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 5.33,
|
| 288 |
+
"eval_loss": 0.1751021295785904,
|
| 289 |
+
"eval_runtime": 170.8086,
|
| 290 |
+
"eval_samples_per_second": 15.339,
|
| 291 |
+
"eval_steps_per_second": 1.92,
|
| 292 |
+
"eval_wer": 0.8714812842361533,
|
| 293 |
+
"step": 9500
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 5.61,
|
| 297 |
+
"learning_rate": 8.287890327494288e-05,
|
| 298 |
+
"loss": 0.0651,
|
| 299 |
+
"step": 10000
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.61,
|
| 303 |
+
"eval_loss": 0.15104886889457703,
|
| 304 |
+
"eval_runtime": 170.1342,
|
| 305 |
+
"eval_samples_per_second": 15.4,
|
| 306 |
+
"eval_steps_per_second": 1.928,
|
| 307 |
+
"eval_wer": 0.8766927875836884,
|
| 308 |
+
"step": 10000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 5.89,
|
| 312 |
+
"learning_rate": 8.192878903274944e-05,
|
| 313 |
+
"loss": 0.0662,
|
| 314 |
+
"step": 10500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 5.89,
|
| 318 |
+
"eval_loss": 0.13780978322029114,
|
| 319 |
+
"eval_runtime": 170.7672,
|
| 320 |
+
"eval_samples_per_second": 15.343,
|
| 321 |
+
"eval_steps_per_second": 1.921,
|
| 322 |
+
"eval_wer": 0.8781763542300669,
|
| 323 |
+
"step": 10500
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 6.17,
|
| 327 |
+
"learning_rate": 8.097677075399847e-05,
|
| 328 |
+
"loss": 0.0601,
|
| 329 |
+
"step": 11000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 6.17,
|
| 333 |
+
"eval_loss": 0.15003788471221924,
|
| 334 |
+
"eval_runtime": 170.8745,
|
| 335 |
+
"eval_samples_per_second": 15.333,
|
| 336 |
+
"eval_steps_per_second": 1.92,
|
| 337 |
+
"eval_wer": 0.852099817407182,
|
| 338 |
+
"step": 11000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.45,
|
| 342 |
+
"learning_rate": 8.002475247524754e-05,
|
| 343 |
+
"loss": 0.0599,
|
| 344 |
+
"step": 11500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 6.45,
|
| 348 |
+
"eval_loss": 0.15895423293113708,
|
| 349 |
+
"eval_runtime": 171.3813,
|
| 350 |
+
"eval_samples_per_second": 15.288,
|
| 351 |
+
"eval_steps_per_second": 1.914,
|
| 352 |
+
"eval_wer": 0.8785947961046866,
|
| 353 |
+
"step": 11500
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 6.73,
|
| 357 |
+
"learning_rate": 7.907273419649657e-05,
|
| 358 |
+
"loss": 0.0592,
|
| 359 |
+
"step": 12000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.73,
|
| 363 |
+
"eval_loss": 0.1461658626794815,
|
| 364 |
+
"eval_runtime": 170.9782,
|
| 365 |
+
"eval_samples_per_second": 15.324,
|
| 366 |
+
"eval_steps_per_second": 1.918,
|
| 367 |
+
"eval_wer": 0.8844910225197808,
|
| 368 |
+
"step": 12000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 7.01,
|
| 372 |
+
"learning_rate": 7.812071591774562e-05,
|
| 373 |
+
"loss": 0.0605,
|
| 374 |
+
"step": 12500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 7.01,
|
| 378 |
+
"eval_loss": 0.1626746654510498,
|
| 379 |
+
"eval_runtime": 171.4984,
|
| 380 |
+
"eval_samples_per_second": 15.277,
|
| 381 |
+
"eval_steps_per_second": 1.913,
|
| 382 |
+
"eval_wer": 0.8526513998782714,
|
| 383 |
+
"step": 12500
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 7.29,
|
| 387 |
+
"learning_rate": 7.716869763899467e-05,
|
| 388 |
+
"loss": 0.0522,
|
| 389 |
+
"step": 13000
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 7.29,
|
| 393 |
+
"eval_loss": 0.13948819041252136,
|
| 394 |
+
"eval_runtime": 172.0496,
|
| 395 |
+
"eval_samples_per_second": 15.228,
|
| 396 |
+
"eval_steps_per_second": 1.906,
|
| 397 |
+
"eval_wer": 0.8781002738892271,
|
| 398 |
+
"step": 13000
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.57,
|
| 402 |
+
"learning_rate": 7.621667936024372e-05,
|
| 403 |
+
"loss": 0.0545,
|
| 404 |
+
"step": 13500
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.57,
|
| 408 |
+
"eval_loss": 0.17240145802497864,
|
| 409 |
+
"eval_runtime": 172.0023,
|
| 410 |
+
"eval_samples_per_second": 15.232,
|
| 411 |
+
"eval_steps_per_second": 1.907,
|
| 412 |
+
"eval_wer": 0.8662127206329885,
|
| 413 |
+
"step": 13500
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 7.85,
|
| 417 |
+
"learning_rate": 7.526466108149277e-05,
|
| 418 |
+
"loss": 0.0557,
|
| 419 |
+
"step": 14000
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.85,
|
| 423 |
+
"eval_loss": 0.15961869060993195,
|
| 424 |
+
"eval_runtime": 172.3758,
|
| 425 |
+
"eval_samples_per_second": 15.199,
|
| 426 |
+
"eval_steps_per_second": 1.903,
|
| 427 |
+
"eval_wer": 0.8705873402312843,
|
| 428 |
+
"step": 14000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 8.13,
|
| 432 |
+
"learning_rate": 7.431264280274181e-05,
|
| 433 |
+
"loss": 0.0532,
|
| 434 |
+
"step": 14500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 8.13,
|
| 438 |
+
"eval_loss": 0.2017827332019806,
|
| 439 |
+
"eval_runtime": 172.5156,
|
| 440 |
+
"eval_samples_per_second": 15.187,
|
| 441 |
+
"eval_steps_per_second": 1.901,
|
| 442 |
+
"eval_wer": 0.8924604382227632,
|
| 443 |
+
"step": 14500
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 8.41,
|
| 447 |
+
"learning_rate": 7.336062452399086e-05,
|
| 448 |
+
"loss": 0.0495,
|
| 449 |
+
"step": 15000
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 8.41,
|
| 453 |
+
"eval_loss": 0.16683323681354523,
|
| 454 |
+
"eval_runtime": 170.6381,
|
| 455 |
+
"eval_samples_per_second": 15.354,
|
| 456 |
+
"eval_steps_per_second": 1.922,
|
| 457 |
+
"eval_wer": 0.8496462264150944,
|
| 458 |
+
"step": 15000
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 8.69,
|
| 462 |
+
"learning_rate": 7.241051028179741e-05,
|
| 463 |
+
"loss": 0.0496,
|
| 464 |
+
"step": 15500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 8.69,
|
| 468 |
+
"eval_loss": 0.1456364542245865,
|
| 469 |
+
"eval_runtime": 171.0658,
|
| 470 |
+
"eval_samples_per_second": 15.316,
|
| 471 |
+
"eval_steps_per_second": 1.917,
|
| 472 |
+
"eval_wer": 0.8635499087035909,
|
| 473 |
+
"step": 15500
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 8.97,
|
| 477 |
+
"learning_rate": 7.145849200304646e-05,
|
| 478 |
+
"loss": 0.0503,
|
| 479 |
+
"step": 16000
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 8.97,
|
| 483 |
+
"eval_loss": 0.14933466911315918,
|
| 484 |
+
"eval_runtime": 171.7717,
|
| 485 |
+
"eval_samples_per_second": 15.253,
|
| 486 |
+
"eval_steps_per_second": 1.91,
|
| 487 |
+
"eval_wer": 0.887572276323798,
|
| 488 |
+
"step": 16000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 9.25,
|
| 492 |
+
"learning_rate": 7.05064737242955e-05,
|
| 493 |
+
"loss": 0.0454,
|
| 494 |
+
"step": 16500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 9.25,
|
| 498 |
+
"eval_loss": 0.16774588823318481,
|
| 499 |
+
"eval_runtime": 172.0824,
|
| 500 |
+
"eval_samples_per_second": 15.225,
|
| 501 |
+
"eval_steps_per_second": 1.906,
|
| 502 |
+
"eval_wer": 0.8762553256238588,
|
| 503 |
+
"step": 16500
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 9.53,
|
| 507 |
+
"learning_rate": 6.955445544554455e-05,
|
| 508 |
+
"loss": 0.046,
|
| 509 |
+
"step": 17000
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 9.53,
|
| 513 |
+
"eval_loss": 0.15530605614185333,
|
| 514 |
+
"eval_runtime": 172.6241,
|
| 515 |
+
"eval_samples_per_second": 15.177,
|
| 516 |
+
"eval_steps_per_second": 1.9,
|
| 517 |
+
"eval_wer": 0.8600311929397444,
|
| 518 |
+
"step": 17000
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 9.81,
|
| 522 |
+
"learning_rate": 6.860243716679361e-05,
|
| 523 |
+
"loss": 0.0459,
|
| 524 |
+
"step": 17500
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 9.81,
|
| 528 |
+
"eval_loss": 0.16235464811325073,
|
| 529 |
+
"eval_runtime": 172.8904,
|
| 530 |
+
"eval_samples_per_second": 15.154,
|
| 531 |
+
"eval_steps_per_second": 1.897,
|
| 532 |
+
"eval_wer": 0.8851567255021302,
|
| 533 |
+
"step": 17500
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 10.09,
|
| 537 |
+
"learning_rate": 6.765041888804265e-05,
|
| 538 |
+
"loss": 0.0449,
|
| 539 |
+
"step": 18000
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 10.09,
|
| 543 |
+
"eval_loss": 0.17912352085113525,
|
| 544 |
+
"eval_runtime": 172.0805,
|
| 545 |
+
"eval_samples_per_second": 15.225,
|
| 546 |
+
"eval_steps_per_second": 1.906,
|
| 547 |
+
"eval_wer": 0.847915398660986,
|
| 548 |
+
"step": 18000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 10.37,
|
| 552 |
+
"learning_rate": 6.66984006092917e-05,
|
| 553 |
+
"loss": 0.0415,
|
| 554 |
+
"step": 18500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 10.37,
|
| 558 |
+
"eval_loss": 0.1881452053785324,
|
| 559 |
+
"eval_runtime": 172.0219,
|
| 560 |
+
"eval_samples_per_second": 15.231,
|
| 561 |
+
"eval_steps_per_second": 1.907,
|
| 562 |
+
"eval_wer": 0.8846051430310408,
|
| 563 |
+
"step": 18500
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 10.65,
|
| 567 |
+
"learning_rate": 6.574638233054075e-05,
|
| 568 |
+
"loss": 0.0416,
|
| 569 |
+
"step": 19000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 10.65,
|
| 573 |
+
"eval_loss": 0.17328748106956482,
|
| 574 |
+
"eval_runtime": 172.4725,
|
| 575 |
+
"eval_samples_per_second": 15.191,
|
| 576 |
+
"eval_steps_per_second": 1.902,
|
| 577 |
+
"eval_wer": 0.8776057516737675,
|
| 578 |
+
"step": 19000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.93,
|
| 582 |
+
"learning_rate": 6.47943640517898e-05,
|
| 583 |
+
"loss": 0.0431,
|
| 584 |
+
"step": 19500
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 10.93,
|
| 588 |
+
"eval_loss": 0.18211907148361206,
|
| 589 |
+
"eval_runtime": 172.9146,
|
| 590 |
+
"eval_samples_per_second": 15.152,
|
| 591 |
+
"eval_steps_per_second": 1.897,
|
| 592 |
+
"eval_wer": 0.8507874315276932,
|
| 593 |
+
"step": 19500
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 11.21,
|
| 597 |
+
"learning_rate": 6.384424980959634e-05,
|
| 598 |
+
"loss": 0.0406,
|
| 599 |
+
"step": 20000
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 11.21,
|
| 603 |
+
"eval_loss": 0.18406778573989868,
|
| 604 |
+
"eval_runtime": 173.76,
|
| 605 |
+
"eval_samples_per_second": 15.078,
|
| 606 |
+
"eval_steps_per_second": 1.888,
|
| 607 |
+
"eval_wer": 0.8517384357881923,
|
| 608 |
+
"step": 20000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 11.49,
|
| 612 |
+
"learning_rate": 6.28941355674029e-05,
|
| 613 |
+
"loss": 0.0397,
|
| 614 |
+
"step": 20500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 11.49,
|
| 618 |
+
"eval_loss": 0.20063810050487518,
|
| 619 |
+
"eval_runtime": 172.4336,
|
| 620 |
+
"eval_samples_per_second": 15.194,
|
| 621 |
+
"eval_steps_per_second": 1.902,
|
| 622 |
+
"eval_wer": 0.872850730371272,
|
| 623 |
+
"step": 20500
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 11.77,
|
| 627 |
+
"learning_rate": 6.194402132520945e-05,
|
| 628 |
+
"loss": 0.0398,
|
| 629 |
+
"step": 21000
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 11.77,
|
| 633 |
+
"eval_loss": 0.1665649265050888,
|
| 634 |
+
"eval_runtime": 170.9877,
|
| 635 |
+
"eval_samples_per_second": 15.323,
|
| 636 |
+
"eval_steps_per_second": 1.918,
|
| 637 |
+
"eval_wer": 0.8569309190505173,
|
| 638 |
+
"step": 21000
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 12.05,
|
| 642 |
+
"learning_rate": 6.099200304645849e-05,
|
| 643 |
+
"loss": 0.0407,
|
| 644 |
+
"step": 21500
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 12.05,
|
| 648 |
+
"eval_loss": 0.18828719854354858,
|
| 649 |
+
"eval_runtime": 172.0882,
|
| 650 |
+
"eval_samples_per_second": 15.225,
|
| 651 |
+
"eval_steps_per_second": 1.906,
|
| 652 |
+
"eval_wer": 0.8598600121728546,
|
| 653 |
+
"step": 21500
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 12.33,
|
| 657 |
+
"learning_rate": 6.003998476770754e-05,
|
| 658 |
+
"loss": 0.037,
|
| 659 |
+
"step": 22000
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 12.33,
|
| 663 |
+
"eval_loss": 0.16707730293273926,
|
| 664 |
+
"eval_runtime": 172.8578,
|
| 665 |
+
"eval_samples_per_second": 15.157,
|
| 666 |
+
"eval_steps_per_second": 1.898,
|
| 667 |
+
"eval_wer": 0.8588139074863055,
|
| 668 |
+
"step": 22000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 12.61,
|
| 672 |
+
"learning_rate": 5.908987052551409e-05,
|
| 673 |
+
"loss": 0.0375,
|
| 674 |
+
"step": 22500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 12.61,
|
| 678 |
+
"eval_loss": 0.17921198904514313,
|
| 679 |
+
"eval_runtime": 173.3898,
|
| 680 |
+
"eval_samples_per_second": 15.11,
|
| 681 |
+
"eval_steps_per_second": 1.892,
|
| 682 |
+
"eval_wer": 0.8597458916615947,
|
| 683 |
+
"step": 22500
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 12.89,
|
| 687 |
+
"learning_rate": 5.8137852246763134e-05,
|
| 688 |
+
"loss": 0.037,
|
| 689 |
+
"step": 23000
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 12.89,
|
| 693 |
+
"eval_loss": 0.18186964094638824,
|
| 694 |
+
"eval_runtime": 173.9272,
|
| 695 |
+
"eval_samples_per_second": 15.064,
|
| 696 |
+
"eval_steps_per_second": 1.886,
|
| 697 |
+
"eval_wer": 0.8751141205112599,
|
| 698 |
+
"step": 23000
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 13.17,
|
| 702 |
+
"learning_rate": 5.718583396801219e-05,
|
| 703 |
+
"loss": 0.0367,
|
| 704 |
+
"step": 23500
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 13.17,
|
| 708 |
+
"eval_loss": 0.22593411803245544,
|
| 709 |
+
"eval_runtime": 173.0314,
|
| 710 |
+
"eval_samples_per_second": 15.142,
|
| 711 |
+
"eval_steps_per_second": 1.896,
|
| 712 |
+
"eval_wer": 0.8539828058429701,
|
| 713 |
+
"step": 23500
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 13.45,
|
| 717 |
+
"learning_rate": 5.623381568926124e-05,
|
| 718 |
+
"loss": 0.0355,
|
| 719 |
+
"step": 24000
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 13.45,
|
| 723 |
+
"eval_loss": 0.19105392694473267,
|
| 724 |
+
"eval_runtime": 172.4407,
|
| 725 |
+
"eval_samples_per_second": 15.194,
|
| 726 |
+
"eval_steps_per_second": 1.902,
|
| 727 |
+
"eval_wer": 0.873915855143031,
|
| 728 |
+
"step": 24000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 13.73,
|
| 732 |
+
"learning_rate": 5.5281797410510284e-05,
|
| 733 |
+
"loss": 0.0344,
|
| 734 |
+
"step": 24500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 13.73,
|
| 738 |
+
"eval_loss": 0.31946560740470886,
|
| 739 |
+
"eval_runtime": 172.356,
|
| 740 |
+
"eval_samples_per_second": 15.201,
|
| 741 |
+
"eval_steps_per_second": 1.903,
|
| 742 |
+
"eval_wer": 0.8714812842361533,
|
| 743 |
+
"step": 24500
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 14.01,
|
| 747 |
+
"learning_rate": 5.432977913175933e-05,
|
| 748 |
+
"loss": 0.0354,
|
| 749 |
+
"step": 25000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 14.01,
|
| 753 |
+
"eval_loss": 0.23475807905197144,
|
| 754 |
+
"eval_runtime": 173.688,
|
| 755 |
+
"eval_samples_per_second": 15.085,
|
| 756 |
+
"eval_steps_per_second": 1.888,
|
| 757 |
+
"eval_wer": 0.8856892878880097,
|
| 758 |
+
"step": 25000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 14.29,
|
| 762 |
+
"learning_rate": 5.3377760853008385e-05,
|
| 763 |
+
"loss": 0.0327,
|
| 764 |
+
"step": 25500
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 14.29,
|
| 768 |
+
"eval_loss": 0.1936648041009903,
|
| 769 |
+
"eval_runtime": 173.1214,
|
| 770 |
+
"eval_samples_per_second": 15.134,
|
| 771 |
+
"eval_steps_per_second": 1.895,
|
| 772 |
+
"eval_wer": 0.8734974132684115,
|
| 773 |
+
"step": 25500
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 14.57,
|
| 777 |
+
"learning_rate": 5.242574257425743e-05,
|
| 778 |
+
"loss": 0.0318,
|
| 779 |
+
"step": 26000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 14.57,
|
| 783 |
+
"eval_loss": 0.264340341091156,
|
| 784 |
+
"eval_runtime": 172.855,
|
| 785 |
+
"eval_samples_per_second": 15.157,
|
| 786 |
+
"eval_steps_per_second": 1.898,
|
| 787 |
+
"eval_wer": 0.8679055082166768,
|
| 788 |
+
"step": 26000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 14.85,
|
| 792 |
+
"learning_rate": 5.147372429550648e-05,
|
| 793 |
+
"loss": 0.032,
|
| 794 |
+
"step": 26500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 14.85,
|
| 798 |
+
"eval_loss": 0.39989423751831055,
|
| 799 |
+
"eval_runtime": 173.1827,
|
| 800 |
+
"eval_samples_per_second": 15.129,
|
| 801 |
+
"eval_steps_per_second": 1.894,
|
| 802 |
+
"eval_wer": 0.8897025258673159,
|
| 803 |
+
"step": 26500
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 15.13,
|
| 807 |
+
"learning_rate": 5.052170601675552e-05,
|
| 808 |
+
"loss": 0.0318,
|
| 809 |
+
"step": 27000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 15.13,
|
| 813 |
+
"eval_loss": 0.17986007034778595,
|
| 814 |
+
"eval_runtime": 172.2454,
|
| 815 |
+
"eval_samples_per_second": 15.211,
|
| 816 |
+
"eval_steps_per_second": 1.904,
|
| 817 |
+
"eval_wer": 0.8881618989653074,
|
| 818 |
+
"step": 27000
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 15.41,
|
| 822 |
+
"learning_rate": 4.956968773800457e-05,
|
| 823 |
+
"loss": 0.0313,
|
| 824 |
+
"step": 27500
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 15.41,
|
| 828 |
+
"eval_loss": 0.198421448469162,
|
| 829 |
+
"eval_runtime": 172.7438,
|
| 830 |
+
"eval_samples_per_second": 15.167,
|
| 831 |
+
"eval_steps_per_second": 1.899,
|
| 832 |
+
"eval_wer": 0.8861647900182593,
|
| 833 |
+
"step": 27500
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 15.7,
|
| 837 |
+
"learning_rate": 4.861957349581112e-05,
|
| 838 |
+
"loss": 0.0307,
|
| 839 |
+
"step": 28000
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 15.7,
|
| 843 |
+
"eval_loss": 0.20344625413417816,
|
| 844 |
+
"eval_runtime": 173.6375,
|
| 845 |
+
"eval_samples_per_second": 15.089,
|
| 846 |
+
"eval_steps_per_second": 1.889,
|
| 847 |
+
"eval_wer": 0.8731550517346318,
|
| 848 |
+
"step": 28000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 15.98,
|
| 852 |
+
"learning_rate": 4.766755521706017e-05,
|
| 853 |
+
"loss": 0.0302,
|
| 854 |
+
"step": 28500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 15.98,
|
| 858 |
+
"eval_loss": 0.17505203187465668,
|
| 859 |
+
"eval_runtime": 173.3071,
|
| 860 |
+
"eval_samples_per_second": 15.118,
|
| 861 |
+
"eval_steps_per_second": 1.893,
|
| 862 |
+
"eval_wer": 0.8726224893487523,
|
| 863 |
+
"step": 28500
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 16.26,
|
| 867 |
+
"learning_rate": 4.671553693830922e-05,
|
| 868 |
+
"loss": 0.0293,
|
| 869 |
+
"step": 29000
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 16.26,
|
| 873 |
+
"eval_loss": 0.2021918147802353,
|
| 874 |
+
"eval_runtime": 173.0402,
|
| 875 |
+
"eval_samples_per_second": 15.141,
|
| 876 |
+
"eval_steps_per_second": 1.896,
|
| 877 |
+
"eval_wer": 0.8536404443091905,
|
| 878 |
+
"step": 29000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 16.54,
|
| 882 |
+
"learning_rate": 4.576351865955826e-05,
|
| 883 |
+
"loss": 0.0284,
|
| 884 |
+
"step": 29500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 16.54,
|
| 888 |
+
"eval_loss": 0.1741979867219925,
|
| 889 |
+
"eval_runtime": 173.6754,
|
| 890 |
+
"eval_samples_per_second": 15.086,
|
| 891 |
+
"eval_steps_per_second": 1.889,
|
| 892 |
+
"eval_wer": 0.8673158855751674,
|
| 893 |
+
"step": 29500
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 16.82,
|
| 897 |
+
"learning_rate": 4.4811500380807313e-05,
|
| 898 |
+
"loss": 0.0293,
|
| 899 |
+
"step": 30000
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 16.82,
|
| 903 |
+
"eval_loss": 0.24190114438533783,
|
| 904 |
+
"eval_runtime": 172.7419,
|
| 905 |
+
"eval_samples_per_second": 15.167,
|
| 906 |
+
"eval_steps_per_second": 1.899,
|
| 907 |
+
"eval_wer": 0.88540398660986,
|
| 908 |
+
"step": 30000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 17.1,
|
| 912 |
+
"learning_rate": 4.385948210205636e-05,
|
| 913 |
+
"loss": 0.027,
|
| 914 |
+
"step": 30500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 17.1,
|
| 918 |
+
"eval_loss": 0.2092837691307068,
|
| 919 |
+
"eval_runtime": 172.4386,
|
| 920 |
+
"eval_samples_per_second": 15.194,
|
| 921 |
+
"eval_steps_per_second": 1.902,
|
| 922 |
+
"eval_wer": 0.8659274193548387,
|
| 923 |
+
"step": 30500
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 17.38,
|
| 927 |
+
"learning_rate": 4.290746382330541e-05,
|
| 928 |
+
"loss": 0.0276,
|
| 929 |
+
"step": 31000
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 17.38,
|
| 933 |
+
"eval_loss": 0.18375042080879211,
|
| 934 |
+
"eval_runtime": 172.5864,
|
| 935 |
+
"eval_samples_per_second": 15.181,
|
| 936 |
+
"eval_steps_per_second": 1.9,
|
| 937 |
+
"eval_wer": 0.872945830797322,
|
| 938 |
+
"step": 31000
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 17.66,
|
| 942 |
+
"learning_rate": 4.195544554455446e-05,
|
| 943 |
+
"loss": 0.026,
|
| 944 |
+
"step": 31500
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 17.66,
|
| 948 |
+
"eval_loss": 0.18148663640022278,
|
| 949 |
+
"eval_runtime": 173.4358,
|
| 950 |
+
"eval_samples_per_second": 15.106,
|
| 951 |
+
"eval_steps_per_second": 1.891,
|
| 952 |
+
"eval_wer": 0.8676392270237371,
|
| 953 |
+
"step": 31500
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 17.94,
|
| 957 |
+
"learning_rate": 4.100342726580351e-05,
|
| 958 |
+
"loss": 0.0278,
|
| 959 |
+
"step": 32000
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 17.94,
|
| 963 |
+
"eval_loss": 0.18632376194000244,
|
| 964 |
+
"eval_runtime": 172.755,
|
| 965 |
+
"eval_samples_per_second": 15.166,
|
| 966 |
+
"eval_steps_per_second": 1.899,
|
| 967 |
+
"eval_wer": 0.8682098295800366,
|
| 968 |
+
"step": 32000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 18.22,
|
| 972 |
+
"learning_rate": 4.005140898705255e-05,
|
| 973 |
+
"loss": 0.0257,
|
| 974 |
+
"step": 32500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 18.22,
|
| 978 |
+
"eval_loss": 0.18216948211193085,
|
| 979 |
+
"eval_runtime": 172.6935,
|
| 980 |
+
"eval_samples_per_second": 15.171,
|
| 981 |
+
"eval_steps_per_second": 1.899,
|
| 982 |
+
"eval_wer": 0.8683429701765064,
|
| 983 |
+
"step": 32500
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 18.5,
|
| 987 |
+
"learning_rate": 3.90993907083016e-05,
|
| 988 |
+
"loss": 0.0257,
|
| 989 |
+
"step": 33000
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 18.5,
|
| 993 |
+
"eval_loss": 0.19133740663528442,
|
| 994 |
+
"eval_runtime": 171.6481,
|
| 995 |
+
"eval_samples_per_second": 15.264,
|
| 996 |
+
"eval_steps_per_second": 1.911,
|
| 997 |
+
"eval_wer": 0.8630744065733414,
|
| 998 |
+
"step": 33000
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 18.78,
|
| 1002 |
+
"learning_rate": 3.814737242955065e-05,
|
| 1003 |
+
"loss": 0.0252,
|
| 1004 |
+
"step": 33500
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 18.78,
|
| 1008 |
+
"eval_loss": 0.174552783370018,
|
| 1009 |
+
"eval_runtime": 172.18,
|
| 1010 |
+
"eval_samples_per_second": 15.217,
|
| 1011 |
+
"eval_steps_per_second": 1.905,
|
| 1012 |
+
"eval_wer": 0.8786708764455264,
|
| 1013 |
+
"step": 33500
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 19.06,
|
| 1017 |
+
"learning_rate": 3.71953541507997e-05,
|
| 1018 |
+
"loss": 0.0244,
|
| 1019 |
+
"step": 34000
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 19.06,
|
| 1023 |
+
"eval_loss": 0.1861806958913803,
|
| 1024 |
+
"eval_runtime": 173.107,
|
| 1025 |
+
"eval_samples_per_second": 15.135,
|
| 1026 |
+
"eval_steps_per_second": 1.895,
|
| 1027 |
+
"eval_wer": 0.8616288800973828,
|
| 1028 |
+
"step": 34000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 19.34,
|
| 1032 |
+
"learning_rate": 3.624904798172125e-05,
|
| 1033 |
+
"loss": 0.0239,
|
| 1034 |
+
"step": 34500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 19.34,
|
| 1038 |
+
"eval_loss": 0.19879654049873352,
|
| 1039 |
+
"eval_runtime": 173.2557,
|
| 1040 |
+
"eval_samples_per_second": 15.122,
|
| 1041 |
+
"eval_steps_per_second": 1.893,
|
| 1042 |
+
"eval_wer": 0.8602404138770542,
|
| 1043 |
+
"step": 34500
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 19.62,
|
| 1047 |
+
"learning_rate": 3.52970297029703e-05,
|
| 1048 |
+
"loss": 0.0243,
|
| 1049 |
+
"step": 35000
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 19.62,
|
| 1053 |
+
"eval_loss": 0.19157561659812927,
|
| 1054 |
+
"eval_runtime": 173.3754,
|
| 1055 |
+
"eval_samples_per_second": 15.112,
|
| 1056 |
+
"eval_steps_per_second": 1.892,
|
| 1057 |
+
"eval_wer": 0.8565885575167377,
|
| 1058 |
+
"step": 35000
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 19.9,
|
| 1062 |
+
"learning_rate": 3.434501142421935e-05,
|
| 1063 |
+
"loss": 0.0242,
|
| 1064 |
+
"step": 35500
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 19.9,
|
| 1068 |
+
"eval_loss": 0.21290895342826843,
|
| 1069 |
+
"eval_runtime": 173.7119,
|
| 1070 |
+
"eval_samples_per_second": 15.082,
|
| 1071 |
+
"eval_steps_per_second": 1.888,
|
| 1072 |
+
"eval_wer": 0.8637781497261108,
|
| 1073 |
+
"step": 35500
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 20.18,
|
| 1077 |
+
"learning_rate": 3.339299314546839e-05,
|
| 1078 |
+
"loss": 0.0242,
|
| 1079 |
+
"step": 36000
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 20.18,
|
| 1083 |
+
"eval_loss": 0.1832270473241806,
|
| 1084 |
+
"eval_runtime": 172.9999,
|
| 1085 |
+
"eval_samples_per_second": 15.145,
|
| 1086 |
+
"eval_steps_per_second": 1.896,
|
| 1087 |
+
"eval_wer": 0.8728126902008521,
|
| 1088 |
+
"step": 36000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 20.46,
|
| 1092 |
+
"learning_rate": 3.244097486671744e-05,
|
| 1093 |
+
"loss": 0.0223,
|
| 1094 |
+
"step": 36500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 20.46,
|
| 1098 |
+
"eval_loss": 0.16877047717571259,
|
| 1099 |
+
"eval_runtime": 171.9283,
|
| 1100 |
+
"eval_samples_per_second": 15.239,
|
| 1101 |
+
"eval_steps_per_second": 1.908,
|
| 1102 |
+
"eval_wer": 0.8582433049300061,
|
| 1103 |
+
"step": 36500
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 20.74,
|
| 1107 |
+
"learning_rate": 3.1488956587966486e-05,
|
| 1108 |
+
"loss": 0.023,
|
| 1109 |
+
"step": 37000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 20.74,
|
| 1113 |
+
"eval_loss": 0.19964033365249634,
|
| 1114 |
+
"eval_runtime": 173.6406,
|
| 1115 |
+
"eval_samples_per_second": 15.089,
|
| 1116 |
+
"eval_steps_per_second": 1.889,
|
| 1117 |
+
"eval_wer": 0.8681147291539866,
|
| 1118 |
+
"step": 37000
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 21.02,
|
| 1122 |
+
"learning_rate": 3.053693830921554e-05,
|
| 1123 |
+
"loss": 0.0229,
|
| 1124 |
+
"step": 37500
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 21.02,
|
| 1128 |
+
"eval_loss": 0.2140882909297943,
|
| 1129 |
+
"eval_runtime": 173.5521,
|
| 1130 |
+
"eval_samples_per_second": 15.096,
|
| 1131 |
+
"eval_steps_per_second": 1.89,
|
| 1132 |
+
"eval_wer": 0.8751141205112599,
|
| 1133 |
+
"step": 37500
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 21.3,
|
| 1137 |
+
"learning_rate": 2.9584920030464587e-05,
|
| 1138 |
+
"loss": 0.0217,
|
| 1139 |
+
"step": 38000
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 21.3,
|
| 1143 |
+
"eval_loss": 0.21021434664726257,
|
| 1144 |
+
"eval_runtime": 172.4762,
|
| 1145 |
+
"eval_samples_per_second": 15.191,
|
| 1146 |
+
"eval_steps_per_second": 1.902,
|
| 1147 |
+
"eval_wer": 0.8711199026171638,
|
| 1148 |
+
"step": 38000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 21.58,
|
| 1152 |
+
"learning_rate": 2.8632901751713637e-05,
|
| 1153 |
+
"loss": 0.0211,
|
| 1154 |
+
"step": 38500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 21.58,
|
| 1158 |
+
"eval_loss": 0.2082085758447647,
|
| 1159 |
+
"eval_runtime": 173.3781,
|
| 1160 |
+
"eval_samples_per_second": 15.111,
|
| 1161 |
+
"eval_steps_per_second": 1.892,
|
| 1162 |
+
"eval_wer": 0.8795458003651856,
|
| 1163 |
+
"step": 38500
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 21.86,
|
| 1167 |
+
"learning_rate": 2.7682787509520187e-05,
|
| 1168 |
+
"loss": 0.0216,
|
| 1169 |
+
"step": 39000
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 21.86,
|
| 1173 |
+
"eval_loss": 0.19715508818626404,
|
| 1174 |
+
"eval_runtime": 172.5766,
|
| 1175 |
+
"eval_samples_per_second": 15.182,
|
| 1176 |
+
"eval_steps_per_second": 1.901,
|
| 1177 |
+
"eval_wer": 0.8637020693852708,
|
| 1178 |
+
"step": 39000
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 22.14,
|
| 1182 |
+
"learning_rate": 2.673076923076923e-05,
|
| 1183 |
+
"loss": 0.0213,
|
| 1184 |
+
"step": 39500
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 22.14,
|
| 1188 |
+
"eval_loss": 0.21034926176071167,
|
| 1189 |
+
"eval_runtime": 173.6164,
|
| 1190 |
+
"eval_samples_per_second": 15.091,
|
| 1191 |
+
"eval_steps_per_second": 1.889,
|
| 1192 |
+
"eval_wer": 0.8720899269628728,
|
| 1193 |
+
"step": 39500
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 22.42,
|
| 1197 |
+
"learning_rate": 2.577875095201828e-05,
|
| 1198 |
+
"loss": 0.0201,
|
| 1199 |
+
"step": 40000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 22.42,
|
| 1203 |
+
"eval_loss": 0.21576687693595886,
|
| 1204 |
+
"eval_runtime": 173.7131,
|
| 1205 |
+
"eval_samples_per_second": 15.082,
|
| 1206 |
+
"eval_steps_per_second": 1.888,
|
| 1207 |
+
"eval_wer": 0.8763884662203286,
|
| 1208 |
+
"step": 40000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 22.7,
|
| 1212 |
+
"learning_rate": 2.4826732673267328e-05,
|
| 1213 |
+
"loss": 0.0206,
|
| 1214 |
+
"step": 40500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 22.7,
|
| 1218 |
+
"eval_loss": 0.21420465409755707,
|
| 1219 |
+
"eval_runtime": 172.3252,
|
| 1220 |
+
"eval_samples_per_second": 15.204,
|
| 1221 |
+
"eval_steps_per_second": 1.903,
|
| 1222 |
+
"eval_wer": 0.8615908399269628,
|
| 1223 |
+
"step": 40500
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 22.98,
|
| 1227 |
+
"learning_rate": 2.3874714394516375e-05,
|
| 1228 |
+
"loss": 0.02,
|
| 1229 |
+
"step": 41000
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 22.98,
|
| 1233 |
+
"eval_loss": 0.21698887646198273,
|
| 1234 |
+
"eval_runtime": 175.0748,
|
| 1235 |
+
"eval_samples_per_second": 14.965,
|
| 1236 |
+
"eval_steps_per_second": 1.873,
|
| 1237 |
+
"eval_wer": 0.8692939744370055,
|
| 1238 |
+
"step": 41000
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 23.26,
|
| 1242 |
+
"learning_rate": 2.2922696115765425e-05,
|
| 1243 |
+
"loss": 0.0199,
|
| 1244 |
+
"step": 41500
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 23.26,
|
| 1248 |
+
"eval_loss": 0.19081184267997742,
|
| 1249 |
+
"eval_runtime": 174.2525,
|
| 1250 |
+
"eval_samples_per_second": 15.036,
|
| 1251 |
+
"eval_steps_per_second": 1.882,
|
| 1252 |
+
"eval_wer": 0.8776247717589775,
|
| 1253 |
+
"step": 41500
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 23.54,
|
| 1257 |
+
"learning_rate": 2.1970677837014472e-05,
|
| 1258 |
+
"loss": 0.0197,
|
| 1259 |
+
"step": 42000
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 23.54,
|
| 1263 |
+
"eval_loss": 0.22879697382450104,
|
| 1264 |
+
"eval_runtime": 173.0989,
|
| 1265 |
+
"eval_samples_per_second": 15.136,
|
| 1266 |
+
"eval_steps_per_second": 1.895,
|
| 1267 |
+
"eval_wer": 0.8709867620206938,
|
| 1268 |
+
"step": 42000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 23.82,
|
| 1272 |
+
"learning_rate": 2.101865955826352e-05,
|
| 1273 |
+
"loss": 0.0191,
|
| 1274 |
+
"step": 42500
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 23.82,
|
| 1278 |
+
"eval_loss": 0.20562587678432465,
|
| 1279 |
+
"eval_runtime": 172.5196,
|
| 1280 |
+
"eval_samples_per_second": 15.187,
|
| 1281 |
+
"eval_steps_per_second": 1.901,
|
| 1282 |
+
"eval_wer": 0.8727366098600122,
|
| 1283 |
+
"step": 42500
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 24.1,
|
| 1287 |
+
"learning_rate": 2.006664127951257e-05,
|
| 1288 |
+
"loss": 0.0186,
|
| 1289 |
+
"step": 43000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 24.1,
|
| 1293 |
+
"eval_loss": 0.23561741411685944,
|
| 1294 |
+
"eval_runtime": 173.2932,
|
| 1295 |
+
"eval_samples_per_second": 15.119,
|
| 1296 |
+
"eval_steps_per_second": 1.893,
|
| 1297 |
+
"eval_wer": 0.8775867315885575,
|
| 1298 |
+
"step": 43000
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 24.38,
|
| 1302 |
+
"learning_rate": 1.911652703731912e-05,
|
| 1303 |
+
"loss": 0.0185,
|
| 1304 |
+
"step": 43500
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 24.38,
|
| 1308 |
+
"eval_loss": 0.25951194763183594,
|
| 1309 |
+
"eval_runtime": 173.3272,
|
| 1310 |
+
"eval_samples_per_second": 15.116,
|
| 1311 |
+
"eval_steps_per_second": 1.892,
|
| 1312 |
+
"eval_wer": 0.8748858794887401,
|
| 1313 |
+
"step": 43500
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 24.66,
|
| 1317 |
+
"learning_rate": 1.8166412795125666e-05,
|
| 1318 |
+
"loss": 0.0185,
|
| 1319 |
+
"step": 44000
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 24.66,
|
| 1323 |
+
"eval_loss": 0.2548380196094513,
|
| 1324 |
+
"eval_runtime": 172.782,
|
| 1325 |
+
"eval_samples_per_second": 15.164,
|
| 1326 |
+
"eval_steps_per_second": 1.898,
|
| 1327 |
+
"eval_wer": 0.8635499087035909,
|
| 1328 |
+
"step": 44000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 24.94,
|
| 1332 |
+
"learning_rate": 1.7214394516374716e-05,
|
| 1333 |
+
"loss": 0.0183,
|
| 1334 |
+
"step": 44500
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 24.94,
|
| 1338 |
+
"eval_loss": 0.20978882908821106,
|
| 1339 |
+
"eval_runtime": 173.1802,
|
| 1340 |
+
"eval_samples_per_second": 15.129,
|
| 1341 |
+
"eval_steps_per_second": 1.894,
|
| 1342 |
+
"eval_wer": 0.8623516433353622,
|
| 1343 |
+
"step": 44500
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 25.22,
|
| 1347 |
+
"learning_rate": 1.6264280274181266e-05,
|
| 1348 |
+
"loss": 0.0175,
|
| 1349 |
+
"step": 45000
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 25.22,
|
| 1353 |
+
"eval_loss": 0.1987256407737732,
|
| 1354 |
+
"eval_runtime": 172.2969,
|
| 1355 |
+
"eval_samples_per_second": 15.206,
|
| 1356 |
+
"eval_steps_per_second": 1.904,
|
| 1357 |
+
"eval_wer": 0.8696933962264151,
|
| 1358 |
+
"step": 45000
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 25.5,
|
| 1362 |
+
"learning_rate": 1.5312261995430313e-05,
|
| 1363 |
+
"loss": 0.0175,
|
| 1364 |
+
"step": 45500
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 25.5,
|
| 1368 |
+
"eval_loss": 0.19228355586528778,
|
| 1369 |
+
"eval_runtime": 173.0272,
|
| 1370 |
+
"eval_samples_per_second": 15.142,
|
| 1371 |
+
"eval_steps_per_second": 1.896,
|
| 1372 |
+
"eval_wer": 0.8653758368837492,
|
| 1373 |
+
"step": 45500
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 25.78,
|
| 1377 |
+
"learning_rate": 1.4360243716679362e-05,
|
| 1378 |
+
"loss": 0.0171,
|
| 1379 |
+
"step": 46000
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 25.78,
|
| 1383 |
+
"eval_loss": 0.18031327426433563,
|
| 1384 |
+
"eval_runtime": 173.6134,
|
| 1385 |
+
"eval_samples_per_second": 15.091,
|
| 1386 |
+
"eval_steps_per_second": 1.889,
|
| 1387 |
+
"eval_wer": 0.8679625684723068,
|
| 1388 |
+
"step": 46000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 26.07,
|
| 1392 |
+
"learning_rate": 1.3408225437928409e-05,
|
| 1393 |
+
"loss": 0.0167,
|
| 1394 |
+
"step": 46500
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 26.07,
|
| 1398 |
+
"eval_loss": 0.18257634341716766,
|
| 1399 |
+
"eval_runtime": 173.628,
|
| 1400 |
+
"eval_samples_per_second": 15.09,
|
| 1401 |
+
"eval_steps_per_second": 1.889,
|
| 1402 |
+
"eval_wer": 0.8620853621424224,
|
| 1403 |
+
"step": 46500
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 26.35,
|
| 1407 |
+
"learning_rate": 1.2456207159177456e-05,
|
| 1408 |
+
"loss": 0.016,
|
| 1409 |
+
"step": 47000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 26.35,
|
| 1413 |
+
"eval_loss": 0.18541069328784943,
|
| 1414 |
+
"eval_runtime": 172.5876,
|
| 1415 |
+
"eval_samples_per_second": 15.181,
|
| 1416 |
+
"eval_steps_per_second": 1.9,
|
| 1417 |
+
"eval_wer": 0.8629412659768716,
|
| 1418 |
+
"step": 47000
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 26.63,
|
| 1422 |
+
"learning_rate": 1.1504188880426504e-05,
|
| 1423 |
+
"loss": 0.0162,
|
| 1424 |
+
"step": 47500
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 26.63,
|
| 1428 |
+
"eval_loss": 0.1826283484697342,
|
| 1429 |
+
"eval_runtime": 173.6818,
|
| 1430 |
+
"eval_samples_per_second": 15.085,
|
| 1431 |
+
"eval_steps_per_second": 1.889,
|
| 1432 |
+
"eval_wer": 0.8728126902008521,
|
| 1433 |
+
"step": 47500
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 26.91,
|
| 1437 |
+
"learning_rate": 1.0552170601675553e-05,
|
| 1438 |
+
"loss": 0.0159,
|
| 1439 |
+
"step": 48000
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 26.91,
|
| 1443 |
+
"eval_loss": 0.18218345940113068,
|
| 1444 |
+
"eval_runtime": 174.9608,
|
| 1445 |
+
"eval_samples_per_second": 14.975,
|
| 1446 |
+
"eval_steps_per_second": 1.875,
|
| 1447 |
+
"eval_wer": 0.8637401095556908,
|
| 1448 |
+
"step": 48000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 27.19,
|
| 1452 |
+
"learning_rate": 9.600152322924601e-06,
|
| 1453 |
+
"loss": 0.0159,
|
| 1454 |
+
"step": 48500
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 27.19,
|
| 1458 |
+
"eval_loss": 0.19467805325984955,
|
| 1459 |
+
"eval_runtime": 173.9602,
|
| 1460 |
+
"eval_samples_per_second": 15.061,
|
| 1461 |
+
"eval_steps_per_second": 1.885,
|
| 1462 |
+
"eval_wer": 0.8611723980523432,
|
| 1463 |
+
"step": 48500
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 27.47,
|
| 1467 |
+
"learning_rate": 8.648134044173648e-06,
|
| 1468 |
+
"loss": 0.0159,
|
| 1469 |
+
"step": 49000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 27.47,
|
| 1473 |
+
"eval_loss": 0.19071811437606812,
|
| 1474 |
+
"eval_runtime": 173.4216,
|
| 1475 |
+
"eval_samples_per_second": 15.108,
|
| 1476 |
+
"eval_steps_per_second": 1.891,
|
| 1477 |
+
"eval_wer": 0.8624087035909921,
|
| 1478 |
+
"step": 49000
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 27.75,
|
| 1482 |
+
"learning_rate": 7.696115765422697e-06,
|
| 1483 |
+
"loss": 0.0152,
|
| 1484 |
+
"step": 49500
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 27.75,
|
| 1488 |
+
"eval_loss": 0.19280757009983063,
|
| 1489 |
+
"eval_runtime": 173.1959,
|
| 1490 |
+
"eval_samples_per_second": 15.127,
|
| 1491 |
+
"eval_steps_per_second": 1.894,
|
| 1492 |
+
"eval_wer": 0.8612484783931832,
|
| 1493 |
+
"step": 49500
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 28.03,
|
| 1497 |
+
"learning_rate": 6.744097486671744e-06,
|
| 1498 |
+
"loss": 0.0155,
|
| 1499 |
+
"step": 50000
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 28.03,
|
| 1503 |
+
"eval_loss": 0.1905270218849182,
|
| 1504 |
+
"eval_runtime": 172.8869,
|
| 1505 |
+
"eval_samples_per_second": 15.154,
|
| 1506 |
+
"eval_steps_per_second": 1.897,
|
| 1507 |
+
"eval_wer": 0.8642726719415703,
|
| 1508 |
+
"step": 50000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 28.31,
|
| 1512 |
+
"learning_rate": 5.792079207920793e-06,
|
| 1513 |
+
"loss": 0.0157,
|
| 1514 |
+
"step": 50500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 28.31,
|
| 1518 |
+
"eval_loss": 0.18496514856815338,
|
| 1519 |
+
"eval_runtime": 174.2609,
|
| 1520 |
+
"eval_samples_per_second": 15.035,
|
| 1521 |
+
"eval_steps_per_second": 1.882,
|
| 1522 |
+
"eval_wer": 0.8687804321363359,
|
| 1523 |
+
"step": 50500
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 28.59,
|
| 1527 |
+
"learning_rate": 4.841964965727342e-06,
|
| 1528 |
+
"loss": 0.0141,
|
| 1529 |
+
"step": 51000
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 28.59,
|
| 1533 |
+
"eval_loss": 0.19027549028396606,
|
| 1534 |
+
"eval_runtime": 174.6402,
|
| 1535 |
+
"eval_samples_per_second": 15.002,
|
| 1536 |
+
"eval_steps_per_second": 1.878,
|
| 1537 |
+
"eval_wer": 0.8673729458307973,
|
| 1538 |
+
"step": 51000
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 28.87,
|
| 1542 |
+
"learning_rate": 3.88994668697639e-06,
|
| 1543 |
+
"loss": 0.0146,
|
| 1544 |
+
"step": 51500
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 28.87,
|
| 1548 |
+
"eval_loss": 0.1888495832681656,
|
| 1549 |
+
"eval_runtime": 173.6727,
|
| 1550 |
+
"eval_samples_per_second": 15.086,
|
| 1551 |
+
"eval_steps_per_second": 1.889,
|
| 1552 |
+
"eval_wer": 0.8696363359707852,
|
| 1553 |
+
"step": 51500
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 29.15,
|
| 1557 |
+
"learning_rate": 2.9379284082254382e-06,
|
| 1558 |
+
"loss": 0.0141,
|
| 1559 |
+
"step": 52000
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 29.15,
|
| 1563 |
+
"eval_loss": 0.19334958493709564,
|
| 1564 |
+
"eval_runtime": 174.9971,
|
| 1565 |
+
"eval_samples_per_second": 14.972,
|
| 1566 |
+
"eval_steps_per_second": 1.874,
|
| 1567 |
+
"eval_wer": 0.8652617163724894,
|
| 1568 |
+
"step": 52000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 29.43,
|
| 1572 |
+
"learning_rate": 1.985910129474486e-06,
|
| 1573 |
+
"loss": 0.0145,
|
| 1574 |
+
"step": 52500
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 29.43,
|
| 1578 |
+
"eval_loss": 0.19555361568927765,
|
| 1579 |
+
"eval_runtime": 173.8338,
|
| 1580 |
+
"eval_samples_per_second": 15.072,
|
| 1581 |
+
"eval_steps_per_second": 1.887,
|
| 1582 |
+
"eval_wer": 0.8642346317711503,
|
| 1583 |
+
"step": 52500
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 29.71,
|
| 1587 |
+
"learning_rate": 1.033891850723534e-06,
|
| 1588 |
+
"loss": 0.0143,
|
| 1589 |
+
"step": 53000
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 29.71,
|
| 1593 |
+
"eval_loss": 0.19510917365550995,
|
| 1594 |
+
"eval_runtime": 173.0428,
|
| 1595 |
+
"eval_samples_per_second": 15.141,
|
| 1596 |
+
"eval_steps_per_second": 1.895,
|
| 1597 |
+
"eval_wer": 0.8620283018867925,
|
| 1598 |
+
"step": 53000
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 29.99,
|
| 1602 |
+
"learning_rate": 8.377760853008378e-08,
|
| 1603 |
+
"loss": 0.0148,
|
| 1604 |
+
"step": 53500
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 29.99,
|
| 1608 |
+
"eval_loss": 0.19494272768497467,
|
| 1609 |
+
"eval_runtime": 173.493,
|
| 1610 |
+
"eval_samples_per_second": 15.101,
|
| 1611 |
+
"eval_steps_per_second": 1.891,
|
| 1612 |
+
"eval_wer": 0.8613245587340231,
|
| 1613 |
+
"step": 53500
|
| 1614 |
+
}
|
| 1615 |
+
],
|
| 1616 |
+
"max_steps": 53520,
|
| 1617 |
+
"num_train_epochs": 30,
|
| 1618 |
+
"total_flos": 9.958592660590779e+19,
|
| 1619 |
+
"trial_name": null,
|
| 1620 |
+
"trial_params": null
|
| 1621 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a49dc6e3d3afe41e51c8be81b90710fab47c46b84e4147a98d6ef9aa1821e728
|
| 3 |
+
size 3003
|