Amiran13 commited on
Commit
03a5a0a
1 Parent(s): b543a46

Training in progress, step 5200, checkpoint

Browse files
.gitattributes CHANGED
@@ -24,3 +24,4 @@ checkpoint-4400/optimizer.pt filter=lfs diff=lfs merge=lfs -text
24
  checkpoint-4600/optimizer.pt filter=lfs diff=lfs merge=lfs -text
25
  checkpoint-4800/optimizer.pt filter=lfs diff=lfs merge=lfs -text
26
  checkpoint-5000/optimizer.pt filter=lfs diff=lfs merge=lfs -text
 
 
24
  checkpoint-4600/optimizer.pt filter=lfs diff=lfs merge=lfs -text
25
  checkpoint-4800/optimizer.pt filter=lfs diff=lfs merge=lfs -text
26
  checkpoint-5000/optimizer.pt filter=lfs diff=lfs merge=lfs -text
27
+ checkpoint-5200/optimizer.pt filter=lfs diff=lfs merge=lfs -text
checkpoint-5200/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.0,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.1,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.1,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 10,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.05,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
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": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 38,
90
+ "proj_codevector_dim": 768,
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": "4.35.2",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 41,
116
+ "xvector_output_dim": 512
117
+ }
checkpoint-5200/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b82e4aa578e5a4a885c0fd5e9a5fa300707b02ee0f929b53c9ade058fd0b6ff2
3
+ size 1261975580
checkpoint-5200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c25b51db653221c8b138d9ced8ae1cb9133e9867e54cad0c36da113d86c3bbfd
3
+ size 2490495926
checkpoint-5200/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": true,
8
+ "sampling_rate": 16000
9
+ }
checkpoint-5200/rng_state.pth ADDED
Binary file (14.2 kB). View file
 
checkpoint-5200/scheduler.pt ADDED
Binary file (1.06 kB). View file
 
checkpoint-5200/trainer_state.json ADDED
@@ -0,0 +1,3607 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 8.467046379170057,
5
+ "eval_steps": 100,
6
+ "global_step": 5200,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02,
13
+ "learning_rate": 1.4999999999999999e-05,
14
+ "loss": 11.8163,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.03,
19
+ "learning_rate": 2.8499999999999998e-05,
20
+ "loss": 11.8929,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.05,
25
+ "learning_rate": 4.3499999999999993e-05,
26
+ "loss": 12.0064,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.07,
31
+ "learning_rate": 5.6999999999999996e-05,
32
+ "loss": 10.5132,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.08,
37
+ "learning_rate": 7.199999999999999e-05,
38
+ "loss": 7.9699,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.1,
43
+ "learning_rate": 8.699999999999999e-05,
44
+ "loss": 4.7703,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.11,
49
+ "learning_rate": 0.000102,
50
+ "loss": 3.7982,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.13,
55
+ "learning_rate": 0.000117,
56
+ "loss": 3.436,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.15,
61
+ "learning_rate": 0.00013199999999999998,
62
+ "loss": 3.2573,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.16,
67
+ "learning_rate": 0.000147,
68
+ "loss": 3.1412,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.16,
73
+ "eval_loss": 3.0855324268341064,
74
+ "eval_runtime": 1038.5545,
75
+ "eval_samples_per_second": 11.178,
76
+ "eval_steps_per_second": 0.699,
77
+ "eval_wer": 1.0,
78
+ "step": 100
79
+ },
80
+ {
81
+ "epoch": 0.18,
82
+ "learning_rate": 0.000162,
83
+ "loss": 3.0577,
84
+ "step": 110
85
+ },
86
+ {
87
+ "epoch": 0.2,
88
+ "learning_rate": 0.00017699999999999997,
89
+ "loss": 3.0443,
90
+ "step": 120
91
+ },
92
+ {
93
+ "epoch": 0.21,
94
+ "learning_rate": 0.00019199999999999998,
95
+ "loss": 3.0597,
96
+ "step": 130
97
+ },
98
+ {
99
+ "epoch": 0.23,
100
+ "learning_rate": 0.00020699999999999996,
101
+ "loss": 3.061,
102
+ "step": 140
103
+ },
104
+ {
105
+ "epoch": 0.24,
106
+ "learning_rate": 0.00022199999999999998,
107
+ "loss": 3.0717,
108
+ "step": 150
109
+ },
110
+ {
111
+ "epoch": 0.26,
112
+ "learning_rate": 0.000237,
113
+ "loss": 3.0299,
114
+ "step": 160
115
+ },
116
+ {
117
+ "epoch": 0.28,
118
+ "learning_rate": 0.00025199999999999995,
119
+ "loss": 3.0295,
120
+ "step": 170
121
+ },
122
+ {
123
+ "epoch": 0.29,
124
+ "learning_rate": 0.000267,
125
+ "loss": 3.0423,
126
+ "step": 180
127
+ },
128
+ {
129
+ "epoch": 0.31,
130
+ "learning_rate": 0.00028199999999999997,
131
+ "loss": 3.0437,
132
+ "step": 190
133
+ },
134
+ {
135
+ "epoch": 0.33,
136
+ "learning_rate": 0.00029699999999999996,
137
+ "loss": 3.0569,
138
+ "step": 200
139
+ },
140
+ {
141
+ "epoch": 0.33,
142
+ "eval_loss": 3.036912679672241,
143
+ "eval_runtime": 1030.0275,
144
+ "eval_samples_per_second": 11.271,
145
+ "eval_steps_per_second": 0.705,
146
+ "eval_wer": 1.0,
147
+ "step": 200
148
+ },
149
+ {
150
+ "epoch": 0.34,
151
+ "learning_rate": 0.00029973362930077685,
152
+ "loss": 3.0243,
153
+ "step": 210
154
+ },
155
+ {
156
+ "epoch": 0.36,
157
+ "learning_rate": 0.000299400665926748,
158
+ "loss": 3.0236,
159
+ "step": 220
160
+ },
161
+ {
162
+ "epoch": 0.37,
163
+ "learning_rate": 0.0002990677025527192,
164
+ "loss": 3.0316,
165
+ "step": 230
166
+ },
167
+ {
168
+ "epoch": 0.39,
169
+ "learning_rate": 0.0002987347391786903,
170
+ "loss": 3.0363,
171
+ "step": 240
172
+ },
173
+ {
174
+ "epoch": 0.41,
175
+ "learning_rate": 0.00029840177580466146,
176
+ "loss": 3.0582,
177
+ "step": 250
178
+ },
179
+ {
180
+ "epoch": 0.42,
181
+ "learning_rate": 0.0002980688124306326,
182
+ "loss": 3.0144,
183
+ "step": 260
184
+ },
185
+ {
186
+ "epoch": 0.44,
187
+ "learning_rate": 0.00029773584905660376,
188
+ "loss": 3.0126,
189
+ "step": 270
190
+ },
191
+ {
192
+ "epoch": 0.46,
193
+ "learning_rate": 0.0002974028856825749,
194
+ "loss": 3.0035,
195
+ "step": 280
196
+ },
197
+ {
198
+ "epoch": 0.47,
199
+ "learning_rate": 0.000297069922308546,
200
+ "loss": 2.9913,
201
+ "step": 290
202
+ },
203
+ {
204
+ "epoch": 0.49,
205
+ "learning_rate": 0.00029673695893451716,
206
+ "loss": 2.9625,
207
+ "step": 300
208
+ },
209
+ {
210
+ "epoch": 0.49,
211
+ "eval_loss": 2.977771520614624,
212
+ "eval_runtime": 1042.3792,
213
+ "eval_samples_per_second": 11.137,
214
+ "eval_steps_per_second": 0.696,
215
+ "eval_wer": 1.0,
216
+ "step": 300
217
+ },
218
+ {
219
+ "epoch": 0.5,
220
+ "learning_rate": 0.0002964039955604883,
221
+ "loss": 2.9686,
222
+ "step": 310
223
+ },
224
+ {
225
+ "epoch": 0.52,
226
+ "learning_rate": 0.00029607103218645946,
227
+ "loss": 2.8773,
228
+ "step": 320
229
+ },
230
+ {
231
+ "epoch": 0.54,
232
+ "learning_rate": 0.0002957380688124306,
233
+ "loss": 2.726,
234
+ "step": 330
235
+ },
236
+ {
237
+ "epoch": 0.55,
238
+ "learning_rate": 0.00029540510543840177,
239
+ "loss": 2.3671,
240
+ "step": 340
241
+ },
242
+ {
243
+ "epoch": 0.57,
244
+ "learning_rate": 0.0002950721420643729,
245
+ "loss": 1.9479,
246
+ "step": 350
247
+ },
248
+ {
249
+ "epoch": 0.59,
250
+ "learning_rate": 0.000294739178690344,
251
+ "loss": 1.5549,
252
+ "step": 360
253
+ },
254
+ {
255
+ "epoch": 0.6,
256
+ "learning_rate": 0.00029440621531631517,
257
+ "loss": 1.1765,
258
+ "step": 370
259
+ },
260
+ {
261
+ "epoch": 0.62,
262
+ "learning_rate": 0.0002940732519422863,
263
+ "loss": 0.9155,
264
+ "step": 380
265
+ },
266
+ {
267
+ "epoch": 0.63,
268
+ "learning_rate": 0.00029374028856825747,
269
+ "loss": 0.7823,
270
+ "step": 390
271
+ },
272
+ {
273
+ "epoch": 0.65,
274
+ "learning_rate": 0.0002934073251942286,
275
+ "loss": 0.7715,
276
+ "step": 400
277
+ },
278
+ {
279
+ "epoch": 0.65,
280
+ "eval_loss": 0.5112669467926025,
281
+ "eval_runtime": 1048.8544,
282
+ "eval_samples_per_second": 11.068,
283
+ "eval_steps_per_second": 0.692,
284
+ "eval_wer": 0.718480591159969,
285
+ "step": 400
286
+ },
287
+ {
288
+ "epoch": 0.67,
289
+ "learning_rate": 0.00029307436182019977,
290
+ "loss": 0.6635,
291
+ "step": 410
292
+ },
293
+ {
294
+ "epoch": 0.68,
295
+ "learning_rate": 0.0002927413984461709,
296
+ "loss": 0.6032,
297
+ "step": 420
298
+ },
299
+ {
300
+ "epoch": 0.7,
301
+ "learning_rate": 0.000292408435072142,
302
+ "loss": 0.5401,
303
+ "step": 430
304
+ },
305
+ {
306
+ "epoch": 0.72,
307
+ "learning_rate": 0.00029207547169811317,
308
+ "loss": 0.5276,
309
+ "step": 440
310
+ },
311
+ {
312
+ "epoch": 0.73,
313
+ "learning_rate": 0.0002917425083240843,
314
+ "loss": 0.5713,
315
+ "step": 450
316
+ },
317
+ {
318
+ "epoch": 0.75,
319
+ "learning_rate": 0.0002914095449500555,
320
+ "loss": 0.46,
321
+ "step": 460
322
+ },
323
+ {
324
+ "epoch": 0.76,
325
+ "learning_rate": 0.0002910765815760266,
326
+ "loss": 0.4508,
327
+ "step": 470
328
+ },
329
+ {
330
+ "epoch": 0.78,
331
+ "learning_rate": 0.0002907436182019977,
332
+ "loss": 0.447,
333
+ "step": 480
334
+ },
335
+ {
336
+ "epoch": 0.8,
337
+ "learning_rate": 0.00029041065482796893,
338
+ "loss": 0.4454,
339
+ "step": 490
340
+ },
341
+ {
342
+ "epoch": 0.81,
343
+ "learning_rate": 0.0002900776914539401,
344
+ "loss": 0.4725,
345
+ "step": 500
346
+ },
347
+ {
348
+ "epoch": 0.81,
349
+ "eval_loss": 0.3072386682033539,
350
+ "eval_runtime": 1051.5656,
351
+ "eval_samples_per_second": 11.04,
352
+ "eval_steps_per_second": 0.69,
353
+ "eval_wer": 0.5137640833827487,
354
+ "step": 500
355
+ },
356
+ {
357
+ "epoch": 0.83,
358
+ "learning_rate": 0.0002897447280799112,
359
+ "loss": 0.3822,
360
+ "step": 510
361
+ },
362
+ {
363
+ "epoch": 0.85,
364
+ "learning_rate": 0.00028941176470588233,
365
+ "loss": 0.3807,
366
+ "step": 520
367
+ },
368
+ {
369
+ "epoch": 0.86,
370
+ "learning_rate": 0.0002890788013318535,
371
+ "loss": 0.3827,
372
+ "step": 530
373
+ },
374
+ {
375
+ "epoch": 0.88,
376
+ "learning_rate": 0.00028874583795782463,
377
+ "loss": 0.3945,
378
+ "step": 540
379
+ },
380
+ {
381
+ "epoch": 0.9,
382
+ "learning_rate": 0.00028841287458379573,
383
+ "loss": 0.4519,
384
+ "step": 550
385
+ },
386
+ {
387
+ "epoch": 0.91,
388
+ "learning_rate": 0.0002880799112097669,
389
+ "loss": 0.3534,
390
+ "step": 560
391
+ },
392
+ {
393
+ "epoch": 0.93,
394
+ "learning_rate": 0.00028774694783573803,
395
+ "loss": 0.3731,
396
+ "step": 570
397
+ },
398
+ {
399
+ "epoch": 0.94,
400
+ "learning_rate": 0.0002874139844617092,
401
+ "loss": 0.3433,
402
+ "step": 580
403
+ },
404
+ {
405
+ "epoch": 0.96,
406
+ "learning_rate": 0.00028708102108768033,
407
+ "loss": 0.3547,
408
+ "step": 590
409
+ },
410
+ {
411
+ "epoch": 0.98,
412
+ "learning_rate": 0.0002867480577136515,
413
+ "loss": 0.4103,
414
+ "step": 600
415
+ },
416
+ {
417
+ "epoch": 0.98,
418
+ "eval_loss": 0.24468904733657837,
419
+ "eval_runtime": 1043.2136,
420
+ "eval_samples_per_second": 11.128,
421
+ "eval_steps_per_second": 0.696,
422
+ "eval_wer": 0.4336997673676048,
423
+ "step": 600
424
+ },
425
+ {
426
+ "epoch": 0.99,
427
+ "learning_rate": 0.00028641509433962264,
428
+ "loss": 0.3484,
429
+ "step": 610
430
+ },
431
+ {
432
+ "epoch": 1.01,
433
+ "learning_rate": 0.00028608213096559373,
434
+ "loss": 0.3775,
435
+ "step": 620
436
+ },
437
+ {
438
+ "epoch": 1.03,
439
+ "learning_rate": 0.0002857491675915649,
440
+ "loss": 0.3207,
441
+ "step": 630
442
+ },
443
+ {
444
+ "epoch": 1.04,
445
+ "learning_rate": 0.00028541620421753604,
446
+ "loss": 0.2971,
447
+ "step": 640
448
+ },
449
+ {
450
+ "epoch": 1.06,
451
+ "learning_rate": 0.0002850832408435072,
452
+ "loss": 0.3129,
453
+ "step": 650
454
+ },
455
+ {
456
+ "epoch": 1.07,
457
+ "learning_rate": 0.00028475027746947834,
458
+ "loss": 0.3143,
459
+ "step": 660
460
+ },
461
+ {
462
+ "epoch": 1.09,
463
+ "learning_rate": 0.0002844173140954495,
464
+ "loss": 0.3179,
465
+ "step": 670
466
+ },
467
+ {
468
+ "epoch": 1.11,
469
+ "learning_rate": 0.00028408435072142064,
470
+ "loss": 0.299,
471
+ "step": 680
472
+ },
473
+ {
474
+ "epoch": 1.12,
475
+ "learning_rate": 0.0002837513873473918,
476
+ "loss": 0.2734,
477
+ "step": 690
478
+ },
479
+ {
480
+ "epoch": 1.14,
481
+ "learning_rate": 0.0002834184239733629,
482
+ "loss": 0.2775,
483
+ "step": 700
484
+ },
485
+ {
486
+ "epoch": 1.14,
487
+ "eval_loss": 0.2055242359638214,
488
+ "eval_runtime": 1053.2634,
489
+ "eval_samples_per_second": 11.022,
490
+ "eval_steps_per_second": 0.689,
491
+ "eval_wer": 0.3768644802262464,
492
+ "step": 700
493
+ },
494
+ {
495
+ "epoch": 1.16,
496
+ "learning_rate": 0.00028308546059933404,
497
+ "loss": 0.2878,
498
+ "step": 710
499
+ },
500
+ {
501
+ "epoch": 1.17,
502
+ "learning_rate": 0.0002827524972253052,
503
+ "loss": 0.307,
504
+ "step": 720
505
+ },
506
+ {
507
+ "epoch": 1.19,
508
+ "learning_rate": 0.00028241953385127634,
509
+ "loss": 0.2848,
510
+ "step": 730
511
+ },
512
+ {
513
+ "epoch": 1.2,
514
+ "learning_rate": 0.00028208657047724744,
515
+ "loss": 0.2744,
516
+ "step": 740
517
+ },
518
+ {
519
+ "epoch": 1.22,
520
+ "learning_rate": 0.00028175360710321865,
521
+ "loss": 0.2839,
522
+ "step": 750
523
+ },
524
+ {
525
+ "epoch": 1.24,
526
+ "learning_rate": 0.0002814206437291898,
527
+ "loss": 0.2888,
528
+ "step": 760
529
+ },
530
+ {
531
+ "epoch": 1.25,
532
+ "learning_rate": 0.0002810876803551609,
533
+ "loss": 0.288,
534
+ "step": 770
535
+ },
536
+ {
537
+ "epoch": 1.27,
538
+ "learning_rate": 0.00028075471698113205,
539
+ "loss": 0.2559,
540
+ "step": 780
541
+ },
542
+ {
543
+ "epoch": 1.29,
544
+ "learning_rate": 0.0002804217536071032,
545
+ "loss": 0.2498,
546
+ "step": 790
547
+ },
548
+ {
549
+ "epoch": 1.3,
550
+ "learning_rate": 0.00028008879023307435,
551
+ "loss": 0.2554,
552
+ "step": 800
553
+ },
554
+ {
555
+ "epoch": 1.3,
556
+ "eval_loss": 0.19498416781425476,
557
+ "eval_runtime": 1045.8615,
558
+ "eval_samples_per_second": 11.1,
559
+ "eval_steps_per_second": 0.694,
560
+ "eval_wer": 0.360272316744971,
561
+ "step": 800
562
+ },
563
+ {
564
+ "epoch": 1.32,
565
+ "learning_rate": 0.00027975582685904545,
566
+ "loss": 0.3059,
567
+ "step": 810
568
+ },
569
+ {
570
+ "epoch": 1.33,
571
+ "learning_rate": 0.0002794228634850166,
572
+ "loss": 0.281,
573
+ "step": 820
574
+ },
575
+ {
576
+ "epoch": 1.35,
577
+ "learning_rate": 0.00027908990011098775,
578
+ "loss": 0.2642,
579
+ "step": 830
580
+ },
581
+ {
582
+ "epoch": 1.37,
583
+ "learning_rate": 0.0002787569367369589,
584
+ "loss": 0.242,
585
+ "step": 840
586
+ },
587
+ {
588
+ "epoch": 1.38,
589
+ "learning_rate": 0.00027842397336293005,
590
+ "loss": 0.2541,
591
+ "step": 850
592
+ },
593
+ {
594
+ "epoch": 1.4,
595
+ "learning_rate": 0.0002780910099889012,
596
+ "loss": 0.2614,
597
+ "step": 860
598
+ },
599
+ {
600
+ "epoch": 1.42,
601
+ "learning_rate": 0.00027775804661487236,
602
+ "loss": 0.3001,
603
+ "step": 870
604
+ },
605
+ {
606
+ "epoch": 1.43,
607
+ "learning_rate": 0.0002774250832408435,
608
+ "loss": 0.2365,
609
+ "step": 880
610
+ },
611
+ {
612
+ "epoch": 1.45,
613
+ "learning_rate": 0.0002770921198668146,
614
+ "loss": 0.2373,
615
+ "step": 890
616
+ },
617
+ {
618
+ "epoch": 1.46,
619
+ "learning_rate": 0.00027675915649278575,
620
+ "loss": 0.263,
621
+ "step": 900
622
+ },
623
+ {
624
+ "epoch": 1.46,
625
+ "eval_loss": 0.1812964379787445,
626
+ "eval_runtime": 1053.0871,
627
+ "eval_samples_per_second": 11.024,
628
+ "eval_steps_per_second": 0.689,
629
+ "eval_wer": 0.3371687269078137,
630
+ "step": 900
631
+ },
632
+ {
633
+ "epoch": 1.48,
634
+ "learning_rate": 0.0002764261931187569,
635
+ "loss": 0.289,
636
+ "step": 910
637
+ },
638
+ {
639
+ "epoch": 1.5,
640
+ "learning_rate": 0.00027609322974472806,
641
+ "loss": 0.2578,
642
+ "step": 920
643
+ },
644
+ {
645
+ "epoch": 1.51,
646
+ "learning_rate": 0.0002757602663706992,
647
+ "loss": 0.2565,
648
+ "step": 930
649
+ },
650
+ {
651
+ "epoch": 1.53,
652
+ "learning_rate": 0.00027542730299667036,
653
+ "loss": 0.2543,
654
+ "step": 940
655
+ },
656
+ {
657
+ "epoch": 1.55,
658
+ "learning_rate": 0.0002750943396226415,
659
+ "loss": 0.2404,
660
+ "step": 950
661
+ },
662
+ {
663
+ "epoch": 1.56,
664
+ "learning_rate": 0.0002747613762486126,
665
+ "loss": 0.2291,
666
+ "step": 960
667
+ },
668
+ {
669
+ "epoch": 1.58,
670
+ "learning_rate": 0.00027442841287458376,
671
+ "loss": 0.2636,
672
+ "step": 970
673
+ },
674
+ {
675
+ "epoch": 1.59,
676
+ "learning_rate": 0.0002740954495005549,
677
+ "loss": 0.2222,
678
+ "step": 980
679
+ },
680
+ {
681
+ "epoch": 1.61,
682
+ "learning_rate": 0.00027376248612652606,
683
+ "loss": 0.2322,
684
+ "step": 990
685
+ },
686
+ {
687
+ "epoch": 1.63,
688
+ "learning_rate": 0.0002734295227524972,
689
+ "loss": 0.2294,
690
+ "step": 1000
691
+ },
692
+ {
693
+ "epoch": 1.63,
694
+ "eval_loss": 0.1664419174194336,
695
+ "eval_runtime": 1047.59,
696
+ "eval_samples_per_second": 11.082,
697
+ "eval_steps_per_second": 0.693,
698
+ "eval_wer": 0.3131756602654746,
699
+ "step": 1000
700
+ },
701
+ {
702
+ "epoch": 1.64,
703
+ "learning_rate": 0.00027309655937846837,
704
+ "loss": 0.249,
705
+ "step": 1010
706
+ },
707
+ {
708
+ "epoch": 1.66,
709
+ "learning_rate": 0.0002727635960044395,
710
+ "loss": 0.2682,
711
+ "step": 1020
712
+ },
713
+ {
714
+ "epoch": 1.68,
715
+ "learning_rate": 0.0002724306326304106,
716
+ "loss": 0.2242,
717
+ "step": 1030
718
+ },
719
+ {
720
+ "epoch": 1.69,
721
+ "learning_rate": 0.00027209766925638177,
722
+ "loss": 0.2333,
723
+ "step": 1040
724
+ },
725
+ {
726
+ "epoch": 1.71,
727
+ "learning_rate": 0.0002717647058823529,
728
+ "loss": 0.2188,
729
+ "step": 1050
730
+ },
731
+ {
732
+ "epoch": 1.72,
733
+ "learning_rate": 0.00027143174250832407,
734
+ "loss": 0.2592,
735
+ "step": 1060
736
+ },
737
+ {
738
+ "epoch": 1.74,
739
+ "learning_rate": 0.0002710987791342952,
740
+ "loss": 0.268,
741
+ "step": 1070
742
+ },
743
+ {
744
+ "epoch": 1.76,
745
+ "learning_rate": 0.0002707658157602663,
746
+ "loss": 0.2091,
747
+ "step": 1080
748
+ },
749
+ {
750
+ "epoch": 1.77,
751
+ "learning_rate": 0.00027043285238623747,
752
+ "loss": 0.2069,
753
+ "step": 1090
754
+ },
755
+ {
756
+ "epoch": 1.79,
757
+ "learning_rate": 0.0002700998890122087,
758
+ "loss": 0.2296,
759
+ "step": 1100
760
+ },
761
+ {
762
+ "epoch": 1.79,
763
+ "eval_loss": 0.15650752186775208,
764
+ "eval_runtime": 1054.82,
765
+ "eval_samples_per_second": 11.006,
766
+ "eval_steps_per_second": 0.688,
767
+ "eval_wer": 0.2962413903206678,
768
+ "step": 1100
769
+ },
770
+ {
771
+ "epoch": 1.81,
772
+ "learning_rate": 0.00026976692563817977,
773
+ "loss": 0.2402,
774
+ "step": 1110
775
+ },
776
+ {
777
+ "epoch": 1.82,
778
+ "learning_rate": 0.0002694339622641509,
779
+ "loss": 0.2258,
780
+ "step": 1120
781
+ },
782
+ {
783
+ "epoch": 1.84,
784
+ "learning_rate": 0.0002691009988901221,
785
+ "loss": 0.2215,
786
+ "step": 1130
787
+ },
788
+ {
789
+ "epoch": 1.86,
790
+ "learning_rate": 0.0002687680355160932,
791
+ "loss": 0.2059,
792
+ "step": 1140
793
+ },
794
+ {
795
+ "epoch": 1.87,
796
+ "learning_rate": 0.0002684350721420643,
797
+ "loss": 0.2022,
798
+ "step": 1150
799
+ },
800
+ {
801
+ "epoch": 1.89,
802
+ "learning_rate": 0.0002681021087680355,
803
+ "loss": 0.2261,
804
+ "step": 1160
805
+ },
806
+ {
807
+ "epoch": 1.9,
808
+ "learning_rate": 0.0002677691453940066,
809
+ "loss": 0.2132,
810
+ "step": 1170
811
+ },
812
+ {
813
+ "epoch": 1.92,
814
+ "learning_rate": 0.0002674361820199778,
815
+ "loss": 0.2245,
816
+ "step": 1180
817
+ },
818
+ {
819
+ "epoch": 1.94,
820
+ "learning_rate": 0.00026710321864594893,
821
+ "loss": 0.2133,
822
+ "step": 1190
823
+ },
824
+ {
825
+ "epoch": 1.95,
826
+ "learning_rate": 0.0002667702552719201,
827
+ "loss": 0.2183,
828
+ "step": 1200
829
+ },
830
+ {
831
+ "epoch": 1.95,
832
+ "eval_loss": 0.14735129475593567,
833
+ "eval_runtime": 1059.1579,
834
+ "eval_samples_per_second": 10.961,
835
+ "eval_steps_per_second": 0.685,
836
+ "eval_wer": 0.29856771427268164,
837
+ "step": 1200
838
+ },
839
+ {
840
+ "epoch": 1.97,
841
+ "learning_rate": 0.00026643729189789123,
842
+ "loss": 0.2204,
843
+ "step": 1210
844
+ },
845
+ {
846
+ "epoch": 1.99,
847
+ "learning_rate": 0.00026610432852386233,
848
+ "loss": 0.2482,
849
+ "step": 1220
850
+ },
851
+ {
852
+ "epoch": 2.0,
853
+ "learning_rate": 0.0002657713651498335,
854
+ "loss": 0.2214,
855
+ "step": 1230
856
+ },
857
+ {
858
+ "epoch": 2.02,
859
+ "learning_rate": 0.00026543840177580463,
860
+ "loss": 0.1779,
861
+ "step": 1240
862
+ },
863
+ {
864
+ "epoch": 2.03,
865
+ "learning_rate": 0.0002651054384017758,
866
+ "loss": 0.1688,
867
+ "step": 1250
868
+ },
869
+ {
870
+ "epoch": 2.05,
871
+ "learning_rate": 0.00026477247502774693,
872
+ "loss": 0.1915,
873
+ "step": 1260
874
+ },
875
+ {
876
+ "epoch": 2.07,
877
+ "learning_rate": 0.0002644395116537181,
878
+ "loss": 0.1942,
879
+ "step": 1270
880
+ },
881
+ {
882
+ "epoch": 2.08,
883
+ "learning_rate": 0.00026410654827968924,
884
+ "loss": 0.2266,
885
+ "step": 1280
886
+ },
887
+ {
888
+ "epoch": 2.1,
889
+ "learning_rate": 0.0002637735849056604,
890
+ "loss": 0.1931,
891
+ "step": 1290
892
+ },
893
+ {
894
+ "epoch": 2.12,
895
+ "learning_rate": 0.0002634406215316315,
896
+ "loss": 0.1822,
897
+ "step": 1300
898
+ },
899
+ {
900
+ "epoch": 2.12,
901
+ "eval_loss": 0.1546158790588379,
902
+ "eval_runtime": 1062.5837,
903
+ "eval_samples_per_second": 10.925,
904
+ "eval_steps_per_second": 0.683,
905
+ "eval_wer": 0.2810974775350089,
906
+ "step": 1300
907
+ },
908
+ {
909
+ "epoch": 2.13,
910
+ "learning_rate": 0.00026310765815760264,
911
+ "loss": 0.1735,
912
+ "step": 1310
913
+ },
914
+ {
915
+ "epoch": 2.15,
916
+ "learning_rate": 0.0002627746947835738,
917
+ "loss": 0.2042,
918
+ "step": 1320
919
+ },
920
+ {
921
+ "epoch": 2.16,
922
+ "learning_rate": 0.00026244173140954494,
923
+ "loss": 0.2347,
924
+ "step": 1330
925
+ },
926
+ {
927
+ "epoch": 2.18,
928
+ "learning_rate": 0.00026210876803551604,
929
+ "loss": 0.1816,
930
+ "step": 1340
931
+ },
932
+ {
933
+ "epoch": 2.2,
934
+ "learning_rate": 0.0002617758046614872,
935
+ "loss": 0.1838,
936
+ "step": 1350
937
+ },
938
+ {
939
+ "epoch": 2.21,
940
+ "learning_rate": 0.0002614428412874584,
941
+ "loss": 0.1856,
942
+ "step": 1360
943
+ },
944
+ {
945
+ "epoch": 2.23,
946
+ "learning_rate": 0.0002611098779134295,
947
+ "loss": 0.1866,
948
+ "step": 1370
949
+ },
950
+ {
951
+ "epoch": 2.25,
952
+ "learning_rate": 0.00026077691453940064,
953
+ "loss": 0.2016,
954
+ "step": 1380
955
+ },
956
+ {
957
+ "epoch": 2.26,
958
+ "learning_rate": 0.0002604439511653718,
959
+ "loss": 0.1813,
960
+ "step": 1390
961
+ },
962
+ {
963
+ "epoch": 2.28,
964
+ "learning_rate": 0.00026011098779134294,
965
+ "loss": 0.1798,
966
+ "step": 1400
967
+ },
968
+ {
969
+ "epoch": 2.28,
970
+ "eval_loss": 0.14420565962791443,
971
+ "eval_runtime": 1057.3911,
972
+ "eval_samples_per_second": 10.979,
973
+ "eval_steps_per_second": 0.687,
974
+ "eval_wer": 0.28108607398622454,
975
+ "step": 1400
976
+ },
977
+ {
978
+ "epoch": 2.29,
979
+ "learning_rate": 0.00025977802441731404,
980
+ "loss": 0.184,
981
+ "step": 1410
982
+ },
983
+ {
984
+ "epoch": 2.31,
985
+ "learning_rate": 0.0002594450610432852,
986
+ "loss": 0.19,
987
+ "step": 1420
988
+ },
989
+ {
990
+ "epoch": 2.33,
991
+ "learning_rate": 0.00025911209766925634,
992
+ "loss": 0.1982,
993
+ "step": 1430
994
+ },
995
+ {
996
+ "epoch": 2.34,
997
+ "learning_rate": 0.0002587791342952275,
998
+ "loss": 0.1627,
999
+ "step": 1440
1000
+ },
1001
+ {
1002
+ "epoch": 2.36,
1003
+ "learning_rate": 0.00025844617092119865,
1004
+ "loss": 0.1914,
1005
+ "step": 1450
1006
+ },
1007
+ {
1008
+ "epoch": 2.38,
1009
+ "learning_rate": 0.0002581132075471698,
1010
+ "loss": 0.171,
1011
+ "step": 1460
1012
+ },
1013
+ {
1014
+ "epoch": 2.39,
1015
+ "learning_rate": 0.00025778024417314095,
1016
+ "loss": 0.1653,
1017
+ "step": 1470
1018
+ },
1019
+ {
1020
+ "epoch": 2.41,
1021
+ "learning_rate": 0.0002574472807991121,
1022
+ "loss": 0.2263,
1023
+ "step": 1480
1024
+ },
1025
+ {
1026
+ "epoch": 2.42,
1027
+ "learning_rate": 0.0002571143174250832,
1028
+ "loss": 0.1801,
1029
+ "step": 1490
1030
+ },
1031
+ {
1032
+ "epoch": 2.44,
1033
+ "learning_rate": 0.00025678135405105435,
1034
+ "loss": 0.179,
1035
+ "step": 1500
1036
+ },
1037
+ {
1038
+ "epoch": 2.44,
1039
+ "eval_loss": 0.14106744527816772,
1040
+ "eval_runtime": 1059.0138,
1041
+ "eval_samples_per_second": 10.962,
1042
+ "eval_steps_per_second": 0.686,
1043
+ "eval_wer": 0.26864480226246407,
1044
+ "step": 1500
1045
+ },
1046
+ {
1047
+ "epoch": 2.46,
1048
+ "learning_rate": 0.0002564483906770255,
1049
+ "loss": 0.1718,
1050
+ "step": 1510
1051
+ },
1052
+ {
1053
+ "epoch": 2.47,
1054
+ "learning_rate": 0.00025611542730299665,
1055
+ "loss": 0.1659,
1056
+ "step": 1520
1057
+ },
1058
+ {
1059
+ "epoch": 2.49,
1060
+ "learning_rate": 0.0002557824639289678,
1061
+ "loss": 0.2146,
1062
+ "step": 1530
1063
+ },
1064
+ {
1065
+ "epoch": 2.51,
1066
+ "learning_rate": 0.00025544950055493896,
1067
+ "loss": 0.1789,
1068
+ "step": 1540
1069
+ },
1070
+ {
1071
+ "epoch": 2.52,
1072
+ "learning_rate": 0.0002551165371809101,
1073
+ "loss": 0.18,
1074
+ "step": 1550
1075
+ },
1076
+ {
1077
+ "epoch": 2.54,
1078
+ "learning_rate": 0.0002547835738068812,
1079
+ "loss": 0.176,
1080
+ "step": 1560
1081
+ },
1082
+ {
1083
+ "epoch": 2.55,
1084
+ "learning_rate": 0.00025445061043285236,
1085
+ "loss": 0.1789,
1086
+ "step": 1570
1087
+ },
1088
+ {
1089
+ "epoch": 2.57,
1090
+ "learning_rate": 0.0002541176470588235,
1091
+ "loss": 0.2318,
1092
+ "step": 1580
1093
+ },
1094
+ {
1095
+ "epoch": 2.59,
1096
+ "learning_rate": 0.00025378468368479466,
1097
+ "loss": 0.161,
1098
+ "step": 1590
1099
+ },
1100
+ {
1101
+ "epoch": 2.6,
1102
+ "learning_rate": 0.00025345172031076576,
1103
+ "loss": 0.1593,
1104
+ "step": 1600
1105
+ },
1106
+ {
1107
+ "epoch": 2.6,
1108
+ "eval_loss": 0.14078468084335327,
1109
+ "eval_runtime": 1052.8874,
1110
+ "eval_samples_per_second": 11.026,
1111
+ "eval_steps_per_second": 0.69,
1112
+ "eval_wer": 0.2739132418008484,
1113
+ "step": 1600
1114
+ },
1115
+ {
1116
+ "epoch": 2.62,
1117
+ "learning_rate": 0.0002531187569367369,
1118
+ "loss": 5.5922,
1119
+ "step": 1610
1120
+ },
1121
+ {
1122
+ "epoch": 2.64,
1123
+ "learning_rate": 0.0002527857935627081,
1124
+ "loss": 3.2059,
1125
+ "step": 1620
1126
+ },
1127
+ {
1128
+ "epoch": 2.65,
1129
+ "learning_rate": 0.0002524528301886792,
1130
+ "loss": 3.0426,
1131
+ "step": 1630
1132
+ },
1133
+ {
1134
+ "epoch": 2.67,
1135
+ "learning_rate": 0.00025211986681465036,
1136
+ "loss": 2.9869,
1137
+ "step": 1640
1138
+ },
1139
+ {
1140
+ "epoch": 2.69,
1141
+ "learning_rate": 0.0002517869034406215,
1142
+ "loss": 2.936,
1143
+ "step": 1650
1144
+ },
1145
+ {
1146
+ "epoch": 2.7,
1147
+ "learning_rate": 0.00025145394006659266,
1148
+ "loss": 2.5037,
1149
+ "step": 1660
1150
+ },
1151
+ {
1152
+ "epoch": 2.72,
1153
+ "learning_rate": 0.0002511209766925638,
1154
+ "loss": 1.297,
1155
+ "step": 1670
1156
+ },
1157
+ {
1158
+ "epoch": 2.74,
1159
+ "learning_rate": 0.0002507880133185349,
1160
+ "loss": 0.56,
1161
+ "step": 1680
1162
+ },
1163
+ {
1164
+ "epoch": 2.75,
1165
+ "learning_rate": 0.00025045504994450606,
1166
+ "loss": 0.3262,
1167
+ "step": 1690
1168
+ },
1169
+ {
1170
+ "epoch": 2.77,
1171
+ "learning_rate": 0.0002501220865704772,
1172
+ "loss": 0.2652,
1173
+ "step": 1700
1174
+ },
1175
+ {
1176
+ "epoch": 2.77,
1177
+ "eval_loss": 0.20742562413215637,
1178
+ "eval_runtime": 1056.4173,
1179
+ "eval_samples_per_second": 10.989,
1180
+ "eval_steps_per_second": 0.687,
1181
+ "eval_wer": 0.4499384208365643,
1182
+ "step": 1700
1183
+ },
1184
+ {
1185
+ "epoch": 2.78,
1186
+ "learning_rate": 0.00024978912319644837,
1187
+ "loss": 0.2585,
1188
+ "step": 1710
1189
+ },
1190
+ {
1191
+ "epoch": 2.8,
1192
+ "learning_rate": 0.0002494561598224195,
1193
+ "loss": 0.2722,
1194
+ "step": 1720
1195
+ },
1196
+ {
1197
+ "epoch": 2.82,
1198
+ "learning_rate": 0.00024912319644839067,
1199
+ "loss": 0.2571,
1200
+ "step": 1730
1201
+ },
1202
+ {
1203
+ "epoch": 2.83,
1204
+ "learning_rate": 0.0002487902330743618,
1205
+ "loss": 0.2256,
1206
+ "step": 1740
1207
+ },
1208
+ {
1209
+ "epoch": 2.85,
1210
+ "learning_rate": 0.0002484572697003329,
1211
+ "loss": 0.2176,
1212
+ "step": 1750
1213
+ },
1214
+ {
1215
+ "epoch": 2.87,
1216
+ "learning_rate": 0.00024812430632630407,
1217
+ "loss": 0.1954,
1218
+ "step": 1760
1219
+ },
1220
+ {
1221
+ "epoch": 2.88,
1222
+ "learning_rate": 0.0002477913429522752,
1223
+ "loss": 0.2099,
1224
+ "step": 1770
1225
+ },
1226
+ {
1227
+ "epoch": 2.9,
1228
+ "learning_rate": 0.00024745837957824637,
1229
+ "loss": 0.2316,
1230
+ "step": 1780
1231
+ },
1232
+ {
1233
+ "epoch": 2.91,
1234
+ "learning_rate": 0.0002471254162042175,
1235
+ "loss": 0.1814,
1236
+ "step": 1790
1237
+ },
1238
+ {
1239
+ "epoch": 2.93,
1240
+ "learning_rate": 0.0002467924528301887,
1241
+ "loss": 0.1834,
1242
+ "step": 1800
1243
+ },
1244
+ {
1245
+ "epoch": 2.93,
1246
+ "eval_loss": 0.15701305866241455,
1247
+ "eval_runtime": 1047.7925,
1248
+ "eval_samples_per_second": 11.079,
1249
+ "eval_steps_per_second": 0.693,
1250
+ "eval_wer": 0.39416366373215345,
1251
+ "step": 1800
1252
+ },
1253
+ {
1254
+ "epoch": 2.95,
1255
+ "learning_rate": 0.0002464594894561598,
1256
+ "loss": 0.2047,
1257
+ "step": 1810
1258
+ },
1259
+ {
1260
+ "epoch": 2.96,
1261
+ "learning_rate": 0.0002461265260821309,
1262
+ "loss": 0.2051,
1263
+ "step": 1820
1264
+ },
1265
+ {
1266
+ "epoch": 2.98,
1267
+ "learning_rate": 0.0002457935627081021,
1268
+ "loss": 0.2246,
1269
+ "step": 1830
1270
+ },
1271
+ {
1272
+ "epoch": 3.0,
1273
+ "learning_rate": 0.0002454605993340732,
1274
+ "loss": 0.1797,
1275
+ "step": 1840
1276
+ },
1277
+ {
1278
+ "epoch": 3.01,
1279
+ "learning_rate": 0.0002451276359600444,
1280
+ "loss": 0.2112,
1281
+ "step": 1850
1282
+ },
1283
+ {
1284
+ "epoch": 3.03,
1285
+ "learning_rate": 0.00024479467258601553,
1286
+ "loss": 0.1914,
1287
+ "step": 1860
1288
+ },
1289
+ {
1290
+ "epoch": 3.04,
1291
+ "learning_rate": 0.0002444617092119866,
1292
+ "loss": 0.1761,
1293
+ "step": 1870
1294
+ },
1295
+ {
1296
+ "epoch": 3.06,
1297
+ "learning_rate": 0.0002441287458379578,
1298
+ "loss": 0.1917,
1299
+ "step": 1880
1300
+ },
1301
+ {
1302
+ "epoch": 3.08,
1303
+ "learning_rate": 0.00024379578246392896,
1304
+ "loss": 0.2106,
1305
+ "step": 1890
1306
+ },
1307
+ {
1308
+ "epoch": 3.09,
1309
+ "learning_rate": 0.0002434628190899001,
1310
+ "loss": 0.2015,
1311
+ "step": 1900
1312
+ },
1313
+ {
1314
+ "epoch": 3.09,
1315
+ "eval_loss": 0.15158045291900635,
1316
+ "eval_runtime": 1057.6306,
1317
+ "eval_samples_per_second": 10.976,
1318
+ "eval_steps_per_second": 0.686,
1319
+ "eval_wer": 0.38587328376590796,
1320
+ "step": 1900
1321
+ },
1322
+ {
1323
+ "epoch": 3.11,
1324
+ "learning_rate": 0.00024312985571587123,
1325
+ "loss": 0.1756,
1326
+ "step": 1910
1327
+ },
1328
+ {
1329
+ "epoch": 3.13,
1330
+ "learning_rate": 0.00024279689234184238,
1331
+ "loss": 0.1733,
1332
+ "step": 1920
1333
+ },
1334
+ {
1335
+ "epoch": 3.14,
1336
+ "learning_rate": 0.0002424639289678135,
1337
+ "loss": 0.1805,
1338
+ "step": 1930
1339
+ },
1340
+ {
1341
+ "epoch": 3.16,
1342
+ "learning_rate": 0.00024213096559378466,
1343
+ "loss": 0.181,
1344
+ "step": 1940
1345
+ },
1346
+ {
1347
+ "epoch": 3.17,
1348
+ "learning_rate": 0.00024179800221975578,
1349
+ "loss": 0.1701,
1350
+ "step": 1950
1351
+ },
1352
+ {
1353
+ "epoch": 3.19,
1354
+ "learning_rate": 0.00024146503884572693,
1355
+ "loss": 0.1894,
1356
+ "step": 1960
1357
+ },
1358
+ {
1359
+ "epoch": 3.21,
1360
+ "learning_rate": 0.0002411320754716981,
1361
+ "loss": 0.1749,
1362
+ "step": 1970
1363
+ },
1364
+ {
1365
+ "epoch": 3.22,
1366
+ "learning_rate": 0.00024079911209766924,
1367
+ "loss": 0.1559,
1368
+ "step": 1980
1369
+ },
1370
+ {
1371
+ "epoch": 3.24,
1372
+ "learning_rate": 0.0002404661487236404,
1373
+ "loss": 0.2001,
1374
+ "step": 1990
1375
+ },
1376
+ {
1377
+ "epoch": 3.26,
1378
+ "learning_rate": 0.0002401331853496115,
1379
+ "loss": 0.1696,
1380
+ "step": 2000
1381
+ },
1382
+ {
1383
+ "epoch": 3.26,
1384
+ "eval_loss": 0.14519302546977997,
1385
+ "eval_runtime": 1046.4919,
1386
+ "eval_samples_per_second": 11.093,
1387
+ "eval_steps_per_second": 0.694,
1388
+ "eval_wer": 0.3826004652647904,
1389
+ "step": 2000
1390
+ },
1391
+ {
1392
+ "epoch": 3.27,
1393
+ "learning_rate": 0.00023980022197558266,
1394
+ "loss": 0.1714,
1395
+ "step": 2010
1396
+ },
1397
+ {
1398
+ "epoch": 3.29,
1399
+ "learning_rate": 0.0002394672586015538,
1400
+ "loss": 0.1527,
1401
+ "step": 2020
1402
+ },
1403
+ {
1404
+ "epoch": 3.31,
1405
+ "learning_rate": 0.00023913429522752494,
1406
+ "loss": 0.1589,
1407
+ "step": 2030
1408
+ },
1409
+ {
1410
+ "epoch": 3.32,
1411
+ "learning_rate": 0.0002388013318534961,
1412
+ "loss": 0.1786,
1413
+ "step": 2040
1414
+ },
1415
+ {
1416
+ "epoch": 3.34,
1417
+ "learning_rate": 0.00023846836847946724,
1418
+ "loss": 0.1958,
1419
+ "step": 2050
1420
+ },
1421
+ {
1422
+ "epoch": 3.35,
1423
+ "learning_rate": 0.0002381354051054384,
1424
+ "loss": 0.1671,
1425
+ "step": 2060
1426
+ },
1427
+ {
1428
+ "epoch": 3.37,
1429
+ "learning_rate": 0.00023780244173140955,
1430
+ "loss": 0.177,
1431
+ "step": 2070
1432
+ },
1433
+ {
1434
+ "epoch": 3.39,
1435
+ "learning_rate": 0.00023746947835738067,
1436
+ "loss": 0.1618,
1437
+ "step": 2080
1438
+ },
1439
+ {
1440
+ "epoch": 3.4,
1441
+ "learning_rate": 0.00023713651498335182,
1442
+ "loss": 0.1859,
1443
+ "step": 2090
1444
+ },
1445
+ {
1446
+ "epoch": 3.42,
1447
+ "learning_rate": 0.00023680355160932295,
1448
+ "loss": 0.1782,
1449
+ "step": 2100
1450
+ },
1451
+ {
1452
+ "epoch": 3.42,
1453
+ "eval_loss": 0.14127187430858612,
1454
+ "eval_runtime": 1047.734,
1455
+ "eval_samples_per_second": 11.08,
1456
+ "eval_steps_per_second": 0.693,
1457
+ "eval_wer": 0.3763285134333805,
1458
+ "step": 2100
1459
+ },
1460
+ {
1461
+ "epoch": 3.44,
1462
+ "learning_rate": 0.0002364705882352941,
1463
+ "loss": 0.1563,
1464
+ "step": 2110
1465
+ },
1466
+ {
1467
+ "epoch": 3.45,
1468
+ "learning_rate": 0.00023613762486126522,
1469
+ "loss": 0.159,
1470
+ "step": 2120
1471
+ },
1472
+ {
1473
+ "epoch": 3.47,
1474
+ "learning_rate": 0.00023580466148723637,
1475
+ "loss": 0.1781,
1476
+ "step": 2130
1477
+ },
1478
+ {
1479
+ "epoch": 3.48,
1480
+ "learning_rate": 0.00023547169811320755,
1481
+ "loss": 0.1841,
1482
+ "step": 2140
1483
+ },
1484
+ {
1485
+ "epoch": 3.5,
1486
+ "learning_rate": 0.00023513873473917868,
1487
+ "loss": 0.1755,
1488
+ "step": 2150
1489
+ },
1490
+ {
1491
+ "epoch": 3.52,
1492
+ "learning_rate": 0.00023480577136514983,
1493
+ "loss": 0.153,
1494
+ "step": 2160
1495
+ },
1496
+ {
1497
+ "epoch": 3.53,
1498
+ "learning_rate": 0.00023447280799112095,
1499
+ "loss": 0.151,
1500
+ "step": 2170
1501
+ },
1502
+ {
1503
+ "epoch": 3.55,
1504
+ "learning_rate": 0.0002341398446170921,
1505
+ "loss": 0.1691,
1506
+ "step": 2180
1507
+ },
1508
+ {
1509
+ "epoch": 3.57,
1510
+ "learning_rate": 0.00023380688124306323,
1511
+ "loss": 0.1685,
1512
+ "step": 2190
1513
+ },
1514
+ {
1515
+ "epoch": 3.58,
1516
+ "learning_rate": 0.00023347391786903438,
1517
+ "loss": 0.1636,
1518
+ "step": 2200
1519
+ },
1520
+ {
1521
+ "epoch": 3.58,
1522
+ "eval_loss": 0.13500617444515228,
1523
+ "eval_runtime": 1049.8895,
1524
+ "eval_samples_per_second": 11.057,
1525
+ "eval_steps_per_second": 0.692,
1526
+ "eval_wer": 0.3761004424576928,
1527
+ "step": 2200
1528
+ },
1529
+ {
1530
+ "epoch": 3.6,
1531
+ "learning_rate": 0.00023314095449500553,
1532
+ "loss": 0.1803,
1533
+ "step": 2210
1534
+ },
1535
+ {
1536
+ "epoch": 3.61,
1537
+ "learning_rate": 0.00023280799112097665,
1538
+ "loss": 0.15,
1539
+ "step": 2220
1540
+ },
1541
+ {
1542
+ "epoch": 3.63,
1543
+ "learning_rate": 0.00023247502774694783,
1544
+ "loss": 0.1721,
1545
+ "step": 2230
1546
+ },
1547
+ {
1548
+ "epoch": 3.65,
1549
+ "learning_rate": 0.00023214206437291896,
1550
+ "loss": 0.1691,
1551
+ "step": 2240
1552
+ },
1553
+ {
1554
+ "epoch": 3.66,
1555
+ "learning_rate": 0.0002318091009988901,
1556
+ "loss": 0.1548,
1557
+ "step": 2250
1558
+ },
1559
+ {
1560
+ "epoch": 3.68,
1561
+ "learning_rate": 0.00023147613762486126,
1562
+ "loss": 0.1616,
1563
+ "step": 2260
1564
+ },
1565
+ {
1566
+ "epoch": 3.7,
1567
+ "learning_rate": 0.00023114317425083238,
1568
+ "loss": 0.159,
1569
+ "step": 2270
1570
+ },
1571
+ {
1572
+ "epoch": 3.71,
1573
+ "learning_rate": 0.00023081021087680353,
1574
+ "loss": 0.1483,
1575
+ "step": 2280
1576
+ },
1577
+ {
1578
+ "epoch": 3.73,
1579
+ "learning_rate": 0.00023047724750277466,
1580
+ "loss": 0.1693,
1581
+ "step": 2290
1582
+ },
1583
+ {
1584
+ "epoch": 3.74,
1585
+ "learning_rate": 0.0002301442841287458,
1586
+ "loss": 0.173,
1587
+ "step": 2300
1588
+ },
1589
+ {
1590
+ "epoch": 3.74,
1591
+ "eval_loss": 0.13234160840511322,
1592
+ "eval_runtime": 1044.2826,
1593
+ "eval_samples_per_second": 11.117,
1594
+ "eval_steps_per_second": 0.695,
1595
+ "eval_wer": 0.3621653058431784,
1596
+ "step": 2300
1597
+ },
1598
+ {
1599
+ "epoch": 3.76,
1600
+ "learning_rate": 0.000229811320754717,
1601
+ "loss": 0.1488,
1602
+ "step": 2310
1603
+ },
1604
+ {
1605
+ "epoch": 3.78,
1606
+ "learning_rate": 0.0002294783573806881,
1607
+ "loss": 0.1491,
1608
+ "step": 2320
1609
+ },
1610
+ {
1611
+ "epoch": 3.79,
1612
+ "learning_rate": 0.00022914539400665926,
1613
+ "loss": 0.1632,
1614
+ "step": 2330
1615
+ },
1616
+ {
1617
+ "epoch": 3.81,
1618
+ "learning_rate": 0.0002288124306326304,
1619
+ "loss": 0.1733,
1620
+ "step": 2340
1621
+ },
1622
+ {
1623
+ "epoch": 3.83,
1624
+ "learning_rate": 0.00022847946725860154,
1625
+ "loss": 0.1648,
1626
+ "step": 2350
1627
+ },
1628
+ {
1629
+ "epoch": 3.84,
1630
+ "learning_rate": 0.00022814650388457266,
1631
+ "loss": 0.1513,
1632
+ "step": 2360
1633
+ },
1634
+ {
1635
+ "epoch": 3.86,
1636
+ "learning_rate": 0.00022781354051054382,
1637
+ "loss": 0.1602,
1638
+ "step": 2370
1639
+ },
1640
+ {
1641
+ "epoch": 3.87,
1642
+ "learning_rate": 0.00022748057713651494,
1643
+ "loss": 0.1528,
1644
+ "step": 2380
1645
+ },
1646
+ {
1647
+ "epoch": 3.89,
1648
+ "learning_rate": 0.0002271476137624861,
1649
+ "loss": 0.1688,
1650
+ "step": 2390
1651
+ },
1652
+ {
1653
+ "epoch": 3.91,
1654
+ "learning_rate": 0.00022681465038845727,
1655
+ "loss": 0.1704,
1656
+ "step": 2400
1657
+ },
1658
+ {
1659
+ "epoch": 3.91,
1660
+ "eval_loss": 0.12890243530273438,
1661
+ "eval_runtime": 1049.8266,
1662
+ "eval_samples_per_second": 11.058,
1663
+ "eval_steps_per_second": 0.692,
1664
+ "eval_wer": 0.364423208502486,
1665
+ "step": 2400
1666
+ },
1667
+ {
1668
+ "epoch": 3.92,
1669
+ "learning_rate": 0.0002264816870144284,
1670
+ "loss": 0.1427,
1671
+ "step": 2410
1672
+ },
1673
+ {
1674
+ "epoch": 3.94,
1675
+ "learning_rate": 0.00022614872364039955,
1676
+ "loss": 0.1585,
1677
+ "step": 2420
1678
+ },
1679
+ {
1680
+ "epoch": 3.96,
1681
+ "learning_rate": 0.00022581576026637067,
1682
+ "loss": 0.1504,
1683
+ "step": 2430
1684
+ },
1685
+ {
1686
+ "epoch": 3.97,
1687
+ "learning_rate": 0.00022548279689234182,
1688
+ "loss": 0.1819,
1689
+ "step": 2440
1690
+ },
1691
+ {
1692
+ "epoch": 3.99,
1693
+ "learning_rate": 0.00022514983351831297,
1694
+ "loss": 0.1679,
1695
+ "step": 2450
1696
+ },
1697
+ {
1698
+ "epoch": 4.0,
1699
+ "learning_rate": 0.0002248168701442841,
1700
+ "loss": 0.1619,
1701
+ "step": 2460
1702
+ },
1703
+ {
1704
+ "epoch": 4.02,
1705
+ "learning_rate": 0.00022448390677025525,
1706
+ "loss": 0.1288,
1707
+ "step": 2470
1708
+ },
1709
+ {
1710
+ "epoch": 4.04,
1711
+ "learning_rate": 0.00022415094339622637,
1712
+ "loss": 0.1288,
1713
+ "step": 2480
1714
+ },
1715
+ {
1716
+ "epoch": 4.05,
1717
+ "learning_rate": 0.00022381798002219755,
1718
+ "loss": 0.1465,
1719
+ "step": 2490
1720
+ },
1721
+ {
1722
+ "epoch": 4.07,
1723
+ "learning_rate": 0.0002234850166481687,
1724
+ "loss": 0.1418,
1725
+ "step": 2500
1726
+ },
1727
+ {
1728
+ "epoch": 4.07,
1729
+ "eval_loss": 0.12660406529903412,
1730
+ "eval_runtime": 1055.0446,
1731
+ "eval_samples_per_second": 11.003,
1732
+ "eval_steps_per_second": 0.688,
1733
+ "eval_wer": 0.34805911599689826,
1734
+ "step": 2500
1735
+ },
1736
+ {
1737
+ "epoch": 4.09,
1738
+ "learning_rate": 0.00022315205327413983,
1739
+ "loss": 0.1618,
1740
+ "step": 2510
1741
+ },
1742
+ {
1743
+ "epoch": 4.1,
1744
+ "learning_rate": 0.00022281908990011098,
1745
+ "loss": 0.1337,
1746
+ "step": 2520
1747
+ },
1748
+ {
1749
+ "epoch": 4.12,
1750
+ "learning_rate": 0.0002224861265260821,
1751
+ "loss": 0.1545,
1752
+ "step": 2530
1753
+ },
1754
+ {
1755
+ "epoch": 4.14,
1756
+ "learning_rate": 0.00022215316315205325,
1757
+ "loss": 0.1343,
1758
+ "step": 2540
1759
+ },
1760
+ {
1761
+ "epoch": 4.15,
1762
+ "learning_rate": 0.00022182019977802438,
1763
+ "loss": 0.1309,
1764
+ "step": 2550
1765
+ },
1766
+ {
1767
+ "epoch": 4.17,
1768
+ "learning_rate": 0.00022148723640399553,
1769
+ "loss": 0.1449,
1770
+ "step": 2560
1771
+ },
1772
+ {
1773
+ "epoch": 4.18,
1774
+ "learning_rate": 0.0002211542730299667,
1775
+ "loss": 0.1214,
1776
+ "step": 2570
1777
+ },
1778
+ {
1779
+ "epoch": 4.2,
1780
+ "learning_rate": 0.00022082130965593783,
1781
+ "loss": 0.1397,
1782
+ "step": 2580
1783
+ },
1784
+ {
1785
+ "epoch": 4.22,
1786
+ "learning_rate": 0.00022048834628190898,
1787
+ "loss": 0.124,
1788
+ "step": 2590
1789
+ },
1790
+ {
1791
+ "epoch": 4.23,
1792
+ "learning_rate": 0.0002201553829078801,
1793
+ "loss": 0.1403,
1794
+ "step": 2600
1795
+ },
1796
+ {
1797
+ "epoch": 4.23,
1798
+ "eval_loss": 0.1273965835571289,
1799
+ "eval_runtime": 1056.7903,
1800
+ "eval_samples_per_second": 10.985,
1801
+ "eval_steps_per_second": 0.687,
1802
+ "eval_wer": 0.34823016922866395,
1803
+ "step": 2600
1804
+ },
1805
+ {
1806
+ "epoch": 4.25,
1807
+ "learning_rate": 0.00021982241953385126,
1808
+ "loss": 0.1616,
1809
+ "step": 2610
1810
+ },
1811
+ {
1812
+ "epoch": 4.27,
1813
+ "learning_rate": 0.00021948945615982238,
1814
+ "loss": 0.1368,
1815
+ "step": 2620
1816
+ },
1817
+ {
1818
+ "epoch": 4.28,
1819
+ "learning_rate": 0.00021915649278579354,
1820
+ "loss": 0.1288,
1821
+ "step": 2630
1822
+ },
1823
+ {
1824
+ "epoch": 4.3,
1825
+ "learning_rate": 0.0002188235294117647,
1826
+ "loss": 0.1299,
1827
+ "step": 2640
1828
+ },
1829
+ {
1830
+ "epoch": 4.31,
1831
+ "learning_rate": 0.0002184905660377358,
1832
+ "loss": 0.1349,
1833
+ "step": 2650
1834
+ },
1835
+ {
1836
+ "epoch": 4.33,
1837
+ "learning_rate": 0.000218157602663707,
1838
+ "loss": 0.166,
1839
+ "step": 2660
1840
+ },
1841
+ {
1842
+ "epoch": 4.35,
1843
+ "learning_rate": 0.00021782463928967811,
1844
+ "loss": 0.1347,
1845
+ "step": 2670
1846
+ },
1847
+ {
1848
+ "epoch": 4.36,
1849
+ "learning_rate": 0.00021749167591564927,
1850
+ "loss": 0.1315,
1851
+ "step": 2680
1852
+ },
1853
+ {
1854
+ "epoch": 4.38,
1855
+ "learning_rate": 0.00021715871254162042,
1856
+ "loss": 0.1311,
1857
+ "step": 2690
1858
+ },
1859
+ {
1860
+ "epoch": 4.4,
1861
+ "learning_rate": 0.00021682574916759154,
1862
+ "loss": 0.1353,
1863
+ "step": 2700
1864
+ },
1865
+ {
1866
+ "epoch": 4.4,
1867
+ "eval_loss": 0.1287122368812561,
1868
+ "eval_runtime": 1067.8021,
1869
+ "eval_samples_per_second": 10.872,
1870
+ "eval_steps_per_second": 0.68,
1871
+ "eval_wer": 0.3488915750581581,
1872
+ "step": 2700
1873
+ },
1874
+ {
1875
+ "epoch": 4.41,
1876
+ "learning_rate": 0.0002164927857935627,
1877
+ "loss": 0.1741,
1878
+ "step": 2710
1879
+ },
1880
+ {
1881
+ "epoch": 4.43,
1882
+ "learning_rate": 0.00021615982241953382,
1883
+ "loss": 0.1331,
1884
+ "step": 2720
1885
+ },
1886
+ {
1887
+ "epoch": 4.44,
1888
+ "learning_rate": 0.00021582685904550497,
1889
+ "loss": 0.1422,
1890
+ "step": 2730
1891
+ },
1892
+ {
1893
+ "epoch": 4.46,
1894
+ "learning_rate": 0.0002154938956714761,
1895
+ "loss": 0.1313,
1896
+ "step": 2740
1897
+ },
1898
+ {
1899
+ "epoch": 4.48,
1900
+ "learning_rate": 0.00021516093229744727,
1901
+ "loss": 0.1508,
1902
+ "step": 2750
1903
+ },
1904
+ {
1905
+ "epoch": 4.49,
1906
+ "learning_rate": 0.00021482796892341842,
1907
+ "loss": 0.1604,
1908
+ "step": 2760
1909
+ },
1910
+ {
1911
+ "epoch": 4.51,
1912
+ "learning_rate": 0.00021449500554938955,
1913
+ "loss": 0.1301,
1914
+ "step": 2770
1915
+ },
1916
+ {
1917
+ "epoch": 4.53,
1918
+ "learning_rate": 0.0002141620421753607,
1919
+ "loss": 0.1432,
1920
+ "step": 2780
1921
+ },
1922
+ {
1923
+ "epoch": 4.54,
1924
+ "learning_rate": 0.00021382907880133182,
1925
+ "loss": 0.1274,
1926
+ "step": 2790
1927
+ },
1928
+ {
1929
+ "epoch": 4.56,
1930
+ "learning_rate": 0.00021349611542730297,
1931
+ "loss": 0.1432,
1932
+ "step": 2800
1933
+ },
1934
+ {
1935
+ "epoch": 4.56,
1936
+ "eval_loss": 0.12929297983646393,
1937
+ "eval_runtime": 1060.4858,
1938
+ "eval_samples_per_second": 10.947,
1939
+ "eval_steps_per_second": 0.685,
1940
+ "eval_wer": 0.35316790585230123,
1941
+ "step": 2800
1942
+ },
1943
+ {
1944
+ "epoch": 4.57,
1945
+ "learning_rate": 0.0002131631520532741,
1946
+ "loss": 0.1652,
1947
+ "step": 2810
1948
+ },
1949
+ {
1950
+ "epoch": 4.59,
1951
+ "learning_rate": 0.00021283018867924525,
1952
+ "loss": 0.1304,
1953
+ "step": 2820
1954
+ },
1955
+ {
1956
+ "epoch": 4.61,
1957
+ "learning_rate": 0.00021249722530521643,
1958
+ "loss": 0.1461,
1959
+ "step": 2830
1960
+ },
1961
+ {
1962
+ "epoch": 4.62,
1963
+ "learning_rate": 0.00021216426193118755,
1964
+ "loss": 0.1342,
1965
+ "step": 2840
1966
+ },
1967
+ {
1968
+ "epoch": 4.64,
1969
+ "learning_rate": 0.0002118312985571587,
1970
+ "loss": 0.1506,
1971
+ "step": 2850
1972
+ },
1973
+ {
1974
+ "epoch": 4.66,
1975
+ "learning_rate": 0.00021149833518312985,
1976
+ "loss": 0.1487,
1977
+ "step": 2860
1978
+ },
1979
+ {
1980
+ "epoch": 4.67,
1981
+ "learning_rate": 0.00021116537180910098,
1982
+ "loss": 0.1291,
1983
+ "step": 2870
1984
+ },
1985
+ {
1986
+ "epoch": 4.69,
1987
+ "learning_rate": 0.00021083240843507213,
1988
+ "loss": 0.1326,
1989
+ "step": 2880
1990
+ },
1991
+ {
1992
+ "epoch": 4.7,
1993
+ "learning_rate": 0.00021049944506104325,
1994
+ "loss": 0.1352,
1995
+ "step": 2890
1996
+ },
1997
+ {
1998
+ "epoch": 4.72,
1999
+ "learning_rate": 0.0002101664816870144,
2000
+ "loss": 0.1283,
2001
+ "step": 2900
2002
+ },
2003
+ {
2004
+ "epoch": 4.72,
2005
+ "eval_loss": 0.12256743013858795,
2006
+ "eval_runtime": 1069.9155,
2007
+ "eval_samples_per_second": 10.85,
2008
+ "eval_steps_per_second": 0.679,
2009
+ "eval_wer": 0.34155909318980066,
2010
+ "step": 2900
2011
+ },
2012
+ {
2013
+ "epoch": 4.74,
2014
+ "learning_rate": 0.00020983351831298553,
2015
+ "loss": 0.1592,
2016
+ "step": 2910
2017
+ },
2018
+ {
2019
+ "epoch": 4.75,
2020
+ "learning_rate": 0.0002095005549389567,
2021
+ "loss": 0.137,
2022
+ "step": 2920
2023
+ },
2024
+ {
2025
+ "epoch": 4.77,
2026
+ "learning_rate": 0.00020916759156492786,
2027
+ "loss": 0.1447,
2028
+ "step": 2930
2029
+ },
2030
+ {
2031
+ "epoch": 4.79,
2032
+ "learning_rate": 0.00020883462819089898,
2033
+ "loss": 0.13,
2034
+ "step": 2940
2035
+ },
2036
+ {
2037
+ "epoch": 4.8,
2038
+ "learning_rate": 0.00020850166481687014,
2039
+ "loss": 0.1336,
2040
+ "step": 2950
2041
+ },
2042
+ {
2043
+ "epoch": 4.82,
2044
+ "learning_rate": 0.00020816870144284126,
2045
+ "loss": 0.1451,
2046
+ "step": 2960
2047
+ },
2048
+ {
2049
+ "epoch": 4.83,
2050
+ "learning_rate": 0.0002078357380688124,
2051
+ "loss": 0.1191,
2052
+ "step": 2970
2053
+ },
2054
+ {
2055
+ "epoch": 4.85,
2056
+ "learning_rate": 0.00020750277469478354,
2057
+ "loss": 0.1266,
2058
+ "step": 2980
2059
+ },
2060
+ {
2061
+ "epoch": 4.87,
2062
+ "learning_rate": 0.0002071698113207547,
2063
+ "loss": 0.1349,
2064
+ "step": 2990
2065
+ },
2066
+ {
2067
+ "epoch": 4.88,
2068
+ "learning_rate": 0.00020683684794672584,
2069
+ "loss": 0.1367,
2070
+ "step": 3000
2071
+ },
2072
+ {
2073
+ "epoch": 4.88,
2074
+ "eval_loss": 0.12064128369092941,
2075
+ "eval_runtime": 1063.0232,
2076
+ "eval_samples_per_second": 10.921,
2077
+ "eval_steps_per_second": 0.683,
2078
+ "eval_wer": 0.3425740090316106,
2079
+ "step": 3000
2080
+ },
2081
+ {
2082
+ "epoch": 4.9,
2083
+ "learning_rate": 0.000206503884572697,
2084
+ "loss": 0.1426,
2085
+ "step": 3010
2086
+ },
2087
+ {
2088
+ "epoch": 4.92,
2089
+ "learning_rate": 0.00020617092119866814,
2090
+ "loss": 0.1221,
2091
+ "step": 3020
2092
+ },
2093
+ {
2094
+ "epoch": 4.93,
2095
+ "learning_rate": 0.00020583795782463927,
2096
+ "loss": 0.1268,
2097
+ "step": 3030
2098
+ },
2099
+ {
2100
+ "epoch": 4.95,
2101
+ "learning_rate": 0.00020550499445061042,
2102
+ "loss": 0.1421,
2103
+ "step": 3040
2104
+ },
2105
+ {
2106
+ "epoch": 4.97,
2107
+ "learning_rate": 0.00020517203107658157,
2108
+ "loss": 0.142,
2109
+ "step": 3050
2110
+ },
2111
+ {
2112
+ "epoch": 4.98,
2113
+ "learning_rate": 0.0002048390677025527,
2114
+ "loss": 0.1659,
2115
+ "step": 3060
2116
+ },
2117
+ {
2118
+ "epoch": 5.0,
2119
+ "learning_rate": 0.00020450610432852384,
2120
+ "loss": 0.1353,
2121
+ "step": 3070
2122
+ },
2123
+ {
2124
+ "epoch": 5.01,
2125
+ "learning_rate": 0.00020417314095449497,
2126
+ "loss": 0.1325,
2127
+ "step": 3080
2128
+ },
2129
+ {
2130
+ "epoch": 5.03,
2131
+ "learning_rate": 0.00020384017758046615,
2132
+ "loss": 0.1191,
2133
+ "step": 3090
2134
+ },
2135
+ {
2136
+ "epoch": 5.05,
2137
+ "learning_rate": 0.0002035072142064373,
2138
+ "loss": 0.1162,
2139
+ "step": 3100
2140
+ },
2141
+ {
2142
+ "epoch": 5.05,
2143
+ "eval_loss": 0.12216832488775253,
2144
+ "eval_runtime": 1074.6528,
2145
+ "eval_samples_per_second": 10.803,
2146
+ "eval_steps_per_second": 0.676,
2147
+ "eval_wer": 0.33939241892076816,
2148
+ "step": 3100
2149
+ },
2150
+ {
2151
+ "epoch": 5.06,
2152
+ "learning_rate": 0.00020317425083240842,
2153
+ "loss": 0.1213,
2154
+ "step": 3110
2155
+ },
2156
+ {
2157
+ "epoch": 5.08,
2158
+ "learning_rate": 0.00020284128745837957,
2159
+ "loss": 0.1631,
2160
+ "step": 3120
2161
+ },
2162
+ {
2163
+ "epoch": 5.1,
2164
+ "learning_rate": 0.0002025083240843507,
2165
+ "loss": 0.1285,
2166
+ "step": 3130
2167
+ },
2168
+ {
2169
+ "epoch": 5.11,
2170
+ "learning_rate": 0.00020217536071032185,
2171
+ "loss": 0.1122,
2172
+ "step": 3140
2173
+ },
2174
+ {
2175
+ "epoch": 5.13,
2176
+ "learning_rate": 0.00020184239733629297,
2177
+ "loss": 0.1202,
2178
+ "step": 3150
2179
+ },
2180
+ {
2181
+ "epoch": 5.14,
2182
+ "learning_rate": 0.00020150943396226413,
2183
+ "loss": 0.1222,
2184
+ "step": 3160
2185
+ },
2186
+ {
2187
+ "epoch": 5.16,
2188
+ "learning_rate": 0.00020117647058823525,
2189
+ "loss": 0.1442,
2190
+ "step": 3170
2191
+ },
2192
+ {
2193
+ "epoch": 5.18,
2194
+ "learning_rate": 0.00020084350721420643,
2195
+ "loss": 0.1197,
2196
+ "step": 3180
2197
+ },
2198
+ {
2199
+ "epoch": 5.19,
2200
+ "learning_rate": 0.00020051054384017758,
2201
+ "loss": 0.1124,
2202
+ "step": 3190
2203
+ },
2204
+ {
2205
+ "epoch": 5.21,
2206
+ "learning_rate": 0.0002001775804661487,
2207
+ "loss": 0.1267,
2208
+ "step": 3200
2209
+ },
2210
+ {
2211
+ "epoch": 5.21,
2212
+ "eval_loss": 0.11833450943231583,
2213
+ "eval_runtime": 1070.4879,
2214
+ "eval_samples_per_second": 10.845,
2215
+ "eval_steps_per_second": 0.678,
2216
+ "eval_wer": 0.33128449573507274,
2217
+ "step": 3200
2218
+ },
2219
+ {
2220
+ "epoch": 5.23,
2221
+ "learning_rate": 0.00019984461709211985,
2222
+ "loss": 0.1266,
2223
+ "step": 3210
2224
+ },
2225
+ {
2226
+ "epoch": 5.24,
2227
+ "learning_rate": 0.00019951165371809098,
2228
+ "loss": 0.1213,
2229
+ "step": 3220
2230
+ },
2231
+ {
2232
+ "epoch": 5.26,
2233
+ "learning_rate": 0.00019917869034406213,
2234
+ "loss": 0.1246,
2235
+ "step": 3230
2236
+ },
2237
+ {
2238
+ "epoch": 5.27,
2239
+ "learning_rate": 0.00019884572697003328,
2240
+ "loss": 0.1117,
2241
+ "step": 3240
2242
+ },
2243
+ {
2244
+ "epoch": 5.29,
2245
+ "learning_rate": 0.0001985127635960044,
2246
+ "loss": 0.1127,
2247
+ "step": 3250
2248
+ },
2249
+ {
2250
+ "epoch": 5.31,
2251
+ "learning_rate": 0.00019817980022197556,
2252
+ "loss": 0.1134,
2253
+ "step": 3260
2254
+ },
2255
+ {
2256
+ "epoch": 5.32,
2257
+ "learning_rate": 0.0001978468368479467,
2258
+ "loss": 0.1505,
2259
+ "step": 3270
2260
+ },
2261
+ {
2262
+ "epoch": 5.34,
2263
+ "learning_rate": 0.00019751387347391786,
2264
+ "loss": 0.1252,
2265
+ "step": 3280
2266
+ },
2267
+ {
2268
+ "epoch": 5.36,
2269
+ "learning_rate": 0.000197180910099889,
2270
+ "loss": 0.1227,
2271
+ "step": 3290
2272
+ },
2273
+ {
2274
+ "epoch": 5.37,
2275
+ "learning_rate": 0.00019684794672586014,
2276
+ "loss": 0.1126,
2277
+ "step": 3300
2278
+ },
2279
+ {
2280
+ "epoch": 5.37,
2281
+ "eval_loss": 0.11798416078090668,
2282
+ "eval_runtime": 1068.0787,
2283
+ "eval_samples_per_second": 10.869,
2284
+ "eval_steps_per_second": 0.68,
2285
+ "eval_wer": 0.32992747342973133,
2286
+ "step": 3300
2287
+ },
2288
+ {
2289
+ "epoch": 5.39,
2290
+ "learning_rate": 0.0001965149833518313,
2291
+ "loss": 0.1238,
2292
+ "step": 3310
2293
+ },
2294
+ {
2295
+ "epoch": 5.4,
2296
+ "learning_rate": 0.0001961820199778024,
2297
+ "loss": 0.1361,
2298
+ "step": 3320
2299
+ },
2300
+ {
2301
+ "epoch": 5.42,
2302
+ "learning_rate": 0.00019584905660377356,
2303
+ "loss": 0.1354,
2304
+ "step": 3330
2305
+ },
2306
+ {
2307
+ "epoch": 5.44,
2308
+ "learning_rate": 0.0001955160932297447,
2309
+ "loss": 0.1225,
2310
+ "step": 3340
2311
+ },
2312
+ {
2313
+ "epoch": 5.45,
2314
+ "learning_rate": 0.00019518312985571587,
2315
+ "loss": 0.1127,
2316
+ "step": 3350
2317
+ },
2318
+ {
2319
+ "epoch": 5.47,
2320
+ "learning_rate": 0.00019485016648168702,
2321
+ "loss": 0.1162,
2322
+ "step": 3360
2323
+ },
2324
+ {
2325
+ "epoch": 5.49,
2326
+ "learning_rate": 0.00019451720310765814,
2327
+ "loss": 0.142,
2328
+ "step": 3370
2329
+ },
2330
+ {
2331
+ "epoch": 5.5,
2332
+ "learning_rate": 0.0001941842397336293,
2333
+ "loss": 0.1239,
2334
+ "step": 3380
2335
+ },
2336
+ {
2337
+ "epoch": 5.52,
2338
+ "learning_rate": 0.00019385127635960042,
2339
+ "loss": 0.1173,
2340
+ "step": 3390
2341
+ },
2342
+ {
2343
+ "epoch": 5.53,
2344
+ "learning_rate": 0.00019351831298557157,
2345
+ "loss": 0.1127,
2346
+ "step": 3400
2347
+ },
2348
+ {
2349
+ "epoch": 5.53,
2350
+ "eval_loss": 0.11773715913295746,
2351
+ "eval_runtime": 1066.7509,
2352
+ "eval_samples_per_second": 10.883,
2353
+ "eval_steps_per_second": 0.681,
2354
+ "eval_wer": 0.330486247320166,
2355
+ "step": 3400
2356
+ },
2357
+ {
2358
+ "epoch": 5.55,
2359
+ "learning_rate": 0.0001931853496115427,
2360
+ "loss": 0.1243,
2361
+ "step": 3410
2362
+ },
2363
+ {
2364
+ "epoch": 5.57,
2365
+ "learning_rate": 0.00019285238623751384,
2366
+ "loss": 0.1427,
2367
+ "step": 3420
2368
+ },
2369
+ {
2370
+ "epoch": 5.58,
2371
+ "learning_rate": 0.000192519422863485,
2372
+ "loss": 0.1182,
2373
+ "step": 3430
2374
+ },
2375
+ {
2376
+ "epoch": 5.6,
2377
+ "learning_rate": 0.00019218645948945615,
2378
+ "loss": 0.1189,
2379
+ "step": 3440
2380
+ },
2381
+ {
2382
+ "epoch": 5.62,
2383
+ "learning_rate": 0.0001918534961154273,
2384
+ "loss": 0.1062,
2385
+ "step": 3450
2386
+ },
2387
+ {
2388
+ "epoch": 5.63,
2389
+ "learning_rate": 0.00019152053274139842,
2390
+ "loss": 0.1192,
2391
+ "step": 3460
2392
+ },
2393
+ {
2394
+ "epoch": 5.65,
2395
+ "learning_rate": 0.00019118756936736957,
2396
+ "loss": 0.1308,
2397
+ "step": 3470
2398
+ },
2399
+ {
2400
+ "epoch": 5.66,
2401
+ "learning_rate": 0.00019085460599334073,
2402
+ "loss": 0.1137,
2403
+ "step": 3480
2404
+ },
2405
+ {
2406
+ "epoch": 5.68,
2407
+ "learning_rate": 0.00019052164261931185,
2408
+ "loss": 0.1087,
2409
+ "step": 3490
2410
+ },
2411
+ {
2412
+ "epoch": 5.7,
2413
+ "learning_rate": 0.000190188679245283,
2414
+ "loss": 0.1155,
2415
+ "step": 3500
2416
+ },
2417
+ {
2418
+ "epoch": 5.7,
2419
+ "eval_loss": 0.11851849406957626,
2420
+ "eval_runtime": 1077.6014,
2421
+ "eval_samples_per_second": 10.773,
2422
+ "eval_steps_per_second": 0.674,
2423
+ "eval_wer": 0.33171783058887927,
2424
+ "step": 3500
2425
+ },
2426
+ {
2427
+ "epoch": 5.71,
2428
+ "learning_rate": 0.00018985571587125413,
2429
+ "loss": 0.1336,
2430
+ "step": 3510
2431
+ },
2432
+ {
2433
+ "epoch": 5.73,
2434
+ "learning_rate": 0.00018952275249722528,
2435
+ "loss": 0.1248,
2436
+ "step": 3520
2437
+ },
2438
+ {
2439
+ "epoch": 5.75,
2440
+ "learning_rate": 0.00018918978912319646,
2441
+ "loss": 0.1247,
2442
+ "step": 3530
2443
+ },
2444
+ {
2445
+ "epoch": 5.76,
2446
+ "learning_rate": 0.00018885682574916758,
2447
+ "loss": 0.1057,
2448
+ "step": 3540
2449
+ },
2450
+ {
2451
+ "epoch": 5.78,
2452
+ "learning_rate": 0.00018852386237513873,
2453
+ "loss": 0.1213,
2454
+ "step": 3550
2455
+ },
2456
+ {
2457
+ "epoch": 5.79,
2458
+ "learning_rate": 0.00018819089900110986,
2459
+ "loss": 0.1122,
2460
+ "step": 3560
2461
+ },
2462
+ {
2463
+ "epoch": 5.81,
2464
+ "learning_rate": 0.000187857935627081,
2465
+ "loss": 0.1251,
2466
+ "step": 3570
2467
+ },
2468
+ {
2469
+ "epoch": 5.83,
2470
+ "learning_rate": 0.00018752497225305213,
2471
+ "loss": 0.1164,
2472
+ "step": 3580
2473
+ },
2474
+ {
2475
+ "epoch": 5.84,
2476
+ "learning_rate": 0.00018719200887902328,
2477
+ "loss": 0.1161,
2478
+ "step": 3590
2479
+ },
2480
+ {
2481
+ "epoch": 5.86,
2482
+ "learning_rate": 0.00018685904550499443,
2483
+ "loss": 0.1086,
2484
+ "step": 3600
2485
+ },
2486
+ {
2487
+ "epoch": 5.86,
2488
+ "eval_loss": 0.11293730139732361,
2489
+ "eval_runtime": 1066.1285,
2490
+ "eval_samples_per_second": 10.889,
2491
+ "eval_steps_per_second": 0.681,
2492
+ "eval_wer": 0.322686219951649,
2493
+ "step": 3600
2494
+ },
2495
+ {
2496
+ "epoch": 5.88,
2497
+ "learning_rate": 0.00018652608213096559,
2498
+ "loss": 0.1214,
2499
+ "step": 3610
2500
+ },
2501
+ {
2502
+ "epoch": 5.89,
2503
+ "learning_rate": 0.00018619311875693674,
2504
+ "loss": 0.1279,
2505
+ "step": 3620
2506
+ },
2507
+ {
2508
+ "epoch": 5.91,
2509
+ "learning_rate": 0.00018586015538290786,
2510
+ "loss": 0.1201,
2511
+ "step": 3630
2512
+ },
2513
+ {
2514
+ "epoch": 5.93,
2515
+ "learning_rate": 0.000185527192008879,
2516
+ "loss": 0.1016,
2517
+ "step": 3640
2518
+ },
2519
+ {
2520
+ "epoch": 5.94,
2521
+ "learning_rate": 0.00018519422863485016,
2522
+ "loss": 0.1116,
2523
+ "step": 3650
2524
+ },
2525
+ {
2526
+ "epoch": 5.96,
2527
+ "learning_rate": 0.0001848612652608213,
2528
+ "loss": 0.1134,
2529
+ "step": 3660
2530
+ },
2531
+ {
2532
+ "epoch": 5.97,
2533
+ "learning_rate": 0.00018452830188679244,
2534
+ "loss": 0.1374,
2535
+ "step": 3670
2536
+ },
2537
+ {
2538
+ "epoch": 5.99,
2539
+ "learning_rate": 0.00018419533851276356,
2540
+ "loss": 0.1045,
2541
+ "step": 3680
2542
+ },
2543
+ {
2544
+ "epoch": 6.01,
2545
+ "learning_rate": 0.00018386237513873471,
2546
+ "loss": 0.1122,
2547
+ "step": 3690
2548
+ },
2549
+ {
2550
+ "epoch": 6.02,
2551
+ "learning_rate": 0.0001835294117647059,
2552
+ "loss": 0.1135,
2553
+ "step": 3700
2554
+ },
2555
+ {
2556
+ "epoch": 6.02,
2557
+ "eval_loss": 0.11180546879768372,
2558
+ "eval_runtime": 1077.7461,
2559
+ "eval_samples_per_second": 10.772,
2560
+ "eval_steps_per_second": 0.674,
2561
+ "eval_wer": 0.32663184783104504,
2562
+ "step": 3700
2563
+ },
2564
+ {
2565
+ "epoch": 6.04,
2566
+ "learning_rate": 0.00018319644839067702,
2567
+ "loss": 0.0972,
2568
+ "step": 3710
2569
+ },
2570
+ {
2571
+ "epoch": 6.06,
2572
+ "learning_rate": 0.00018286348501664817,
2573
+ "loss": 0.1149,
2574
+ "step": 3720
2575
+ },
2576
+ {
2577
+ "epoch": 6.07,
2578
+ "learning_rate": 0.0001825305216426193,
2579
+ "loss": 0.1192,
2580
+ "step": 3730
2581
+ },
2582
+ {
2583
+ "epoch": 6.09,
2584
+ "learning_rate": 0.00018219755826859044,
2585
+ "loss": 0.1311,
2586
+ "step": 3740
2587
+ },
2588
+ {
2589
+ "epoch": 6.1,
2590
+ "learning_rate": 0.00018186459489456157,
2591
+ "loss": 0.1149,
2592
+ "step": 3750
2593
+ },
2594
+ {
2595
+ "epoch": 6.12,
2596
+ "learning_rate": 0.00018153163152053272,
2597
+ "loss": 0.0989,
2598
+ "step": 3760
2599
+ },
2600
+ {
2601
+ "epoch": 6.14,
2602
+ "learning_rate": 0.00018119866814650384,
2603
+ "loss": 0.1192,
2604
+ "step": 3770
2605
+ },
2606
+ {
2607
+ "epoch": 6.15,
2608
+ "learning_rate": 0.000180865704772475,
2609
+ "loss": 0.1011,
2610
+ "step": 3780
2611
+ },
2612
+ {
2613
+ "epoch": 6.17,
2614
+ "learning_rate": 0.00018053274139844617,
2615
+ "loss": 0.1315,
2616
+ "step": 3790
2617
+ },
2618
+ {
2619
+ "epoch": 6.19,
2620
+ "learning_rate": 0.0001801997780244173,
2621
+ "loss": 0.1112,
2622
+ "step": 3800
2623
+ },
2624
+ {
2625
+ "epoch": 6.19,
2626
+ "eval_loss": 0.1141621470451355,
2627
+ "eval_runtime": 1071.2114,
2628
+ "eval_samples_per_second": 10.837,
2629
+ "eval_steps_per_second": 0.678,
2630
+ "eval_wer": 0.32282306253706156,
2631
+ "step": 3800
2632
+ },
2633
+ {
2634
+ "epoch": 6.2,
2635
+ "learning_rate": 0.00017986681465038845,
2636
+ "loss": 0.12,
2637
+ "step": 3810
2638
+ },
2639
+ {
2640
+ "epoch": 6.22,
2641
+ "learning_rate": 0.00017953385127635957,
2642
+ "loss": 0.1038,
2643
+ "step": 3820
2644
+ },
2645
+ {
2646
+ "epoch": 6.23,
2647
+ "learning_rate": 0.00017920088790233073,
2648
+ "loss": 0.1138,
2649
+ "step": 3830
2650
+ },
2651
+ {
2652
+ "epoch": 6.25,
2653
+ "learning_rate": 0.00017886792452830188,
2654
+ "loss": 0.1376,
2655
+ "step": 3840
2656
+ },
2657
+ {
2658
+ "epoch": 6.27,
2659
+ "learning_rate": 0.000178534961154273,
2660
+ "loss": 0.1072,
2661
+ "step": 3850
2662
+ },
2663
+ {
2664
+ "epoch": 6.28,
2665
+ "learning_rate": 0.00017820199778024415,
2666
+ "loss": 0.1063,
2667
+ "step": 3860
2668
+ },
2669
+ {
2670
+ "epoch": 6.3,
2671
+ "learning_rate": 0.0001778690344062153,
2672
+ "loss": 0.1071,
2673
+ "step": 3870
2674
+ },
2675
+ {
2676
+ "epoch": 6.32,
2677
+ "learning_rate": 0.00017753607103218646,
2678
+ "loss": 0.1016,
2679
+ "step": 3880
2680
+ },
2681
+ {
2682
+ "epoch": 6.33,
2683
+ "learning_rate": 0.0001772031076581576,
2684
+ "loss": 0.1246,
2685
+ "step": 3890
2686
+ },
2687
+ {
2688
+ "epoch": 6.35,
2689
+ "learning_rate": 0.00017687014428412873,
2690
+ "loss": 0.0866,
2691
+ "step": 3900
2692
+ },
2693
+ {
2694
+ "epoch": 6.35,
2695
+ "eval_loss": 0.11722169816493988,
2696
+ "eval_runtime": 1077.3899,
2697
+ "eval_samples_per_second": 10.775,
2698
+ "eval_steps_per_second": 0.674,
2699
+ "eval_wer": 0.32837659079505543,
2700
+ "step": 3900
2701
+ },
2702
+ {
2703
+ "epoch": 6.36,
2704
+ "learning_rate": 0.00017653718091009988,
2705
+ "loss": 0.1062,
2706
+ "step": 3910
2707
+ },
2708
+ {
2709
+ "epoch": 6.38,
2710
+ "learning_rate": 0.000176204217536071,
2711
+ "loss": 0.1214,
2712
+ "step": 3920
2713
+ },
2714
+ {
2715
+ "epoch": 6.4,
2716
+ "learning_rate": 0.00017587125416204216,
2717
+ "loss": 0.1138,
2718
+ "step": 3930
2719
+ },
2720
+ {
2721
+ "epoch": 6.41,
2722
+ "learning_rate": 0.00017553829078801328,
2723
+ "loss": 0.1216,
2724
+ "step": 3940
2725
+ },
2726
+ {
2727
+ "epoch": 6.43,
2728
+ "learning_rate": 0.00017520532741398443,
2729
+ "loss": 0.1111,
2730
+ "step": 3950
2731
+ },
2732
+ {
2733
+ "epoch": 6.45,
2734
+ "learning_rate": 0.0001748723640399556,
2735
+ "loss": 0.1044,
2736
+ "step": 3960
2737
+ },
2738
+ {
2739
+ "epoch": 6.46,
2740
+ "learning_rate": 0.00017453940066592674,
2741
+ "loss": 0.1108,
2742
+ "step": 3970
2743
+ },
2744
+ {
2745
+ "epoch": 6.48,
2746
+ "learning_rate": 0.0001742064372918979,
2747
+ "loss": 0.1083,
2748
+ "step": 3980
2749
+ },
2750
+ {
2751
+ "epoch": 6.49,
2752
+ "learning_rate": 0.000173873473917869,
2753
+ "loss": 0.136,
2754
+ "step": 3990
2755
+ },
2756
+ {
2757
+ "epoch": 6.51,
2758
+ "learning_rate": 0.00017354051054384016,
2759
+ "loss": 0.1003,
2760
+ "step": 4000
2761
+ },
2762
+ {
2763
+ "epoch": 6.51,
2764
+ "eval_loss": 0.11332289129495621,
2765
+ "eval_runtime": 1068.9642,
2766
+ "eval_samples_per_second": 10.86,
2767
+ "eval_steps_per_second": 0.679,
2768
+ "eval_wer": 0.32444236646444374,
2769
+ "step": 4000
2770
+ },
2771
+ {
2772
+ "epoch": 6.53,
2773
+ "learning_rate": 0.0001732075471698113,
2774
+ "loss": 6.155,
2775
+ "step": 4010
2776
+ },
2777
+ {
2778
+ "epoch": 6.55,
2779
+ "learning_rate": 0.00017287458379578244,
2780
+ "loss": 3.1349,
2781
+ "step": 4020
2782
+ },
2783
+ {
2784
+ "epoch": 6.56,
2785
+ "learning_rate": 0.0001725416204217536,
2786
+ "loss": 3.0542,
2787
+ "step": 4030
2788
+ },
2789
+ {
2790
+ "epoch": 6.58,
2791
+ "learning_rate": 0.00017220865704772474,
2792
+ "loss": 3.0084,
2793
+ "step": 4040
2794
+ },
2795
+ {
2796
+ "epoch": 6.6,
2797
+ "learning_rate": 0.0001718756936736959,
2798
+ "loss": 2.9937,
2799
+ "step": 4050
2800
+ },
2801
+ {
2802
+ "epoch": 6.61,
2803
+ "learning_rate": 0.00017154273029966702,
2804
+ "loss": 2.9765,
2805
+ "step": 4060
2806
+ },
2807
+ {
2808
+ "epoch": 6.63,
2809
+ "learning_rate": 0.00017120976692563817,
2810
+ "loss": 2.9408,
2811
+ "step": 4070
2812
+ },
2813
+ {
2814
+ "epoch": 6.64,
2815
+ "learning_rate": 0.00017087680355160932,
2816
+ "loss": 2.6233,
2817
+ "step": 4080
2818
+ },
2819
+ {
2820
+ "epoch": 6.66,
2821
+ "learning_rate": 0.00017054384017758045,
2822
+ "loss": 1.4399,
2823
+ "step": 4090
2824
+ },
2825
+ {
2826
+ "epoch": 6.68,
2827
+ "learning_rate": 0.0001702108768035516,
2828
+ "loss": 0.4366,
2829
+ "step": 4100
2830
+ },
2831
+ {
2832
+ "epoch": 6.68,
2833
+ "eval_loss": 0.24362534284591675,
2834
+ "eval_runtime": 1064.9477,
2835
+ "eval_samples_per_second": 10.901,
2836
+ "eval_steps_per_second": 0.682,
2837
+ "eval_wer": 0.4587077498517539,
2838
+ "step": 4100
2839
+ },
2840
+ {
2841
+ "epoch": 6.69,
2842
+ "learning_rate": 0.00016987791342952272,
2843
+ "loss": 0.2313,
2844
+ "step": 4110
2845
+ },
2846
+ {
2847
+ "epoch": 6.71,
2848
+ "learning_rate": 0.00016954495005549387,
2849
+ "loss": 0.1763,
2850
+ "step": 4120
2851
+ },
2852
+ {
2853
+ "epoch": 6.73,
2854
+ "learning_rate": 0.00016921198668146505,
2855
+ "loss": 0.1956,
2856
+ "step": 4130
2857
+ },
2858
+ {
2859
+ "epoch": 6.74,
2860
+ "learning_rate": 0.00016887902330743617,
2861
+ "loss": 0.1825,
2862
+ "step": 4140
2863
+ },
2864
+ {
2865
+ "epoch": 6.76,
2866
+ "learning_rate": 0.00016854605993340733,
2867
+ "loss": 0.1533,
2868
+ "step": 4150
2869
+ },
2870
+ {
2871
+ "epoch": 6.77,
2872
+ "learning_rate": 0.00016821309655937845,
2873
+ "loss": 0.1459,
2874
+ "step": 4160
2875
+ },
2876
+ {
2877
+ "epoch": 6.79,
2878
+ "learning_rate": 0.0001678801331853496,
2879
+ "loss": 0.1547,
2880
+ "step": 4170
2881
+ },
2882
+ {
2883
+ "epoch": 6.81,
2884
+ "learning_rate": 0.00016754716981132073,
2885
+ "loss": 0.1347,
2886
+ "step": 4180
2887
+ },
2888
+ {
2889
+ "epoch": 6.82,
2890
+ "learning_rate": 0.00016721420643729188,
2891
+ "loss": 0.1387,
2892
+ "step": 4190
2893
+ },
2894
+ {
2895
+ "epoch": 6.84,
2896
+ "learning_rate": 0.000166881243063263,
2897
+ "loss": 0.1216,
2898
+ "step": 4200
2899
+ },
2900
+ {
2901
+ "epoch": 6.84,
2902
+ "eval_loss": 0.1344495564699173,
2903
+ "eval_runtime": 1042.0044,
2904
+ "eval_samples_per_second": 11.141,
2905
+ "eval_steps_per_second": 0.697,
2906
+ "eval_wer": 0.33859417050586144,
2907
+ "step": 4200
2908
+ },
2909
+ {
2910
+ "epoch": 6.86,
2911
+ "learning_rate": 0.00016654827968923415,
2912
+ "loss": 0.1096,
2913
+ "step": 4210
2914
+ },
2915
+ {
2916
+ "epoch": 6.87,
2917
+ "learning_rate": 0.00016621531631520533,
2918
+ "loss": 0.1429,
2919
+ "step": 4220
2920
+ },
2921
+ {
2922
+ "epoch": 6.89,
2923
+ "learning_rate": 0.00016588235294117646,
2924
+ "loss": 0.139,
2925
+ "step": 4230
2926
+ },
2927
+ {
2928
+ "epoch": 6.9,
2929
+ "learning_rate": 0.0001655493895671476,
2930
+ "loss": 0.1343,
2931
+ "step": 4240
2932
+ },
2933
+ {
2934
+ "epoch": 6.92,
2935
+ "learning_rate": 0.00016521642619311873,
2936
+ "loss": 0.1223,
2937
+ "step": 4250
2938
+ },
2939
+ {
2940
+ "epoch": 6.94,
2941
+ "learning_rate": 0.00016488346281908988,
2942
+ "loss": 0.1298,
2943
+ "step": 4260
2944
+ },
2945
+ {
2946
+ "epoch": 6.95,
2947
+ "learning_rate": 0.00016455049944506103,
2948
+ "loss": 0.1549,
2949
+ "step": 4270
2950
+ },
2951
+ {
2952
+ "epoch": 6.97,
2953
+ "learning_rate": 0.00016421753607103216,
2954
+ "loss": 0.1378,
2955
+ "step": 4280
2956
+ },
2957
+ {
2958
+ "epoch": 6.99,
2959
+ "learning_rate": 0.0001638845726970033,
2960
+ "loss": 0.1492,
2961
+ "step": 4290
2962
+ },
2963
+ {
2964
+ "epoch": 7.0,
2965
+ "learning_rate": 0.00016355160932297446,
2966
+ "loss": 0.1165,
2967
+ "step": 4300
2968
+ },
2969
+ {
2970
+ "epoch": 7.0,
2971
+ "eval_loss": 0.12804068624973297,
2972
+ "eval_runtime": 1059.2138,
2973
+ "eval_samples_per_second": 10.96,
2974
+ "eval_steps_per_second": 0.685,
2975
+ "eval_wer": 0.3323564293208046,
2976
+ "step": 4300
2977
+ },
2978
+ {
2979
+ "epoch": 7.02,
2980
+ "learning_rate": 0.0001632186459489456,
2981
+ "loss": 0.1136,
2982
+ "step": 4310
2983
+ },
2984
+ {
2985
+ "epoch": 7.03,
2986
+ "learning_rate": 0.00016288568257491676,
2987
+ "loss": 0.1041,
2988
+ "step": 4320
2989
+ },
2990
+ {
2991
+ "epoch": 7.05,
2992
+ "learning_rate": 0.0001625527192008879,
2993
+ "loss": 0.1216,
2994
+ "step": 4330
2995
+ },
2996
+ {
2997
+ "epoch": 7.07,
2998
+ "learning_rate": 0.00016221975582685904,
2999
+ "loss": 0.1156,
3000
+ "step": 4340
3001
+ },
3002
+ {
3003
+ "epoch": 7.08,
3004
+ "learning_rate": 0.00016188679245283016,
3005
+ "loss": 0.1433,
3006
+ "step": 4350
3007
+ },
3008
+ {
3009
+ "epoch": 7.1,
3010
+ "learning_rate": 0.00016155382907880132,
3011
+ "loss": 0.0961,
3012
+ "step": 4360
3013
+ },
3014
+ {
3015
+ "epoch": 7.12,
3016
+ "learning_rate": 0.00016122086570477244,
3017
+ "loss": 0.1292,
3018
+ "step": 4370
3019
+ },
3020
+ {
3021
+ "epoch": 7.13,
3022
+ "learning_rate": 0.0001608879023307436,
3023
+ "loss": 0.1118,
3024
+ "step": 4380
3025
+ },
3026
+ {
3027
+ "epoch": 7.15,
3028
+ "learning_rate": 0.00016055493895671477,
3029
+ "loss": 0.1076,
3030
+ "step": 4390
3031
+ },
3032
+ {
3033
+ "epoch": 7.17,
3034
+ "learning_rate": 0.0001602219755826859,
3035
+ "loss": 0.131,
3036
+ "step": 4400
3037
+ },
3038
+ {
3039
+ "epoch": 7.17,
3040
+ "eval_loss": 0.12520039081573486,
3041
+ "eval_runtime": 1045.0008,
3042
+ "eval_samples_per_second": 11.109,
3043
+ "eval_steps_per_second": 0.695,
3044
+ "eval_wer": 0.3244879806595813,
3045
+ "step": 4400
3046
+ },
3047
+ {
3048
+ "epoch": 7.18,
3049
+ "learning_rate": 0.00015988901220865705,
3050
+ "loss": 0.1108,
3051
+ "step": 4410
3052
+ },
3053
+ {
3054
+ "epoch": 7.2,
3055
+ "learning_rate": 0.00015955604883462817,
3056
+ "loss": 0.1011,
3057
+ "step": 4420
3058
+ },
3059
+ {
3060
+ "epoch": 7.21,
3061
+ "learning_rate": 0.00015922308546059932,
3062
+ "loss": 0.1123,
3063
+ "step": 4430
3064
+ },
3065
+ {
3066
+ "epoch": 7.23,
3067
+ "learning_rate": 0.00015889012208657047,
3068
+ "loss": 0.119,
3069
+ "step": 4440
3070
+ },
3071
+ {
3072
+ "epoch": 7.25,
3073
+ "learning_rate": 0.0001585571587125416,
3074
+ "loss": 0.1318,
3075
+ "step": 4450
3076
+ },
3077
+ {
3078
+ "epoch": 7.26,
3079
+ "learning_rate": 0.00015822419533851275,
3080
+ "loss": 0.1212,
3081
+ "step": 4460
3082
+ },
3083
+ {
3084
+ "epoch": 7.28,
3085
+ "learning_rate": 0.00015789123196448387,
3086
+ "loss": 0.1036,
3087
+ "step": 4470
3088
+ },
3089
+ {
3090
+ "epoch": 7.3,
3091
+ "learning_rate": 0.00015755826859045505,
3092
+ "loss": 0.1113,
3093
+ "step": 4480
3094
+ },
3095
+ {
3096
+ "epoch": 7.31,
3097
+ "learning_rate": 0.0001572253052164262,
3098
+ "loss": 0.1028,
3099
+ "step": 4490
3100
+ },
3101
+ {
3102
+ "epoch": 7.33,
3103
+ "learning_rate": 0.00015689234184239733,
3104
+ "loss": 0.1407,
3105
+ "step": 4500
3106
+ },
3107
+ {
3108
+ "epoch": 7.33,
3109
+ "eval_loss": 0.12339464575052261,
3110
+ "eval_runtime": 1067.1403,
3111
+ "eval_samples_per_second": 10.879,
3112
+ "eval_steps_per_second": 0.68,
3113
+ "eval_wer": 0.32521780778178166,
3114
+ "step": 4500
3115
+ },
3116
+ {
3117
+ "epoch": 7.34,
3118
+ "learning_rate": 0.00015655937846836848,
3119
+ "loss": 0.1006,
3120
+ "step": 4510
3121
+ },
3122
+ {
3123
+ "epoch": 7.36,
3124
+ "learning_rate": 0.0001562264150943396,
3125
+ "loss": 0.099,
3126
+ "step": 4520
3127
+ },
3128
+ {
3129
+ "epoch": 7.38,
3130
+ "learning_rate": 0.00015589345172031075,
3131
+ "loss": 0.1091,
3132
+ "step": 4530
3133
+ },
3134
+ {
3135
+ "epoch": 7.39,
3136
+ "learning_rate": 0.00015556048834628188,
3137
+ "loss": 0.1087,
3138
+ "step": 4540
3139
+ },
3140
+ {
3141
+ "epoch": 7.41,
3142
+ "learning_rate": 0.00015522752497225303,
3143
+ "loss": 0.1429,
3144
+ "step": 4550
3145
+ },
3146
+ {
3147
+ "epoch": 7.43,
3148
+ "learning_rate": 0.0001548945615982242,
3149
+ "loss": 0.1038,
3150
+ "step": 4560
3151
+ },
3152
+ {
3153
+ "epoch": 7.44,
3154
+ "learning_rate": 0.00015456159822419533,
3155
+ "loss": 0.1086,
3156
+ "step": 4570
3157
+ },
3158
+ {
3159
+ "epoch": 7.46,
3160
+ "learning_rate": 0.00015422863485016648,
3161
+ "loss": 0.1144,
3162
+ "step": 4580
3163
+ },
3164
+ {
3165
+ "epoch": 7.47,
3166
+ "learning_rate": 0.0001538956714761376,
3167
+ "loss": 0.1135,
3168
+ "step": 4590
3169
+ },
3170
+ {
3171
+ "epoch": 7.49,
3172
+ "learning_rate": 0.00015356270810210876,
3173
+ "loss": 0.1394,
3174
+ "step": 4600
3175
+ },
3176
+ {
3177
+ "epoch": 7.49,
3178
+ "eval_loss": 0.12083031237125397,
3179
+ "eval_runtime": 1054.1862,
3180
+ "eval_samples_per_second": 11.012,
3181
+ "eval_steps_per_second": 0.689,
3182
+ "eval_wer": 0.3176800620353054,
3183
+ "step": 4600
3184
+ },
3185
+ {
3186
+ "epoch": 7.51,
3187
+ "learning_rate": 0.00015322974472807988,
3188
+ "loss": 0.0979,
3189
+ "step": 4610
3190
+ },
3191
+ {
3192
+ "epoch": 7.52,
3193
+ "learning_rate": 0.00015289678135405103,
3194
+ "loss": 0.1036,
3195
+ "step": 4620
3196
+ },
3197
+ {
3198
+ "epoch": 7.54,
3199
+ "learning_rate": 0.00015256381798002219,
3200
+ "loss": 0.1145,
3201
+ "step": 4630
3202
+ },
3203
+ {
3204
+ "epoch": 7.56,
3205
+ "learning_rate": 0.0001522308546059933,
3206
+ "loss": 0.1173,
3207
+ "step": 4640
3208
+ },
3209
+ {
3210
+ "epoch": 7.57,
3211
+ "learning_rate": 0.0001518978912319645,
3212
+ "loss": 0.1222,
3213
+ "step": 4650
3214
+ },
3215
+ {
3216
+ "epoch": 7.59,
3217
+ "learning_rate": 0.0001515649278579356,
3218
+ "loss": 0.0967,
3219
+ "step": 4660
3220
+ },
3221
+ {
3222
+ "epoch": 7.6,
3223
+ "learning_rate": 0.00015123196448390676,
3224
+ "loss": 0.1034,
3225
+ "step": 4670
3226
+ },
3227
+ {
3228
+ "epoch": 7.62,
3229
+ "learning_rate": 0.00015089900110987792,
3230
+ "loss": 0.1085,
3231
+ "step": 4680
3232
+ },
3233
+ {
3234
+ "epoch": 7.64,
3235
+ "learning_rate": 0.00015056603773584904,
3236
+ "loss": 0.1081,
3237
+ "step": 4690
3238
+ },
3239
+ {
3240
+ "epoch": 7.65,
3241
+ "learning_rate": 0.0001502330743618202,
3242
+ "loss": 0.1449,
3243
+ "step": 4700
3244
+ },
3245
+ {
3246
+ "epoch": 7.65,
3247
+ "eval_loss": 0.11798720061779022,
3248
+ "eval_runtime": 1055.6152,
3249
+ "eval_samples_per_second": 10.997,
3250
+ "eval_steps_per_second": 0.688,
3251
+ "eval_wer": 0.31648268941294533,
3252
+ "step": 4700
3253
+ },
3254
+ {
3255
+ "epoch": 7.67,
3256
+ "learning_rate": 0.00014990011098779132,
3257
+ "loss": 0.0958,
3258
+ "step": 4710
3259
+ },
3260
+ {
3261
+ "epoch": 7.69,
3262
+ "learning_rate": 0.00014956714761376247,
3263
+ "loss": 0.0984,
3264
+ "step": 4720
3265
+ },
3266
+ {
3267
+ "epoch": 7.7,
3268
+ "learning_rate": 0.00014923418423973362,
3269
+ "loss": 0.11,
3270
+ "step": 4730
3271
+ },
3272
+ {
3273
+ "epoch": 7.72,
3274
+ "learning_rate": 0.00014890122086570477,
3275
+ "loss": 0.1064,
3276
+ "step": 4740
3277
+ },
3278
+ {
3279
+ "epoch": 7.73,
3280
+ "learning_rate": 0.0001485682574916759,
3281
+ "loss": 0.1332,
3282
+ "step": 4750
3283
+ },
3284
+ {
3285
+ "epoch": 7.75,
3286
+ "learning_rate": 0.00014823529411764705,
3287
+ "loss": 0.1068,
3288
+ "step": 4760
3289
+ },
3290
+ {
3291
+ "epoch": 7.77,
3292
+ "learning_rate": 0.0001479023307436182,
3293
+ "loss": 0.097,
3294
+ "step": 4770
3295
+ },
3296
+ {
3297
+ "epoch": 7.78,
3298
+ "learning_rate": 0.00014756936736958932,
3299
+ "loss": 0.1058,
3300
+ "step": 4780
3301
+ },
3302
+ {
3303
+ "epoch": 7.8,
3304
+ "learning_rate": 0.00014723640399556047,
3305
+ "loss": 0.1074,
3306
+ "step": 4790
3307
+ },
3308
+ {
3309
+ "epoch": 7.82,
3310
+ "learning_rate": 0.00014690344062153162,
3311
+ "loss": 0.1295,
3312
+ "step": 4800
3313
+ },
3314
+ {
3315
+ "epoch": 7.82,
3316
+ "eval_loss": 0.11703231930732727,
3317
+ "eval_runtime": 1061.0638,
3318
+ "eval_samples_per_second": 10.941,
3319
+ "eval_steps_per_second": 0.684,
3320
+ "eval_wer": 0.3151712813027414,
3321
+ "step": 4800
3322
+ },
3323
+ {
3324
+ "epoch": 7.83,
3325
+ "learning_rate": 0.00014657047724750278,
3326
+ "loss": 0.104,
3327
+ "step": 4810
3328
+ },
3329
+ {
3330
+ "epoch": 7.85,
3331
+ "learning_rate": 0.0001462375138734739,
3332
+ "loss": 0.1086,
3333
+ "step": 4820
3334
+ },
3335
+ {
3336
+ "epoch": 7.86,
3337
+ "learning_rate": 0.00014590455049944505,
3338
+ "loss": 0.106,
3339
+ "step": 4830
3340
+ },
3341
+ {
3342
+ "epoch": 7.88,
3343
+ "learning_rate": 0.00014557158712541618,
3344
+ "loss": 0.1115,
3345
+ "step": 4840
3346
+ },
3347
+ {
3348
+ "epoch": 7.9,
3349
+ "learning_rate": 0.00014523862375138733,
3350
+ "loss": 0.1274,
3351
+ "step": 4850
3352
+ },
3353
+ {
3354
+ "epoch": 7.91,
3355
+ "learning_rate": 0.00014490566037735848,
3356
+ "loss": 0.0944,
3357
+ "step": 4860
3358
+ },
3359
+ {
3360
+ "epoch": 7.93,
3361
+ "learning_rate": 0.00014457269700332963,
3362
+ "loss": 0.1054,
3363
+ "step": 4870
3364
+ },
3365
+ {
3366
+ "epoch": 7.95,
3367
+ "learning_rate": 0.00014423973362930075,
3368
+ "loss": 0.1,
3369
+ "step": 4880
3370
+ },
3371
+ {
3372
+ "epoch": 7.96,
3373
+ "learning_rate": 0.0001439067702552719,
3374
+ "loss": 0.098,
3375
+ "step": 4890
3376
+ },
3377
+ {
3378
+ "epoch": 7.98,
3379
+ "learning_rate": 0.00014357380688124306,
3380
+ "loss": 0.1228,
3381
+ "step": 4900
3382
+ },
3383
+ {
3384
+ "epoch": 7.98,
3385
+ "eval_loss": 0.11823362857103348,
3386
+ "eval_runtime": 1053.6032,
3387
+ "eval_samples_per_second": 11.018,
3388
+ "eval_steps_per_second": 0.689,
3389
+ "eval_wer": 0.31600374036400125,
3390
+ "step": 4900
3391
+ },
3392
+ {
3393
+ "epoch": 8.0,
3394
+ "learning_rate": 0.00014324084350721418,
3395
+ "loss": 0.0909,
3396
+ "step": 4910
3397
+ },
3398
+ {
3399
+ "epoch": 8.01,
3400
+ "learning_rate": 0.00014290788013318533,
3401
+ "loss": 0.1087,
3402
+ "step": 4920
3403
+ },
3404
+ {
3405
+ "epoch": 8.03,
3406
+ "learning_rate": 0.00014257491675915648,
3407
+ "loss": 0.0955,
3408
+ "step": 4930
3409
+ },
3410
+ {
3411
+ "epoch": 8.04,
3412
+ "learning_rate": 0.00014224195338512764,
3413
+ "loss": 0.0875,
3414
+ "step": 4940
3415
+ },
3416
+ {
3417
+ "epoch": 8.06,
3418
+ "learning_rate": 0.00014190899001109876,
3419
+ "loss": 0.0944,
3420
+ "step": 4950
3421
+ },
3422
+ {
3423
+ "epoch": 8.08,
3424
+ "learning_rate": 0.0001415760266370699,
3425
+ "loss": 0.1136,
3426
+ "step": 4960
3427
+ },
3428
+ {
3429
+ "epoch": 8.09,
3430
+ "learning_rate": 0.00014124306326304104,
3431
+ "loss": 0.0962,
3432
+ "step": 4970
3433
+ },
3434
+ {
3435
+ "epoch": 8.11,
3436
+ "learning_rate": 0.00014091009988901221,
3437
+ "loss": 0.0968,
3438
+ "step": 4980
3439
+ },
3440
+ {
3441
+ "epoch": 8.13,
3442
+ "learning_rate": 0.00014057713651498334,
3443
+ "loss": 0.0902,
3444
+ "step": 4990
3445
+ },
3446
+ {
3447
+ "epoch": 8.14,
3448
+ "learning_rate": 0.0001402441731409545,
3449
+ "loss": 0.0913,
3450
+ "step": 5000
3451
+ },
3452
+ {
3453
+ "epoch": 8.14,
3454
+ "eval_loss": 0.11221442371606827,
3455
+ "eval_runtime": 1054.3263,
3456
+ "eval_samples_per_second": 11.011,
3457
+ "eval_steps_per_second": 0.689,
3458
+ "eval_wer": 0.30864845139807506,
3459
+ "step": 5000
3460
+ },
3461
+ {
3462
+ "epoch": 8.16,
3463
+ "learning_rate": 0.00013991120976692561,
3464
+ "loss": 0.0965,
3465
+ "step": 5010
3466
+ },
3467
+ {
3468
+ "epoch": 8.17,
3469
+ "learning_rate": 0.00013957824639289677,
3470
+ "loss": 0.1048,
3471
+ "step": 5020
3472
+ },
3473
+ {
3474
+ "epoch": 8.19,
3475
+ "learning_rate": 0.00013924528301886792,
3476
+ "loss": 0.0965,
3477
+ "step": 5030
3478
+ },
3479
+ {
3480
+ "epoch": 8.21,
3481
+ "learning_rate": 0.00013891231964483904,
3482
+ "loss": 0.0891,
3483
+ "step": 5040
3484
+ },
3485
+ {
3486
+ "epoch": 8.22,
3487
+ "learning_rate": 0.0001385793562708102,
3488
+ "loss": 0.0898,
3489
+ "step": 5050
3490
+ },
3491
+ {
3492
+ "epoch": 8.24,
3493
+ "learning_rate": 0.00013824639289678134,
3494
+ "loss": 0.1044,
3495
+ "step": 5060
3496
+ },
3497
+ {
3498
+ "epoch": 8.26,
3499
+ "learning_rate": 0.0001379134295227525,
3500
+ "loss": 0.1056,
3501
+ "step": 5070
3502
+ },
3503
+ {
3504
+ "epoch": 8.27,
3505
+ "learning_rate": 0.00013758046614872362,
3506
+ "loss": 0.0937,
3507
+ "step": 5080
3508
+ },
3509
+ {
3510
+ "epoch": 8.29,
3511
+ "learning_rate": 0.00013724750277469477,
3512
+ "loss": 0.0898,
3513
+ "step": 5090
3514
+ },
3515
+ {
3516
+ "epoch": 8.3,
3517
+ "learning_rate": 0.0001369145394006659,
3518
+ "loss": 0.1014,
3519
+ "step": 5100
3520
+ },
3521
+ {
3522
+ "epoch": 8.3,
3523
+ "eval_loss": 0.11184623092412949,
3524
+ "eval_runtime": 1060.664,
3525
+ "eval_samples_per_second": 10.945,
3526
+ "eval_steps_per_second": 0.684,
3527
+ "eval_wer": 0.3099826666058477,
3528
+ "step": 5100
3529
+ },
3530
+ {
3531
+ "epoch": 8.32,
3532
+ "learning_rate": 0.00013658157602663707,
3533
+ "loss": 0.0992,
3534
+ "step": 5110
3535
+ },
3536
+ {
3537
+ "epoch": 8.34,
3538
+ "learning_rate": 0.0001362486126526082,
3539
+ "loss": 0.0992,
3540
+ "step": 5120
3541
+ },
3542
+ {
3543
+ "epoch": 8.35,
3544
+ "learning_rate": 0.00013591564927857935,
3545
+ "loss": 0.0922,
3546
+ "step": 5130
3547
+ },
3548
+ {
3549
+ "epoch": 8.37,
3550
+ "learning_rate": 0.00013558268590455047,
3551
+ "loss": 0.0861,
3552
+ "step": 5140
3553
+ },
3554
+ {
3555
+ "epoch": 8.39,
3556
+ "learning_rate": 0.00013524972253052162,
3557
+ "loss": 0.0847,
3558
+ "step": 5150
3559
+ },
3560
+ {
3561
+ "epoch": 8.4,
3562
+ "learning_rate": 0.00013491675915649278,
3563
+ "loss": 0.1016,
3564
+ "step": 5160
3565
+ },
3566
+ {
3567
+ "epoch": 8.42,
3568
+ "learning_rate": 0.00013458379578246393,
3569
+ "loss": 0.0985,
3570
+ "step": 5170
3571
+ },
3572
+ {
3573
+ "epoch": 8.43,
3574
+ "learning_rate": 0.00013425083240843505,
3575
+ "loss": 0.0808,
3576
+ "step": 5180
3577
+ },
3578
+ {
3579
+ "epoch": 8.45,
3580
+ "learning_rate": 0.0001339178690344062,
3581
+ "loss": 0.1095,
3582
+ "step": 5190
3583
+ },
3584
+ {
3585
+ "epoch": 8.47,
3586
+ "learning_rate": 0.00013358490566037735,
3587
+ "loss": 0.0861,
3588
+ "step": 5200
3589
+ },
3590
+ {
3591
+ "epoch": 8.47,
3592
+ "eval_loss": 0.11255118250846863,
3593
+ "eval_runtime": 1046.9073,
3594
+ "eval_samples_per_second": 11.089,
3595
+ "eval_steps_per_second": 0.693,
3596
+ "eval_wer": 0.3073940610317931,
3597
+ "step": 5200
3598
+ }
3599
+ ],
3600
+ "logging_steps": 10,
3601
+ "max_steps": 9210,
3602
+ "num_train_epochs": 15,
3603
+ "save_steps": 200,
3604
+ "total_flos": 5.4544347330059256e+19,
3605
+ "trial_name": null,
3606
+ "trial_params": null
3607
+ }
checkpoint-5200/training_args.bin ADDED
Binary file (4.6 kB). View file