DewiBrynJones
commited on
Commit
•
8f1b3d0
1
Parent(s):
ba38c1c
Gorffenaf/July 2024
Browse files- added_tokens.json +3 -0
- all_results.json +15 -0
- config.json +10 -9
- eval_results.json +9 -0
- kenlm/en/kenlm.tar.gz +0 -3
- kenlm/cy/kenlm.tar.gz → model.safetensors +2 -2
- preprocessor_config.json +1 -1
- pytorch_model.bin +0 -3
- special_tokens_map.json +30 -1
- tokenizer_config.json +48 -1
- train_results.json +9 -0
- trainer_state.json +2110 -0
- training_args.bin +2 -2
- vocab.json +46 -1
added_tokens.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"</s>": 44
|
3 |
+
}
|
all_results.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.8714672861014323,
|
3 |
+
"eval_loss": 0.38884833455085754,
|
4 |
+
"eval_runtime": 156.3404,
|
5 |
+
"eval_samples": 5656,
|
6 |
+
"eval_samples_per_second": 36.177,
|
7 |
+
"eval_steps_per_second": 4.522,
|
8 |
+
"eval_wer": 0.2884883888879973,
|
9 |
+
"total_flos": 2.249387574100498e+19,
|
10 |
+
"train_loss": 0.15996522521972656,
|
11 |
+
"train_runtime": 19346.8732,
|
12 |
+
"train_samples": 41326,
|
13 |
+
"train_samples_per_second": 8.27,
|
14 |
+
"train_steps_per_second": 1.034
|
15 |
+
}
|
config.json
CHANGED
@@ -1,6 +1,7 @@
|
|
1 |
{
|
2 |
"_name_or_path": "facebook/wav2vec2-large-xlsr-53",
|
3 |
-
"activation_dropout": 0.
|
|
|
4 |
"adapter_kernel_size": 3,
|
5 |
"adapter_stride": 2,
|
6 |
"add_adapter": false,
|
@@ -8,7 +9,7 @@
|
|
8 |
"architectures": [
|
9 |
"Wav2Vec2ForCTC"
|
10 |
],
|
11 |
-
"attention_dropout": 0.
|
12 |
"bos_token_id": 1,
|
13 |
"classifier_proj_size": 256,
|
14 |
"codevector_dim": 768,
|
@@ -49,16 +50,16 @@
|
|
49 |
"feat_extract_activation": "gelu",
|
50 |
"feat_extract_dropout": 0.0,
|
51 |
"feat_extract_norm": "layer",
|
52 |
-
"feat_proj_dropout": 0.
|
53 |
"feat_quantizer_dropout": 0.0,
|
54 |
"final_dropout": 0.0,
|
55 |
"hidden_act": "gelu",
|
56 |
-
"hidden_dropout": 0.
|
57 |
"hidden_size": 1024,
|
58 |
"initializer_range": 0.02,
|
59 |
"intermediate_size": 4096,
|
60 |
"layer_norm_eps": 1e-05,
|
61 |
-
"layerdrop": 0.
|
62 |
"mask_channel_length": 10,
|
63 |
"mask_channel_min_space": 1,
|
64 |
"mask_channel_other": 0.0,
|
@@ -71,7 +72,7 @@
|
|
71 |
"mask_time_min_masks": 2,
|
72 |
"mask_time_min_space": 1,
|
73 |
"mask_time_other": 0.0,
|
74 |
-
"mask_time_prob": 0.
|
75 |
"mask_time_selection": "static",
|
76 |
"model_type": "wav2vec2",
|
77 |
"num_adapter_layers": 3,
|
@@ -84,7 +85,7 @@
|
|
84 |
"num_hidden_layers": 24,
|
85 |
"num_negatives": 100,
|
86 |
"output_hidden_size": 1024,
|
87 |
-
"pad_token_id":
|
88 |
"proj_codevector_dim": 768,
|
89 |
"tdnn_dilation": [
|
90 |
1,
|
@@ -108,8 +109,8 @@
|
|
108 |
1
|
109 |
],
|
110 |
"torch_dtype": "float32",
|
111 |
-
"transformers_version": "4.
|
112 |
"use_weighted_layer_sum": false,
|
113 |
-
"vocab_size":
|
114 |
"xvector_output_dim": 512
|
115 |
}
|
|
|
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,
|
|
|
9 |
"architectures": [
|
10 |
"Wav2Vec2ForCTC"
|
11 |
],
|
12 |
+
"attention_dropout": 0.0,
|
13 |
"bos_token_id": 1,
|
14 |
"classifier_proj_size": 256,
|
15 |
"codevector_dim": 768,
|
|
|
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 |
"hidden_act": "gelu",
|
57 |
+
"hidden_dropout": 0.0,
|
58 |
"hidden_size": 1024,
|
59 |
"initializer_range": 0.02,
|
60 |
"intermediate_size": 4096,
|
61 |
"layer_norm_eps": 1e-05,
|
62 |
+
"layerdrop": 0.0,
|
63 |
"mask_channel_length": 10,
|
64 |
"mask_channel_min_space": 1,
|
65 |
"mask_channel_other": 0.0,
|
|
|
72 |
"mask_time_min_masks": 2,
|
73 |
"mask_time_min_space": 1,
|
74 |
"mask_time_other": 0.0,
|
75 |
+
"mask_time_prob": 0.05,
|
76 |
"mask_time_selection": "static",
|
77 |
"model_type": "wav2vec2",
|
78 |
"num_adapter_layers": 3,
|
|
|
85 |
"num_hidden_layers": 24,
|
86 |
"num_negatives": 100,
|
87 |
"output_hidden_size": 1024,
|
88 |
+
"pad_token_id": 42,
|
89 |
"proj_codevector_dim": 768,
|
90 |
"tdnn_dilation": [
|
91 |
1,
|
|
|
109 |
1
|
110 |
],
|
111 |
"torch_dtype": "float32",
|
112 |
+
"transformers_version": "4.40.2",
|
113 |
"use_weighted_layer_sum": false,
|
114 |
+
"vocab_size": 46,
|
115 |
"xvector_output_dim": 512
|
116 |
}
|
eval_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.8714672861014323,
|
3 |
+
"eval_loss": 0.38884833455085754,
|
4 |
+
"eval_runtime": 156.3404,
|
5 |
+
"eval_samples": 5656,
|
6 |
+
"eval_samples_per_second": 36.177,
|
7 |
+
"eval_steps_per_second": 4.522,
|
8 |
+
"eval_wer": 0.2884883888879973
|
9 |
+
}
|
kenlm/en/kenlm.tar.gz
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:cf2d457a6e31282c1e3f0973ae03d0e46aad7b0f9b35ea8fb927692299036189
|
3 |
-
size 909740909
|
|
|
|
|
|
|
|
kenlm/cy/kenlm.tar.gz → model.safetensors
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ccf348c9cd5b591752a4066f0fac73c1bca7151275cd9c3703a0b83922ba2564
|
3 |
+
size 1261996080
|
preprocessor_config.json
CHANGED
@@ -3,7 +3,7 @@
|
|
3 |
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
"feature_size": 1,
|
5 |
"padding_side": "right",
|
6 |
-
"padding_value": 0
|
7 |
"processor_class": "Wav2Vec2Processor",
|
8 |
"return_attention_mask": true,
|
9 |
"sampling_rate": 16000
|
|
|
3 |
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
"feature_size": 1,
|
5 |
"padding_side": "right",
|
6 |
+
"padding_value": 0,
|
7 |
"processor_class": "Wav2Vec2Processor",
|
8 |
"return_attention_mask": true,
|
9 |
"sampling_rate": 16000
|
pytorch_model.bin
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:f35898183d31e3c5e0a3648ef3900ecf8672ab8e45737b0526585f19353e3af2
|
3 |
-
size 1262091377
|
|
|
|
|
|
|
|
special_tokens_map.json
CHANGED
@@ -1 +1,30 @@
|
|
1 |
-
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": false,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"eos_token": {
|
10 |
+
"content": "</s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"pad_token": {
|
17 |
+
"content": "[PAD]",
|
18 |
+
"lstrip": true,
|
19 |
+
"normalized": false,
|
20 |
+
"rstrip": true,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
+
"unk_token": {
|
24 |
+
"content": "[UNK]",
|
25 |
+
"lstrip": true,
|
26 |
+
"normalized": false,
|
27 |
+
"rstrip": true,
|
28 |
+
"single_word": false
|
29 |
+
}
|
30 |
+
}
|
tokenizer_config.json
CHANGED
@@ -1 +1,48 @@
|
|
1 |
-
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"added_tokens_decoder": {
|
3 |
+
"41": {
|
4 |
+
"content": "[UNK]",
|
5 |
+
"lstrip": true,
|
6 |
+
"normalized": false,
|
7 |
+
"rstrip": true,
|
8 |
+
"single_word": false,
|
9 |
+
"special": false
|
10 |
+
},
|
11 |
+
"42": {
|
12 |
+
"content": "[PAD]",
|
13 |
+
"lstrip": true,
|
14 |
+
"normalized": false,
|
15 |
+
"rstrip": true,
|
16 |
+
"single_word": false,
|
17 |
+
"special": false
|
18 |
+
},
|
19 |
+
"43": {
|
20 |
+
"content": "<s>",
|
21 |
+
"lstrip": false,
|
22 |
+
"normalized": false,
|
23 |
+
"rstrip": false,
|
24 |
+
"single_word": false,
|
25 |
+
"special": true
|
26 |
+
},
|
27 |
+
"44": {
|
28 |
+
"content": "</s>",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false,
|
33 |
+
"special": true
|
34 |
+
}
|
35 |
+
},
|
36 |
+
"bos_token": "<s>",
|
37 |
+
"clean_up_tokenization_spaces": true,
|
38 |
+
"do_lower_case": false,
|
39 |
+
"eos_token": "</s>",
|
40 |
+
"model_max_length": 1000000000000000019884624838656,
|
41 |
+
"pad_token": "[PAD]",
|
42 |
+
"processor_class": "Wav2Vec2Processor",
|
43 |
+
"replace_word_delimiter_char": " ",
|
44 |
+
"target_lang": null,
|
45 |
+
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
46 |
+
"unk_token": "[UNK]",
|
47 |
+
"word_delimiter_token": "|"
|
48 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.8714672861014323,
|
3 |
+
"total_flos": 2.249387574100498e+19,
|
4 |
+
"train_loss": 0.15996522521972656,
|
5 |
+
"train_runtime": 19346.8732,
|
6 |
+
"train_samples": 41326,
|
7 |
+
"train_samples_per_second": 8.27,
|
8 |
+
"train_steps_per_second": 1.034
|
9 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,2110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.8714672861014323,
|
5 |
+
"eval_steps": 100,
|
6 |
+
"global_step": 20000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.019357336430507164,
|
13 |
+
"eval_loss": 3.5565404891967773,
|
14 |
+
"eval_runtime": 151.5266,
|
15 |
+
"eval_samples_per_second": 37.327,
|
16 |
+
"eval_steps_per_second": 4.666,
|
17 |
+
"eval_wer": 1.0,
|
18 |
+
"step": 100
|
19 |
+
},
|
20 |
+
{
|
21 |
+
"epoch": 0.03871467286101433,
|
22 |
+
"eval_loss": 3.0301756858825684,
|
23 |
+
"eval_runtime": 150.582,
|
24 |
+
"eval_samples_per_second": 37.561,
|
25 |
+
"eval_steps_per_second": 4.695,
|
26 |
+
"eval_wer": 1.0,
|
27 |
+
"step": 200
|
28 |
+
},
|
29 |
+
{
|
30 |
+
"epoch": 0.05807200929152149,
|
31 |
+
"eval_loss": 2.9460911750793457,
|
32 |
+
"eval_runtime": 148.9065,
|
33 |
+
"eval_samples_per_second": 37.984,
|
34 |
+
"eval_steps_per_second": 4.748,
|
35 |
+
"eval_wer": 1.0,
|
36 |
+
"step": 300
|
37 |
+
},
|
38 |
+
{
|
39 |
+
"epoch": 0.07742934572202866,
|
40 |
+
"eval_loss": 1.8142520189285278,
|
41 |
+
"eval_runtime": 149.8655,
|
42 |
+
"eval_samples_per_second": 37.741,
|
43 |
+
"eval_steps_per_second": 4.718,
|
44 |
+
"eval_wer": 0.940732775914365,
|
45 |
+
"step": 400
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"epoch": 0.09678668215253582,
|
49 |
+
"grad_norm": 3.132490396499634,
|
50 |
+
"learning_rate": 0.00029759999999999997,
|
51 |
+
"loss": 3.9521,
|
52 |
+
"step": 500
|
53 |
+
},
|
54 |
+
{
|
55 |
+
"epoch": 0.09678668215253582,
|
56 |
+
"eval_loss": 1.4195518493652344,
|
57 |
+
"eval_runtime": 150.5171,
|
58 |
+
"eval_samples_per_second": 37.577,
|
59 |
+
"eval_steps_per_second": 4.697,
|
60 |
+
"eval_wer": 0.8693007655149171,
|
61 |
+
"step": 500
|
62 |
+
},
|
63 |
+
{
|
64 |
+
"epoch": 0.11614401858304298,
|
65 |
+
"eval_loss": 1.16689133644104,
|
66 |
+
"eval_runtime": 150.5387,
|
67 |
+
"eval_samples_per_second": 37.572,
|
68 |
+
"eval_steps_per_second": 4.696,
|
69 |
+
"eval_wer": 0.8055239042865626,
|
70 |
+
"step": 600
|
71 |
+
},
|
72 |
+
{
|
73 |
+
"epoch": 0.13550135501355012,
|
74 |
+
"eval_loss": 1.0756505727767944,
|
75 |
+
"eval_runtime": 151.2385,
|
76 |
+
"eval_samples_per_second": 37.398,
|
77 |
+
"eval_steps_per_second": 4.675,
|
78 |
+
"eval_wer": 0.7596251063215163,
|
79 |
+
"step": 700
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.1548586914440573,
|
83 |
+
"eval_loss": 0.9944618344306946,
|
84 |
+
"eval_runtime": 151.1646,
|
85 |
+
"eval_samples_per_second": 37.416,
|
86 |
+
"eval_steps_per_second": 4.677,
|
87 |
+
"eval_wer": 0.7223925149652549,
|
88 |
+
"step": 800
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 0.17421602787456447,
|
92 |
+
"eval_loss": 0.9381263256072998,
|
93 |
+
"eval_runtime": 151.6289,
|
94 |
+
"eval_samples_per_second": 37.302,
|
95 |
+
"eval_steps_per_second": 4.663,
|
96 |
+
"eval_wer": 0.6870857472998347,
|
97 |
+
"step": 900
|
98 |
+
},
|
99 |
+
{
|
100 |
+
"epoch": 0.19357336430507163,
|
101 |
+
"grad_norm": 7.335289001464844,
|
102 |
+
"learning_rate": 0.0002844,
|
103 |
+
"loss": 1.0266,
|
104 |
+
"step": 1000
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 0.19357336430507163,
|
108 |
+
"eval_loss": 0.8977694511413574,
|
109 |
+
"eval_runtime": 156.0202,
|
110 |
+
"eval_samples_per_second": 36.252,
|
111 |
+
"eval_steps_per_second": 4.531,
|
112 |
+
"eval_wer": 0.661472292211648,
|
113 |
+
"step": 1000
|
114 |
+
},
|
115 |
+
{
|
116 |
+
"epoch": 0.2129307007355788,
|
117 |
+
"eval_loss": 0.8770694136619568,
|
118 |
+
"eval_runtime": 151.6589,
|
119 |
+
"eval_samples_per_second": 37.294,
|
120 |
+
"eval_steps_per_second": 4.662,
|
121 |
+
"eval_wer": 0.6450385967164706,
|
122 |
+
"step": 1100
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 0.23228803716608595,
|
126 |
+
"eval_loss": 0.851553201675415,
|
127 |
+
"eval_runtime": 151.5945,
|
128 |
+
"eval_samples_per_second": 37.31,
|
129 |
+
"eval_steps_per_second": 4.664,
|
130 |
+
"eval_wer": 0.640432668389209,
|
131 |
+
"step": 1200
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 0.2516453735965931,
|
135 |
+
"eval_loss": 0.8273979425430298,
|
136 |
+
"eval_runtime": 151.4524,
|
137 |
+
"eval_samples_per_second": 37.345,
|
138 |
+
"eval_steps_per_second": 4.668,
|
139 |
+
"eval_wer": 0.6138081558633307,
|
140 |
+
"step": 1300
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"epoch": 0.27100271002710025,
|
144 |
+
"eval_loss": 0.7992698550224304,
|
145 |
+
"eval_runtime": 152.8076,
|
146 |
+
"eval_samples_per_second": 37.014,
|
147 |
+
"eval_steps_per_second": 4.627,
|
148 |
+
"eval_wer": 0.596973247099228,
|
149 |
+
"step": 1400
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.29036004645760743,
|
153 |
+
"grad_norm": 4.0737223625183105,
|
154 |
+
"learning_rate": 0.00026861052631578947,
|
155 |
+
"loss": 0.8454,
|
156 |
+
"step": 1500
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.29036004645760743,
|
160 |
+
"eval_loss": 0.7768516540527344,
|
161 |
+
"eval_runtime": 152.3743,
|
162 |
+
"eval_samples_per_second": 37.119,
|
163 |
+
"eval_steps_per_second": 4.64,
|
164 |
+
"eval_wer": 0.5887563993516394,
|
165 |
+
"step": 1500
|
166 |
+
},
|
167 |
+
{
|
168 |
+
"epoch": 0.3097173828881146,
|
169 |
+
"eval_loss": 0.7664207220077515,
|
170 |
+
"eval_runtime": 154.3668,
|
171 |
+
"eval_samples_per_second": 36.64,
|
172 |
+
"eval_steps_per_second": 4.58,
|
173 |
+
"eval_wer": 0.5997977885124617,
|
174 |
+
"step": 1600
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 0.32907471931862176,
|
178 |
+
"eval_loss": 0.7400562763214111,
|
179 |
+
"eval_runtime": 153.7228,
|
180 |
+
"eval_samples_per_second": 36.793,
|
181 |
+
"eval_steps_per_second": 4.599,
|
182 |
+
"eval_wer": 0.5592110542279854,
|
183 |
+
"step": 1700
|
184 |
+
},
|
185 |
+
{
|
186 |
+
"epoch": 0.34843205574912894,
|
187 |
+
"eval_loss": 0.746478796005249,
|
188 |
+
"eval_runtime": 151.7535,
|
189 |
+
"eval_samples_per_second": 37.271,
|
190 |
+
"eval_steps_per_second": 4.659,
|
191 |
+
"eval_wer": 0.5650206223620228,
|
192 |
+
"step": 1800
|
193 |
+
},
|
194 |
+
{
|
195 |
+
"epoch": 0.3677893921796361,
|
196 |
+
"eval_loss": 0.7252949476242065,
|
197 |
+
"eval_runtime": 151.7548,
|
198 |
+
"eval_samples_per_second": 37.271,
|
199 |
+
"eval_steps_per_second": 4.659,
|
200 |
+
"eval_wer": 0.5791272808974338,
|
201 |
+
"step": 1900
|
202 |
+
},
|
203 |
+
{
|
204 |
+
"epoch": 0.38714672861014326,
|
205 |
+
"grad_norm": 2.4802448749542236,
|
206 |
+
"learning_rate": 0.0002528210526315789,
|
207 |
+
"loss": 0.7537,
|
208 |
+
"step": 2000
|
209 |
+
},
|
210 |
+
{
|
211 |
+
"epoch": 0.38714672861014326,
|
212 |
+
"eval_loss": 0.7039346098899841,
|
213 |
+
"eval_runtime": 152.7969,
|
214 |
+
"eval_samples_per_second": 37.016,
|
215 |
+
"eval_steps_per_second": 4.627,
|
216 |
+
"eval_wer": 0.5343518800853782,
|
217 |
+
"step": 2000
|
218 |
+
},
|
219 |
+
{
|
220 |
+
"epoch": 0.4065040650406504,
|
221 |
+
"eval_loss": 0.6932350397109985,
|
222 |
+
"eval_runtime": 152.4406,
|
223 |
+
"eval_samples_per_second": 37.103,
|
224 |
+
"eval_steps_per_second": 4.638,
|
225 |
+
"eval_wer": 0.5168429330294811,
|
226 |
+
"step": 2100
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.4258614014711576,
|
230 |
+
"eval_loss": 0.696869432926178,
|
231 |
+
"eval_runtime": 153.0527,
|
232 |
+
"eval_samples_per_second": 36.955,
|
233 |
+
"eval_steps_per_second": 4.619,
|
234 |
+
"eval_wer": 0.5364381890837894,
|
235 |
+
"step": 2200
|
236 |
+
},
|
237 |
+
{
|
238 |
+
"epoch": 0.4452187379016647,
|
239 |
+
"eval_loss": 0.6781283617019653,
|
240 |
+
"eval_runtime": 152.1378,
|
241 |
+
"eval_samples_per_second": 37.177,
|
242 |
+
"eval_steps_per_second": 4.647,
|
243 |
+
"eval_wer": 0.5173725345444624,
|
244 |
+
"step": 2300
|
245 |
+
},
|
246 |
+
{
|
247 |
+
"epoch": 0.4645760743321719,
|
248 |
+
"eval_loss": 0.6760829091072083,
|
249 |
+
"eval_runtime": 151.9712,
|
250 |
+
"eval_samples_per_second": 37.218,
|
251 |
+
"eval_steps_per_second": 4.652,
|
252 |
+
"eval_wer": 0.5050312143923223,
|
253 |
+
"step": 2400
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 0.48393341076267904,
|
257 |
+
"grad_norm": 3.791292667388916,
|
258 |
+
"learning_rate": 0.0002370315789473684,
|
259 |
+
"loss": 0.681,
|
260 |
+
"step": 2500
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 0.48393341076267904,
|
264 |
+
"eval_loss": 0.6720712780952454,
|
265 |
+
"eval_runtime": 152.2414,
|
266 |
+
"eval_samples_per_second": 37.152,
|
267 |
+
"eval_steps_per_second": 4.644,
|
268 |
+
"eval_wer": 0.528718845789668,
|
269 |
+
"step": 2500
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 0.5032907471931862,
|
273 |
+
"eval_loss": 0.6598270535469055,
|
274 |
+
"eval_runtime": 151.7192,
|
275 |
+
"eval_samples_per_second": 37.279,
|
276 |
+
"eval_steps_per_second": 4.66,
|
277 |
+
"eval_wer": 0.5195069891351447,
|
278 |
+
"step": 2600
|
279 |
+
},
|
280 |
+
{
|
281 |
+
"epoch": 0.5226480836236934,
|
282 |
+
"eval_loss": 0.6555168628692627,
|
283 |
+
"eval_runtime": 152.5678,
|
284 |
+
"eval_samples_per_second": 37.072,
|
285 |
+
"eval_steps_per_second": 4.634,
|
286 |
+
"eval_wer": 0.4975846961210701,
|
287 |
+
"step": 2700
|
288 |
+
},
|
289 |
+
{
|
290 |
+
"epoch": 0.5420054200542005,
|
291 |
+
"eval_loss": 0.6535276770591736,
|
292 |
+
"eval_runtime": 152.5246,
|
293 |
+
"eval_samples_per_second": 37.083,
|
294 |
+
"eval_steps_per_second": 4.635,
|
295 |
+
"eval_wer": 0.49936608303509816,
|
296 |
+
"step": 2800
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.5613627564847077,
|
300 |
+
"eval_loss": 0.6258506178855896,
|
301 |
+
"eval_runtime": 151.843,
|
302 |
+
"eval_samples_per_second": 37.249,
|
303 |
+
"eval_steps_per_second": 4.656,
|
304 |
+
"eval_wer": 0.48192133010222915,
|
305 |
+
"step": 2900
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 0.5807200929152149,
|
309 |
+
"grad_norm": 9.4619779586792,
|
310 |
+
"learning_rate": 0.00022124210526315786,
|
311 |
+
"loss": 0.6737,
|
312 |
+
"step": 3000
|
313 |
+
},
|
314 |
+
{
|
315 |
+
"epoch": 0.5807200929152149,
|
316 |
+
"eval_loss": 0.629943311214447,
|
317 |
+
"eval_runtime": 151.8389,
|
318 |
+
"eval_samples_per_second": 37.25,
|
319 |
+
"eval_steps_per_second": 4.656,
|
320 |
+
"eval_wer": 0.48022018584198617,
|
321 |
+
"step": 3000
|
322 |
+
},
|
323 |
+
{
|
324 |
+
"epoch": 0.6000774293457221,
|
325 |
+
"eval_loss": 0.6378594636917114,
|
326 |
+
"eval_runtime": 151.6255,
|
327 |
+
"eval_samples_per_second": 37.302,
|
328 |
+
"eval_steps_per_second": 4.663,
|
329 |
+
"eval_wer": 0.4893197027812104,
|
330 |
+
"step": 3100
|
331 |
+
},
|
332 |
+
{
|
333 |
+
"epoch": 0.6194347657762292,
|
334 |
+
"eval_loss": 0.6225672364234924,
|
335 |
+
"eval_runtime": 153.0144,
|
336 |
+
"eval_samples_per_second": 36.964,
|
337 |
+
"eval_steps_per_second": 4.62,
|
338 |
+
"eval_wer": 0.4806053505801544,
|
339 |
+
"step": 3200
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 0.6387921022067363,
|
343 |
+
"eval_loss": 0.6088670492172241,
|
344 |
+
"eval_runtime": 152.2222,
|
345 |
+
"eval_samples_per_second": 37.156,
|
346 |
+
"eval_steps_per_second": 4.645,
|
347 |
+
"eval_wer": 0.4627112387860891,
|
348 |
+
"step": 3300
|
349 |
+
},
|
350 |
+
{
|
351 |
+
"epoch": 0.6581494386372435,
|
352 |
+
"eval_loss": 0.6028585433959961,
|
353 |
+
"eval_runtime": 153.0615,
|
354 |
+
"eval_samples_per_second": 36.952,
|
355 |
+
"eval_steps_per_second": 4.619,
|
356 |
+
"eval_wer": 0.47354399704707034,
|
357 |
+
"step": 3400
|
358 |
+
},
|
359 |
+
{
|
360 |
+
"epoch": 0.6775067750677507,
|
361 |
+
"grad_norm": 3.4705822467803955,
|
362 |
+
"learning_rate": 0.00020545263157894736,
|
363 |
+
"loss": 0.6419,
|
364 |
+
"step": 3500
|
365 |
+
},
|
366 |
+
{
|
367 |
+
"epoch": 0.6775067750677507,
|
368 |
+
"eval_loss": 0.5871421694755554,
|
369 |
+
"eval_runtime": 152.5739,
|
370 |
+
"eval_samples_per_second": 37.071,
|
371 |
+
"eval_steps_per_second": 4.634,
|
372 |
+
"eval_wer": 0.4592126590810611,
|
373 |
+
"step": 3500
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.6968641114982579,
|
377 |
+
"eval_loss": 0.6001027226448059,
|
378 |
+
"eval_runtime": 152.1697,
|
379 |
+
"eval_samples_per_second": 37.169,
|
380 |
+
"eval_steps_per_second": 4.646,
|
381 |
+
"eval_wer": 0.4610742886488742,
|
382 |
+
"step": 3600
|
383 |
+
},
|
384 |
+
{
|
385 |
+
"epoch": 0.716221447928765,
|
386 |
+
"eval_loss": 0.5848923921585083,
|
387 |
+
"eval_runtime": 152.6563,
|
388 |
+
"eval_samples_per_second": 37.051,
|
389 |
+
"eval_steps_per_second": 4.631,
|
390 |
+
"eval_wer": 0.4472565036670893,
|
391 |
+
"step": 3700
|
392 |
+
},
|
393 |
+
{
|
394 |
+
"epoch": 0.7355787843592722,
|
395 |
+
"eval_loss": 0.5923960208892822,
|
396 |
+
"eval_runtime": 152.6559,
|
397 |
+
"eval_samples_per_second": 37.051,
|
398 |
+
"eval_steps_per_second": 4.631,
|
399 |
+
"eval_wer": 0.46377044181605176,
|
400 |
+
"step": 3800
|
401 |
+
},
|
402 |
+
{
|
403 |
+
"epoch": 0.7549361207897793,
|
404 |
+
"eval_loss": 0.5767965316772461,
|
405 |
+
"eval_runtime": 152.1652,
|
406 |
+
"eval_samples_per_second": 37.17,
|
407 |
+
"eval_steps_per_second": 4.646,
|
408 |
+
"eval_wer": 0.4584904751969957,
|
409 |
+
"step": 3900
|
410 |
+
},
|
411 |
+
{
|
412 |
+
"epoch": 0.7742934572202865,
|
413 |
+
"grad_norm": 3.628082275390625,
|
414 |
+
"learning_rate": 0.00018966315789473683,
|
415 |
+
"loss": 0.6183,
|
416 |
+
"step": 4000
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 0.7742934572202865,
|
420 |
+
"eval_loss": 0.5672534704208374,
|
421 |
+
"eval_runtime": 152.4329,
|
422 |
+
"eval_samples_per_second": 37.105,
|
423 |
+
"eval_steps_per_second": 4.638,
|
424 |
+
"eval_wer": 0.44531463144549116,
|
425 |
+
"step": 4000
|
426 |
+
},
|
427 |
+
{
|
428 |
+
"epoch": 0.7936507936507936,
|
429 |
+
"eval_loss": 0.5575382113456726,
|
430 |
+
"eval_runtime": 152.2388,
|
431 |
+
"eval_samples_per_second": 37.152,
|
432 |
+
"eval_steps_per_second": 4.644,
|
433 |
+
"eval_wer": 0.4451862431994351,
|
434 |
+
"step": 4100
|
435 |
+
},
|
436 |
+
{
|
437 |
+
"epoch": 0.8130081300813008,
|
438 |
+
"eval_loss": 0.5631808042526245,
|
439 |
+
"eval_runtime": 152.7545,
|
440 |
+
"eval_samples_per_second": 37.027,
|
441 |
+
"eval_steps_per_second": 4.628,
|
442 |
+
"eval_wer": 0.4474972316284444,
|
443 |
+
"step": 4200
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 0.832365466511808,
|
447 |
+
"eval_loss": 0.5498641729354858,
|
448 |
+
"eval_runtime": 153.7788,
|
449 |
+
"eval_samples_per_second": 36.78,
|
450 |
+
"eval_steps_per_second": 4.598,
|
451 |
+
"eval_wer": 0.44008281041870617,
|
452 |
+
"step": 4300
|
453 |
+
},
|
454 |
+
{
|
455 |
+
"epoch": 0.8517228029423152,
|
456 |
+
"eval_loss": 0.5662574172019958,
|
457 |
+
"eval_runtime": 152.5034,
|
458 |
+
"eval_samples_per_second": 37.088,
|
459 |
+
"eval_steps_per_second": 4.636,
|
460 |
+
"eval_wer": 0.43101539054099597,
|
461 |
+
"step": 4400
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 0.8710801393728222,
|
465 |
+
"grad_norm": 2.376349925994873,
|
466 |
+
"learning_rate": 0.0001738736842105263,
|
467 |
+
"loss": 0.5877,
|
468 |
+
"step": 4500
|
469 |
+
},
|
470 |
+
{
|
471 |
+
"epoch": 0.8710801393728222,
|
472 |
+
"eval_loss": 0.5584732294082642,
|
473 |
+
"eval_runtime": 152.1714,
|
474 |
+
"eval_samples_per_second": 37.169,
|
475 |
+
"eval_steps_per_second": 4.646,
|
476 |
+
"eval_wer": 0.4317215258943044,
|
477 |
+
"step": 4500
|
478 |
+
},
|
479 |
+
{
|
480 |
+
"epoch": 0.8904374758033294,
|
481 |
+
"eval_loss": 0.5463821291923523,
|
482 |
+
"eval_runtime": 152.4923,
|
483 |
+
"eval_samples_per_second": 37.09,
|
484 |
+
"eval_steps_per_second": 4.636,
|
485 |
+
"eval_wer": 0.41997400138017366,
|
486 |
+
"step": 4600
|
487 |
+
},
|
488 |
+
{
|
489 |
+
"epoch": 0.9097948122338366,
|
490 |
+
"eval_loss": 0.5381494164466858,
|
491 |
+
"eval_runtime": 153.2139,
|
492 |
+
"eval_samples_per_second": 36.916,
|
493 |
+
"eval_steps_per_second": 4.614,
|
494 |
+
"eval_wer": 0.4192197204345942,
|
495 |
+
"step": 4700
|
496 |
+
},
|
497 |
+
{
|
498 |
+
"epoch": 0.9291521486643438,
|
499 |
+
"eval_loss": 0.5453722476959229,
|
500 |
+
"eval_runtime": 151.9737,
|
501 |
+
"eval_samples_per_second": 37.217,
|
502 |
+
"eval_steps_per_second": 4.652,
|
503 |
+
"eval_wer": 0.4201986808107718,
|
504 |
+
"step": 4800
|
505 |
+
},
|
506 |
+
{
|
507 |
+
"epoch": 0.948509485094851,
|
508 |
+
"eval_loss": 0.5237515568733215,
|
509 |
+
"eval_runtime": 151.8558,
|
510 |
+
"eval_samples_per_second": 37.246,
|
511 |
+
"eval_steps_per_second": 4.656,
|
512 |
+
"eval_wer": 0.41241514339362234,
|
513 |
+
"step": 4900
|
514 |
+
},
|
515 |
+
{
|
516 |
+
"epoch": 0.9678668215253581,
|
517 |
+
"grad_norm": 2.5489518642425537,
|
518 |
+
"learning_rate": 0.0001581157894736842,
|
519 |
+
"loss": 0.5621,
|
520 |
+
"step": 5000
|
521 |
+
},
|
522 |
+
{
|
523 |
+
"epoch": 0.9678668215253581,
|
524 |
+
"eval_loss": 0.5303541421890259,
|
525 |
+
"eval_runtime": 152.515,
|
526 |
+
"eval_samples_per_second": 37.085,
|
527 |
+
"eval_steps_per_second": 4.636,
|
528 |
+
"eval_wer": 0.41353854054661293,
|
529 |
+
"step": 5000
|
530 |
+
},
|
531 |
+
{
|
532 |
+
"epoch": 0.9872241579558653,
|
533 |
+
"eval_loss": 0.5163344740867615,
|
534 |
+
"eval_runtime": 156.7945,
|
535 |
+
"eval_samples_per_second": 36.073,
|
536 |
+
"eval_steps_per_second": 4.509,
|
537 |
+
"eval_wer": 0.4061080708061177,
|
538 |
+
"step": 5100
|
539 |
+
},
|
540 |
+
{
|
541 |
+
"epoch": 1.0065814943863725,
|
542 |
+
"eval_loss": 0.51596599817276,
|
543 |
+
"eval_runtime": 153.2891,
|
544 |
+
"eval_samples_per_second": 36.898,
|
545 |
+
"eval_steps_per_second": 4.612,
|
546 |
+
"eval_wer": 0.39927139670363176,
|
547 |
+
"step": 5200
|
548 |
+
},
|
549 |
+
{
|
550 |
+
"epoch": 1.0259388308168795,
|
551 |
+
"eval_loss": 0.5088583827018738,
|
552 |
+
"eval_runtime": 152.7112,
|
553 |
+
"eval_samples_per_second": 37.037,
|
554 |
+
"eval_steps_per_second": 4.63,
|
555 |
+
"eval_wer": 0.3898509091492674,
|
556 |
+
"step": 5300
|
557 |
+
},
|
558 |
+
{
|
559 |
+
"epoch": 1.0452961672473868,
|
560 |
+
"eval_loss": 0.5110610723495483,
|
561 |
+
"eval_runtime": 152.5555,
|
562 |
+
"eval_samples_per_second": 37.075,
|
563 |
+
"eval_steps_per_second": 4.634,
|
564 |
+
"eval_wer": 0.3985652613503234,
|
565 |
+
"step": 5400
|
566 |
+
},
|
567 |
+
{
|
568 |
+
"epoch": 1.064653503677894,
|
569 |
+
"grad_norm": 1.1362248659133911,
|
570 |
+
"learning_rate": 0.0001423578947368421,
|
571 |
+
"loss": 0.4882,
|
572 |
+
"step": 5500
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 1.064653503677894,
|
576 |
+
"eval_loss": 0.5010027885437012,
|
577 |
+
"eval_runtime": 152.1249,
|
578 |
+
"eval_samples_per_second": 37.18,
|
579 |
+
"eval_steps_per_second": 4.647,
|
580 |
+
"eval_wer": 0.38574248527547306,
|
581 |
+
"step": 5500
|
582 |
+
},
|
583 |
+
{
|
584 |
+
"epoch": 1.084010840108401,
|
585 |
+
"eval_loss": 0.49406561255455017,
|
586 |
+
"eval_runtime": 151.5623,
|
587 |
+
"eval_samples_per_second": 37.318,
|
588 |
+
"eval_steps_per_second": 4.665,
|
589 |
+
"eval_wer": 0.3858548249907721,
|
590 |
+
"step": 5600
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 1.1033681765389083,
|
594 |
+
"eval_loss": 0.49403733015060425,
|
595 |
+
"eval_runtime": 152.7631,
|
596 |
+
"eval_samples_per_second": 37.025,
|
597 |
+
"eval_steps_per_second": 4.628,
|
598 |
+
"eval_wer": 0.3813451878480525,
|
599 |
+
"step": 5700
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 1.1227255129694154,
|
603 |
+
"eval_loss": 0.4913772642612457,
|
604 |
+
"eval_runtime": 152.1406,
|
605 |
+
"eval_samples_per_second": 37.176,
|
606 |
+
"eval_steps_per_second": 4.647,
|
607 |
+
"eval_wer": 0.37815153022740766,
|
608 |
+
"step": 5800
|
609 |
+
},
|
610 |
+
{
|
611 |
+
"epoch": 1.1420828493999227,
|
612 |
+
"eval_loss": 0.48747047781944275,
|
613 |
+
"eval_runtime": 151.3195,
|
614 |
+
"eval_samples_per_second": 37.378,
|
615 |
+
"eval_steps_per_second": 4.672,
|
616 |
+
"eval_wer": 0.3745406108070806,
|
617 |
+
"step": 5900
|
618 |
+
},
|
619 |
+
{
|
620 |
+
"epoch": 1.1614401858304297,
|
621 |
+
"grad_norm": 1.0150744915008545,
|
622 |
+
"learning_rate": 0.00012656842105263156,
|
623 |
+
"loss": 0.4569,
|
624 |
+
"step": 6000
|
625 |
+
},
|
626 |
+
{
|
627 |
+
"epoch": 1.1614401858304297,
|
628 |
+
"eval_loss": 0.4841971695423126,
|
629 |
+
"eval_runtime": 151.8567,
|
630 |
+
"eval_samples_per_second": 37.246,
|
631 |
+
"eval_steps_per_second": 4.656,
|
632 |
+
"eval_wer": 0.38071929514852915,
|
633 |
+
"step": 6000
|
634 |
+
},
|
635 |
+
{
|
636 |
+
"epoch": 1.1807975222609368,
|
637 |
+
"eval_loss": 0.48611822724342346,
|
638 |
+
"eval_runtime": 150.971,
|
639 |
+
"eval_samples_per_second": 37.464,
|
640 |
+
"eval_steps_per_second": 4.683,
|
641 |
+
"eval_wer": 0.37370608720771614,
|
642 |
+
"step": 6100
|
643 |
+
},
|
644 |
+
{
|
645 |
+
"epoch": 1.2001548586914441,
|
646 |
+
"eval_loss": 0.48144644498825073,
|
647 |
+
"eval_runtime": 151.4548,
|
648 |
+
"eval_samples_per_second": 37.344,
|
649 |
+
"eval_steps_per_second": 4.668,
|
650 |
+
"eval_wer": 0.3760973182905105,
|
651 |
+
"step": 6200
|
652 |
+
},
|
653 |
+
{
|
654 |
+
"epoch": 1.2195121951219512,
|
655 |
+
"eval_loss": 0.47813892364501953,
|
656 |
+
"eval_runtime": 151.1935,
|
657 |
+
"eval_samples_per_second": 37.409,
|
658 |
+
"eval_steps_per_second": 4.676,
|
659 |
+
"eval_wer": 0.37409125194588433,
|
660 |
+
"step": 6300
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 1.2388695315524583,
|
664 |
+
"eval_loss": 0.4771001935005188,
|
665 |
+
"eval_runtime": 151.1732,
|
666 |
+
"eval_samples_per_second": 37.414,
|
667 |
+
"eval_steps_per_second": 4.677,
|
668 |
+
"eval_wer": 0.36815329556579096,
|
669 |
+
"step": 6400
|
670 |
+
},
|
671 |
+
{
|
672 |
+
"epoch": 1.2582268679829656,
|
673 |
+
"grad_norm": 1.3292571306228638,
|
674 |
+
"learning_rate": 0.00011077894736842105,
|
675 |
+
"loss": 0.4416,
|
676 |
+
"step": 6500
|
677 |
+
},
|
678 |
+
{
|
679 |
+
"epoch": 1.2582268679829656,
|
680 |
+
"eval_loss": 0.47095027565956116,
|
681 |
+
"eval_runtime": 151.5037,
|
682 |
+
"eval_samples_per_second": 37.332,
|
683 |
+
"eval_steps_per_second": 4.667,
|
684 |
+
"eval_wer": 0.37338511659257595,
|
685 |
+
"step": 6500
|
686 |
+
},
|
687 |
+
{
|
688 |
+
"epoch": 1.2775842044134726,
|
689 |
+
"eval_loss": 0.47211408615112305,
|
690 |
+
"eval_runtime": 150.9455,
|
691 |
+
"eval_samples_per_second": 37.47,
|
692 |
+
"eval_steps_per_second": 4.684,
|
693 |
+
"eval_wer": 0.3659706953828377,
|
694 |
+
"step": 6600
|
695 |
+
},
|
696 |
+
{
|
697 |
+
"epoch": 1.29694154084398,
|
698 |
+
"eval_loss": 0.4679400622844696,
|
699 |
+
"eval_runtime": 151.4191,
|
700 |
+
"eval_samples_per_second": 37.353,
|
701 |
+
"eval_steps_per_second": 4.669,
|
702 |
+
"eval_wer": 0.3638843863844265,
|
703 |
+
"step": 6700
|
704 |
+
},
|
705 |
+
{
|
706 |
+
"epoch": 1.316298877274487,
|
707 |
+
"eval_loss": 0.46228036284446716,
|
708 |
+
"eval_runtime": 151.3839,
|
709 |
+
"eval_samples_per_second": 37.362,
|
710 |
+
"eval_steps_per_second": 4.67,
|
711 |
+
"eval_wer": 0.366532393959333,
|
712 |
+
"step": 6800
|
713 |
+
},
|
714 |
+
{
|
715 |
+
"epoch": 1.3356562137049943,
|
716 |
+
"eval_loss": 0.46108925342559814,
|
717 |
+
"eval_runtime": 151.8163,
|
718 |
+
"eval_samples_per_second": 37.256,
|
719 |
+
"eval_steps_per_second": 4.657,
|
720 |
+
"eval_wer": 0.3601771757795574,
|
721 |
+
"step": 6900
|
722 |
+
},
|
723 |
+
{
|
724 |
+
"epoch": 1.3550135501355014,
|
725 |
+
"grad_norm": 0.8062695860862732,
|
726 |
+
"learning_rate": 9.498947368421052e-05,
|
727 |
+
"loss": 0.4324,
|
728 |
+
"step": 7000
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 1.3550135501355014,
|
732 |
+
"eval_loss": 0.46888086199760437,
|
733 |
+
"eval_runtime": 152.4379,
|
734 |
+
"eval_samples_per_second": 37.104,
|
735 |
+
"eval_steps_per_second": 4.638,
|
736 |
+
"eval_wer": 0.3609314567251368,
|
737 |
+
"step": 7000
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 1.3743708865660085,
|
741 |
+
"eval_loss": 0.4573034346103668,
|
742 |
+
"eval_runtime": 151.3077,
|
743 |
+
"eval_samples_per_second": 37.381,
|
744 |
+
"eval_steps_per_second": 4.673,
|
745 |
+
"eval_wer": 0.3602574184333424,
|
746 |
+
"step": 7100
|
747 |
+
},
|
748 |
+
{
|
749 |
+
"epoch": 1.3937282229965158,
|
750 |
+
"eval_loss": 0.45749789476394653,
|
751 |
+
"eval_runtime": 151.5824,
|
752 |
+
"eval_samples_per_second": 37.313,
|
753 |
+
"eval_steps_per_second": 4.664,
|
754 |
+
"eval_wer": 0.3546083356068752,
|
755 |
+
"step": 7200
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 1.4130855594270229,
|
759 |
+
"eval_loss": 0.4555954933166504,
|
760 |
+
"eval_runtime": 151.6035,
|
761 |
+
"eval_samples_per_second": 37.308,
|
762 |
+
"eval_steps_per_second": 4.663,
|
763 |
+
"eval_wer": 0.35836369180401534,
|
764 |
+
"step": 7300
|
765 |
+
},
|
766 |
+
{
|
767 |
+
"epoch": 1.43244289585753,
|
768 |
+
"eval_loss": 0.4495578408241272,
|
769 |
+
"eval_runtime": 152.5621,
|
770 |
+
"eval_samples_per_second": 37.073,
|
771 |
+
"eval_steps_per_second": 4.634,
|
772 |
+
"eval_wer": 0.350724591163679,
|
773 |
+
"step": 7400
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 1.4518002322880372,
|
777 |
+
"grad_norm": 0.7916799187660217,
|
778 |
+
"learning_rate": 7.92e-05,
|
779 |
+
"loss": 0.4255,
|
780 |
+
"step": 7500
|
781 |
+
},
|
782 |
+
{
|
783 |
+
"epoch": 1.4518002322880372,
|
784 |
+
"eval_loss": 0.44609567523002625,
|
785 |
+
"eval_runtime": 151.8498,
|
786 |
+
"eval_samples_per_second": 37.247,
|
787 |
+
"eval_steps_per_second": 4.656,
|
788 |
+
"eval_wer": 0.34671245847442667,
|
789 |
+
"step": 7500
|
790 |
+
},
|
791 |
+
{
|
792 |
+
"epoch": 1.4711575687185443,
|
793 |
+
"eval_loss": 0.44341230392456055,
|
794 |
+
"eval_runtime": 152.528,
|
795 |
+
"eval_samples_per_second": 37.082,
|
796 |
+
"eval_steps_per_second": 4.635,
|
797 |
+
"eval_wer": 0.3462470510824734,
|
798 |
+
"step": 7600
|
799 |
+
},
|
800 |
+
{
|
801 |
+
"epoch": 1.4905149051490514,
|
802 |
+
"eval_loss": 0.44362780451774597,
|
803 |
+
"eval_runtime": 152.5253,
|
804 |
+
"eval_samples_per_second": 37.082,
|
805 |
+
"eval_steps_per_second": 4.635,
|
806 |
+
"eval_wer": 0.3516393574168285,
|
807 |
+
"step": 7700
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 1.5098722415795587,
|
811 |
+
"eval_loss": 0.4406072199344635,
|
812 |
+
"eval_runtime": 152.4039,
|
813 |
+
"eval_samples_per_second": 37.112,
|
814 |
+
"eval_steps_per_second": 4.639,
|
815 |
+
"eval_wer": 0.34579769222127715,
|
816 |
+
"step": 7800
|
817 |
+
},
|
818 |
+
{
|
819 |
+
"epoch": 1.5292295780100658,
|
820 |
+
"eval_loss": 0.43874725699424744,
|
821 |
+
"eval_runtime": 152.6604,
|
822 |
+
"eval_samples_per_second": 37.05,
|
823 |
+
"eval_steps_per_second": 4.631,
|
824 |
+
"eval_wer": 0.3439360626534641,
|
825 |
+
"step": 7900
|
826 |
+
},
|
827 |
+
{
|
828 |
+
"epoch": 1.5485869144405728,
|
829 |
+
"grad_norm": 0.7491864562034607,
|
830 |
+
"learning_rate": 6.344210526315788e-05,
|
831 |
+
"loss": 0.4094,
|
832 |
+
"step": 8000
|
833 |
+
},
|
834 |
+
{
|
835 |
+
"epoch": 1.5485869144405728,
|
836 |
+
"eval_loss": 0.43253499269485474,
|
837 |
+
"eval_runtime": 153.8006,
|
838 |
+
"eval_samples_per_second": 36.775,
|
839 |
+
"eval_steps_per_second": 4.597,
|
840 |
+
"eval_wer": 0.3409831329941744,
|
841 |
+
"step": 8000
|
842 |
+
},
|
843 |
+
{
|
844 |
+
"epoch": 1.5679442508710801,
|
845 |
+
"eval_loss": 0.4359830617904663,
|
846 |
+
"eval_runtime": 153.3674,
|
847 |
+
"eval_samples_per_second": 36.879,
|
848 |
+
"eval_steps_per_second": 4.61,
|
849 |
+
"eval_wer": 0.3419299963088379,
|
850 |
+
"step": 8100
|
851 |
+
},
|
852 |
+
{
|
853 |
+
"epoch": 1.5873015873015874,
|
854 |
+
"eval_loss": 0.4285949170589447,
|
855 |
+
"eval_runtime": 153.3711,
|
856 |
+
"eval_samples_per_second": 36.878,
|
857 |
+
"eval_steps_per_second": 4.61,
|
858 |
+
"eval_wer": 0.3377252812505015,
|
859 |
+
"step": 8200
|
860 |
+
},
|
861 |
+
{
|
862 |
+
"epoch": 1.6066589237320945,
|
863 |
+
"eval_loss": 0.43007034063339233,
|
864 |
+
"eval_runtime": 152.2201,
|
865 |
+
"eval_samples_per_second": 37.157,
|
866 |
+
"eval_steps_per_second": 4.645,
|
867 |
+
"eval_wer": 0.3335526632536791,
|
868 |
+
"step": 8300
|
869 |
+
},
|
870 |
+
{
|
871 |
+
"epoch": 1.6260162601626016,
|
872 |
+
"eval_loss": 0.42966797947883606,
|
873 |
+
"eval_runtime": 152.0163,
|
874 |
+
"eval_samples_per_second": 37.207,
|
875 |
+
"eval_steps_per_second": 4.651,
|
876 |
+
"eval_wer": 0.3322848293238754,
|
877 |
+
"step": 8400
|
878 |
+
},
|
879 |
+
{
|
880 |
+
"epoch": 1.645373596593109,
|
881 |
+
"grad_norm": 1.047472596168518,
|
882 |
+
"learning_rate": 4.765263157894736e-05,
|
883 |
+
"loss": 0.4018,
|
884 |
+
"step": 8500
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 1.645373596593109,
|
888 |
+
"eval_loss": 0.4270441234111786,
|
889 |
+
"eval_runtime": 152.8058,
|
890 |
+
"eval_samples_per_second": 37.014,
|
891 |
+
"eval_steps_per_second": 4.627,
|
892 |
+
"eval_wer": 0.3338575853380623,
|
893 |
+
"step": 8500
|
894 |
+
},
|
895 |
+
{
|
896 |
+
"epoch": 1.664730933023616,
|
897 |
+
"eval_loss": 0.4267289638519287,
|
898 |
+
"eval_runtime": 152.5032,
|
899 |
+
"eval_samples_per_second": 37.088,
|
900 |
+
"eval_steps_per_second": 4.636,
|
901 |
+
"eval_wer": 0.3319959557702492,
|
902 |
+
"step": 8600
|
903 |
+
},
|
904 |
+
{
|
905 |
+
"epoch": 1.684088269454123,
|
906 |
+
"eval_loss": 0.4224300980567932,
|
907 |
+
"eval_runtime": 152.5862,
|
908 |
+
"eval_samples_per_second": 37.068,
|
909 |
+
"eval_steps_per_second": 4.633,
|
910 |
+
"eval_wer": 0.33275023671582865,
|
911 |
+
"step": 8700
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 1.7034456058846303,
|
915 |
+
"eval_loss": 0.4207303822040558,
|
916 |
+
"eval_runtime": 154.5205,
|
917 |
+
"eval_samples_per_second": 36.604,
|
918 |
+
"eval_steps_per_second": 4.575,
|
919 |
+
"eval_wer": 0.32984545264881,
|
920 |
+
"step": 8800
|
921 |
+
},
|
922 |
+
{
|
923 |
+
"epoch": 1.7228029423151374,
|
924 |
+
"eval_loss": 0.4197385013103485,
|
925 |
+
"eval_runtime": 152.0624,
|
926 |
+
"eval_samples_per_second": 37.195,
|
927 |
+
"eval_steps_per_second": 4.649,
|
928 |
+
"eval_wer": 0.32978125852578194,
|
929 |
+
"step": 8900
|
930 |
+
},
|
931 |
+
{
|
932 |
+
"epoch": 1.7421602787456445,
|
933 |
+
"grad_norm": 1.4507739543914795,
|
934 |
+
"learning_rate": 3.189473684210526e-05,
|
935 |
+
"loss": 0.3899,
|
936 |
+
"step": 9000
|
937 |
+
},
|
938 |
+
{
|
939 |
+
"epoch": 1.7421602787456445,
|
940 |
+
"eval_loss": 0.4183507561683655,
|
941 |
+
"eval_runtime": 157.4278,
|
942 |
+
"eval_samples_per_second": 35.928,
|
943 |
+
"eval_steps_per_second": 4.491,
|
944 |
+
"eval_wer": 0.3258493684903147,
|
945 |
+
"step": 9000
|
946 |
+
},
|
947 |
+
{
|
948 |
+
"epoch": 1.7615176151761518,
|
949 |
+
"eval_loss": 0.4164830148220062,
|
950 |
+
"eval_runtime": 153.0475,
|
951 |
+
"eval_samples_per_second": 36.956,
|
952 |
+
"eval_steps_per_second": 4.619,
|
953 |
+
"eval_wer": 0.3262024361669689,
|
954 |
+
"step": 9100
|
955 |
+
},
|
956 |
+
{
|
957 |
+
"epoch": 1.7808749516066589,
|
958 |
+
"eval_loss": 0.41182050108909607,
|
959 |
+
"eval_runtime": 152.4839,
|
960 |
+
"eval_samples_per_second": 37.092,
|
961 |
+
"eval_steps_per_second": 4.637,
|
962 |
+
"eval_wer": 0.322864341769511,
|
963 |
+
"step": 9200
|
964 |
+
},
|
965 |
+
{
|
966 |
+
"epoch": 1.800232288037166,
|
967 |
+
"eval_loss": 0.4134317636489868,
|
968 |
+
"eval_runtime": 152.6353,
|
969 |
+
"eval_samples_per_second": 37.056,
|
970 |
+
"eval_steps_per_second": 4.632,
|
971 |
+
"eval_wer": 0.3232334579769222,
|
972 |
+
"step": 9300
|
973 |
+
},
|
974 |
+
{
|
975 |
+
"epoch": 1.8195896244676733,
|
976 |
+
"eval_loss": 0.4126824736595154,
|
977 |
+
"eval_runtime": 152.5246,
|
978 |
+
"eval_samples_per_second": 37.083,
|
979 |
+
"eval_steps_per_second": 4.635,
|
980 |
+
"eval_wer": 0.3209064210171559,
|
981 |
+
"step": 9400
|
982 |
+
},
|
983 |
+
{
|
984 |
+
"epoch": 1.8389469608981805,
|
985 |
+
"grad_norm": 1.0012460947036743,
|
986 |
+
"learning_rate": 1.6105263157894736e-05,
|
987 |
+
"loss": 0.3665,
|
988 |
+
"step": 9500
|
989 |
+
},
|
990 |
+
{
|
991 |
+
"epoch": 1.8389469608981805,
|
992 |
+
"eval_loss": 0.41083237528800964,
|
993 |
+
"eval_runtime": 152.9993,
|
994 |
+
"eval_samples_per_second": 36.967,
|
995 |
+
"eval_steps_per_second": 4.621,
|
996 |
+
"eval_wer": 0.32109900338624,
|
997 |
+
"step": 9500
|
998 |
+
},
|
999 |
+
{
|
1000 |
+
"epoch": 1.8583042973286876,
|
1001 |
+
"eval_loss": 0.4090138077735901,
|
1002 |
+
"eval_runtime": 152.5291,
|
1003 |
+
"eval_samples_per_second": 37.081,
|
1004 |
+
"eval_steps_per_second": 4.635,
|
1005 |
+
"eval_wer": 0.3199114121102213,
|
1006 |
+
"step": 9600
|
1007 |
+
},
|
1008 |
+
{
|
1009 |
+
"epoch": 1.8776616337591947,
|
1010 |
+
"eval_loss": 0.407578706741333,
|
1011 |
+
"eval_runtime": 153.0711,
|
1012 |
+
"eval_samples_per_second": 36.95,
|
1013 |
+
"eval_steps_per_second": 4.619,
|
1014 |
+
"eval_wer": 0.32087432395564186,
|
1015 |
+
"step": 9700
|
1016 |
+
},
|
1017 |
+
{
|
1018 |
+
"epoch": 1.897018970189702,
|
1019 |
+
"eval_loss": 0.40649694204330444,
|
1020 |
+
"eval_runtime": 154.4136,
|
1021 |
+
"eval_samples_per_second": 36.629,
|
1022 |
+
"eval_steps_per_second": 4.579,
|
1023 |
+
"eval_wer": 0.31981512092567926,
|
1024 |
+
"step": 9800
|
1025 |
+
},
|
1026 |
+
{
|
1027 |
+
"epoch": 1.916376306620209,
|
1028 |
+
"eval_loss": 0.40620651841163635,
|
1029 |
+
"eval_runtime": 153.7508,
|
1030 |
+
"eval_samples_per_second": 36.787,
|
1031 |
+
"eval_steps_per_second": 4.598,
|
1032 |
+
"eval_wer": 0.31923737381842693,
|
1033 |
+
"step": 9900
|
1034 |
+
},
|
1035 |
+
{
|
1036 |
+
"epoch": 1.9357336430507162,
|
1037 |
+
"grad_norm": 0.7244949340820312,
|
1038 |
+
"learning_rate": 3.157894736842105e-07,
|
1039 |
+
"loss": 0.3698,
|
1040 |
+
"step": 10000
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 1.9357336430507162,
|
1044 |
+
"eval_loss": 0.4060620963573456,
|
1045 |
+
"eval_runtime": 153.976,
|
1046 |
+
"eval_samples_per_second": 36.733,
|
1047 |
+
"eval_steps_per_second": 4.592,
|
1048 |
+
"eval_wer": 0.31928551941069794,
|
1049 |
+
"step": 10000
|
1050 |
+
},
|
1051 |
+
{
|
1052 |
+
"epoch": 1.9550909794812235,
|
1053 |
+
"eval_loss": 0.45229342579841614,
|
1054 |
+
"eval_runtime": 154.2948,
|
1055 |
+
"eval_samples_per_second": 36.657,
|
1056 |
+
"eval_steps_per_second": 4.582,
|
1057 |
+
"eval_wer": 0.3406140167867632,
|
1058 |
+
"step": 10100
|
1059 |
+
},
|
1060 |
+
{
|
1061 |
+
"epoch": 1.9744483159117305,
|
1062 |
+
"eval_loss": 0.4579542577266693,
|
1063 |
+
"eval_runtime": 151.5074,
|
1064 |
+
"eval_samples_per_second": 37.331,
|
1065 |
+
"eval_steps_per_second": 4.666,
|
1066 |
+
"eval_wer": 0.3517837941936416,
|
1067 |
+
"step": 10200
|
1068 |
+
},
|
1069 |
+
{
|
1070 |
+
"epoch": 1.9938056523422376,
|
1071 |
+
"eval_loss": 0.46043792366981506,
|
1072 |
+
"eval_runtime": 151.4438,
|
1073 |
+
"eval_samples_per_second": 37.347,
|
1074 |
+
"eval_steps_per_second": 4.668,
|
1075 |
+
"eval_wer": 0.35115790149411824,
|
1076 |
+
"step": 10300
|
1077 |
+
},
|
1078 |
+
{
|
1079 |
+
"epoch": 2.013162988772745,
|
1080 |
+
"eval_loss": 0.46549099683761597,
|
1081 |
+
"eval_runtime": 151.5994,
|
1082 |
+
"eval_samples_per_second": 37.309,
|
1083 |
+
"eval_steps_per_second": 4.664,
|
1084 |
+
"eval_wer": 0.3552181797756415,
|
1085 |
+
"step": 10400
|
1086 |
+
},
|
1087 |
+
{
|
1088 |
+
"epoch": 2.032520325203252,
|
1089 |
+
"grad_norm": 0.703632652759552,
|
1090 |
+
"learning_rate": 0.0001463076923076923,
|
1091 |
+
"loss": 0.3624,
|
1092 |
+
"step": 10500
|
1093 |
+
},
|
1094 |
+
{
|
1095 |
+
"epoch": 2.032520325203252,
|
1096 |
+
"eval_loss": 0.4670031666755676,
|
1097 |
+
"eval_runtime": 151.5063,
|
1098 |
+
"eval_samples_per_second": 37.332,
|
1099 |
+
"eval_steps_per_second": 4.666,
|
1100 |
+
"eval_wer": 0.35144677504774435,
|
1101 |
+
"step": 10500
|
1102 |
+
},
|
1103 |
+
{
|
1104 |
+
"epoch": 2.051877661633759,
|
1105 |
+
"eval_loss": 0.459250271320343,
|
1106 |
+
"eval_runtime": 153.0971,
|
1107 |
+
"eval_samples_per_second": 36.944,
|
1108 |
+
"eval_steps_per_second": 4.618,
|
1109 |
+
"eval_wer": 0.3628251833544639,
|
1110 |
+
"step": 10600
|
1111 |
+
},
|
1112 |
+
{
|
1113 |
+
"epoch": 2.0712349980642664,
|
1114 |
+
"eval_loss": 0.46061432361602783,
|
1115 |
+
"eval_runtime": 152.0732,
|
1116 |
+
"eval_samples_per_second": 37.193,
|
1117 |
+
"eval_steps_per_second": 4.649,
|
1118 |
+
"eval_wer": 0.3545922870761182,
|
1119 |
+
"step": 10700
|
1120 |
+
},
|
1121 |
+
{
|
1122 |
+
"epoch": 2.0905923344947737,
|
1123 |
+
"eval_loss": 0.46500489115715027,
|
1124 |
+
"eval_runtime": 151.985,
|
1125 |
+
"eval_samples_per_second": 37.214,
|
1126 |
+
"eval_steps_per_second": 4.652,
|
1127 |
+
"eval_wer": 0.35905377862656673,
|
1128 |
+
"step": 10800
|
1129 |
+
},
|
1130 |
+
{
|
1131 |
+
"epoch": 2.1099496709252805,
|
1132 |
+
"eval_loss": 0.46085453033447266,
|
1133 |
+
"eval_runtime": 152.4835,
|
1134 |
+
"eval_samples_per_second": 37.093,
|
1135 |
+
"eval_steps_per_second": 4.637,
|
1136 |
+
"eval_wer": 0.35483301503747333,
|
1137 |
+
"step": 10900
|
1138 |
+
},
|
1139 |
+
{
|
1140 |
+
"epoch": 2.129307007355788,
|
1141 |
+
"grad_norm": 0.5008242726325989,
|
1142 |
+
"learning_rate": 0.00013863076923076922,
|
1143 |
+
"loss": 0.3755,
|
1144 |
+
"step": 11000
|
1145 |
+
},
|
1146 |
+
{
|
1147 |
+
"epoch": 2.129307007355788,
|
1148 |
+
"eval_loss": 0.4708138406276703,
|
1149 |
+
"eval_runtime": 152.3457,
|
1150 |
+
"eval_samples_per_second": 37.126,
|
1151 |
+
"eval_steps_per_second": 4.641,
|
1152 |
+
"eval_wer": 0.35573173275986586,
|
1153 |
+
"step": 11000
|
1154 |
+
},
|
1155 |
+
{
|
1156 |
+
"epoch": 2.148664343786295,
|
1157 |
+
"eval_loss": 0.4649392366409302,
|
1158 |
+
"eval_runtime": 152.7087,
|
1159 |
+
"eval_samples_per_second": 37.038,
|
1160 |
+
"eval_steps_per_second": 4.63,
|
1161 |
+
"eval_wer": 0.3548009179759593,
|
1162 |
+
"step": 11100
|
1163 |
+
},
|
1164 |
+
{
|
1165 |
+
"epoch": 2.168021680216802,
|
1166 |
+
"eval_loss": 0.4624271094799042,
|
1167 |
+
"eval_runtime": 153.153,
|
1168 |
+
"eval_samples_per_second": 36.93,
|
1169 |
+
"eval_steps_per_second": 4.616,
|
1170 |
+
"eval_wer": 0.355956412190464,
|
1171 |
+
"step": 11200
|
1172 |
+
},
|
1173 |
+
{
|
1174 |
+
"epoch": 2.1873790166473093,
|
1175 |
+
"eval_loss": 0.45822229981422424,
|
1176 |
+
"eval_runtime": 156.1964,
|
1177 |
+
"eval_samples_per_second": 36.211,
|
1178 |
+
"eval_steps_per_second": 4.526,
|
1179 |
+
"eval_wer": 0.35229734717786587,
|
1180 |
+
"step": 11300
|
1181 |
+
},
|
1182 |
+
{
|
1183 |
+
"epoch": 2.2067363530778166,
|
1184 |
+
"eval_loss": 0.466250479221344,
|
1185 |
+
"eval_runtime": 152.6707,
|
1186 |
+
"eval_samples_per_second": 37.047,
|
1187 |
+
"eval_steps_per_second": 4.631,
|
1188 |
+
"eval_wer": 0.3586044197653705,
|
1189 |
+
"step": 11400
|
1190 |
+
},
|
1191 |
+
{
|
1192 |
+
"epoch": 2.226093689508324,
|
1193 |
+
"grad_norm": 0.9631055593490601,
|
1194 |
+
"learning_rate": 0.00013093846153846151,
|
1195 |
+
"loss": 0.3891,
|
1196 |
+
"step": 11500
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 2.226093689508324,
|
1200 |
+
"eval_loss": 0.46153655648231506,
|
1201 |
+
"eval_runtime": 153.1909,
|
1202 |
+
"eval_samples_per_second": 36.921,
|
1203 |
+
"eval_steps_per_second": 4.615,
|
1204 |
+
"eval_wer": 0.3552181797756415,
|
1205 |
+
"step": 11500
|
1206 |
+
},
|
1207 |
+
{
|
1208 |
+
"epoch": 2.2454510259388307,
|
1209 |
+
"eval_loss": 0.4631531238555908,
|
1210 |
+
"eval_runtime": 152.9395,
|
1211 |
+
"eval_samples_per_second": 36.982,
|
1212 |
+
"eval_steps_per_second": 4.623,
|
1213 |
+
"eval_wer": 0.35886119625748264,
|
1214 |
+
"step": 11600
|
1215 |
+
},
|
1216 |
+
{
|
1217 |
+
"epoch": 2.264808362369338,
|
1218 |
+
"eval_loss": 0.4495234191417694,
|
1219 |
+
"eval_runtime": 153.0237,
|
1220 |
+
"eval_samples_per_second": 36.962,
|
1221 |
+
"eval_steps_per_second": 4.62,
|
1222 |
+
"eval_wer": 0.3425398404776043,
|
1223 |
+
"step": 11700
|
1224 |
+
},
|
1225 |
+
{
|
1226 |
+
"epoch": 2.2841656987998453,
|
1227 |
+
"eval_loss": 0.462666779756546,
|
1228 |
+
"eval_runtime": 152.4714,
|
1229 |
+
"eval_samples_per_second": 37.095,
|
1230 |
+
"eval_steps_per_second": 4.637,
|
1231 |
+
"eval_wer": 0.34942466017236123,
|
1232 |
+
"step": 11800
|
1233 |
+
},
|
1234 |
+
{
|
1235 |
+
"epoch": 2.303523035230352,
|
1236 |
+
"eval_loss": 0.4550352096557617,
|
1237 |
+
"eval_runtime": 152.8072,
|
1238 |
+
"eval_samples_per_second": 37.014,
|
1239 |
+
"eval_steps_per_second": 4.627,
|
1240 |
+
"eval_wer": 0.3451717995217538,
|
1241 |
+
"step": 11900
|
1242 |
+
},
|
1243 |
+
{
|
1244 |
+
"epoch": 2.3228803716608595,
|
1245 |
+
"grad_norm": 0.7961182594299316,
|
1246 |
+
"learning_rate": 0.00012324615384615384,
|
1247 |
+
"loss": 0.3946,
|
1248 |
+
"step": 12000
|
1249 |
+
},
|
1250 |
+
{
|
1251 |
+
"epoch": 2.3228803716608595,
|
1252 |
+
"eval_loss": 0.44988927245140076,
|
1253 |
+
"eval_runtime": 152.9644,
|
1254 |
+
"eval_samples_per_second": 36.976,
|
1255 |
+
"eval_steps_per_second": 4.622,
|
1256 |
+
"eval_wer": 0.3462310025517164,
|
1257 |
+
"step": 12000
|
1258 |
+
},
|
1259 |
+
{
|
1260 |
+
"epoch": 2.3422377080913668,
|
1261 |
+
"eval_loss": 0.4501667320728302,
|
1262 |
+
"eval_runtime": 153.061,
|
1263 |
+
"eval_samples_per_second": 36.953,
|
1264 |
+
"eval_steps_per_second": 4.619,
|
1265 |
+
"eval_wer": 0.341978141901109,
|
1266 |
+
"step": 12100
|
1267 |
+
},
|
1268 |
+
{
|
1269 |
+
"epoch": 2.3615950445218736,
|
1270 |
+
"eval_loss": 0.4580215513706207,
|
1271 |
+
"eval_runtime": 153.2108,
|
1272 |
+
"eval_samples_per_second": 36.916,
|
1273 |
+
"eval_steps_per_second": 4.615,
|
1274 |
+
"eval_wer": 0.3412399094862865,
|
1275 |
+
"step": 12200
|
1276 |
+
},
|
1277 |
+
{
|
1278 |
+
"epoch": 2.380952380952381,
|
1279 |
+
"eval_loss": 0.4506891667842865,
|
1280 |
+
"eval_runtime": 153.6611,
|
1281 |
+
"eval_samples_per_second": 36.808,
|
1282 |
+
"eval_steps_per_second": 4.601,
|
1283 |
+
"eval_wer": 0.34339041260772574,
|
1284 |
+
"step": 12300
|
1285 |
+
},
|
1286 |
+
{
|
1287 |
+
"epoch": 2.4003097173828882,
|
1288 |
+
"eval_loss": 0.44618555903434753,
|
1289 |
+
"eval_runtime": 153.273,
|
1290 |
+
"eval_samples_per_second": 36.901,
|
1291 |
+
"eval_steps_per_second": 4.613,
|
1292 |
+
"eval_wer": 0.34475453772207154,
|
1293 |
+
"step": 12400
|
1294 |
+
},
|
1295 |
+
{
|
1296 |
+
"epoch": 2.419667053813395,
|
1297 |
+
"grad_norm": 0.828158974647522,
|
1298 |
+
"learning_rate": 0.00011556923076923076,
|
1299 |
+
"loss": 0.3824,
|
1300 |
+
"step": 12500
|
1301 |
+
},
|
1302 |
+
{
|
1303 |
+
"epoch": 2.419667053813395,
|
1304 |
+
"eval_loss": 0.44126543402671814,
|
1305 |
+
"eval_runtime": 153.3979,
|
1306 |
+
"eval_samples_per_second": 36.871,
|
1307 |
+
"eval_steps_per_second": 4.609,
|
1308 |
+
"eval_wer": 0.34127200654780054,
|
1309 |
+
"step": 12500
|
1310 |
+
},
|
1311 |
+
{
|
1312 |
+
"epoch": 2.4390243902439024,
|
1313 |
+
"eval_loss": 0.44880929589271545,
|
1314 |
+
"eval_runtime": 153.7143,
|
1315 |
+
"eval_samples_per_second": 36.796,
|
1316 |
+
"eval_steps_per_second": 4.599,
|
1317 |
+
"eval_wer": 0.3443212273916323,
|
1318 |
+
"step": 12600
|
1319 |
+
},
|
1320 |
+
{
|
1321 |
+
"epoch": 2.4583817266744097,
|
1322 |
+
"eval_loss": 0.44148463010787964,
|
1323 |
+
"eval_runtime": 153.647,
|
1324 |
+
"eval_samples_per_second": 36.812,
|
1325 |
+
"eval_steps_per_second": 4.601,
|
1326 |
+
"eval_wer": 0.3431657331771276,
|
1327 |
+
"step": 12700
|
1328 |
+
},
|
1329 |
+
{
|
1330 |
+
"epoch": 2.4777390631049165,
|
1331 |
+
"eval_loss": 0.44202256202697754,
|
1332 |
+
"eval_runtime": 153.5743,
|
1333 |
+
"eval_samples_per_second": 36.829,
|
1334 |
+
"eval_steps_per_second": 4.604,
|
1335 |
+
"eval_wer": 0.34093498740190337,
|
1336 |
+
"step": 12800
|
1337 |
+
},
|
1338 |
+
{
|
1339 |
+
"epoch": 2.497096399535424,
|
1340 |
+
"eval_loss": 0.4379221200942993,
|
1341 |
+
"eval_runtime": 153.5736,
|
1342 |
+
"eval_samples_per_second": 36.829,
|
1343 |
+
"eval_steps_per_second": 4.604,
|
1344 |
+
"eval_wer": 0.3361204281748006,
|
1345 |
+
"step": 12900
|
1346 |
+
},
|
1347 |
+
{
|
1348 |
+
"epoch": 2.516453735965931,
|
1349 |
+
"grad_norm": 1.2163615226745605,
|
1350 |
+
"learning_rate": 0.00010787692307692307,
|
1351 |
+
"loss": 0.372,
|
1352 |
+
"step": 13000
|
1353 |
+
},
|
1354 |
+
{
|
1355 |
+
"epoch": 2.516453735965931,
|
1356 |
+
"eval_loss": 0.43855318427085876,
|
1357 |
+
"eval_runtime": 153.5476,
|
1358 |
+
"eval_samples_per_second": 36.835,
|
1359 |
+
"eval_steps_per_second": 4.604,
|
1360 |
+
"eval_wer": 0.3334884691306511,
|
1361 |
+
"step": 13000
|
1362 |
+
},
|
1363 |
+
{
|
1364 |
+
"epoch": 2.535811072396438,
|
1365 |
+
"eval_loss": 0.44449883699417114,
|
1366 |
+
"eval_runtime": 153.7016,
|
1367 |
+
"eval_samples_per_second": 36.799,
|
1368 |
+
"eval_steps_per_second": 4.6,
|
1369 |
+
"eval_wer": 0.3397794931873987,
|
1370 |
+
"step": 13100
|
1371 |
+
},
|
1372 |
+
{
|
1373 |
+
"epoch": 2.5551684088269453,
|
1374 |
+
"eval_loss": 0.4401286542415619,
|
1375 |
+
"eval_runtime": 154.2488,
|
1376 |
+
"eval_samples_per_second": 36.668,
|
1377 |
+
"eval_steps_per_second": 4.584,
|
1378 |
+
"eval_wer": 0.3392819887339314,
|
1379 |
+
"step": 13200
|
1380 |
+
},
|
1381 |
+
{
|
1382 |
+
"epoch": 2.5745257452574526,
|
1383 |
+
"eval_loss": 0.437770813703537,
|
1384 |
+
"eval_runtime": 153.8927,
|
1385 |
+
"eval_samples_per_second": 36.753,
|
1386 |
+
"eval_steps_per_second": 4.594,
|
1387 |
+
"eval_wer": 0.335077273675595,
|
1388 |
+
"step": 13300
|
1389 |
+
},
|
1390 |
+
{
|
1391 |
+
"epoch": 2.59388308168796,
|
1392 |
+
"eval_loss": 0.4315861463546753,
|
1393 |
+
"eval_runtime": 153.7886,
|
1394 |
+
"eval_samples_per_second": 36.778,
|
1395 |
+
"eval_steps_per_second": 4.597,
|
1396 |
+
"eval_wer": 0.33517356486013705,
|
1397 |
+
"step": 13400
|
1398 |
+
},
|
1399 |
+
{
|
1400 |
+
"epoch": 2.6132404181184667,
|
1401 |
+
"grad_norm": 1.084632158279419,
|
1402 |
+
"learning_rate": 0.0001002,
|
1403 |
+
"loss": 0.3521,
|
1404 |
+
"step": 13500
|
1405 |
+
},
|
1406 |
+
{
|
1407 |
+
"epoch": 2.6132404181184667,
|
1408 |
+
"eval_loss": 0.43864014744758606,
|
1409 |
+
"eval_runtime": 153.9711,
|
1410 |
+
"eval_samples_per_second": 36.734,
|
1411 |
+
"eval_steps_per_second": 4.592,
|
1412 |
+
"eval_wer": 0.33398597358411836,
|
1413 |
+
"step": 13500
|
1414 |
+
},
|
1415 |
+
{
|
1416 |
+
"epoch": 2.632597754548974,
|
1417 |
+
"eval_loss": 0.43551018834114075,
|
1418 |
+
"eval_runtime": 154.3017,
|
1419 |
+
"eval_samples_per_second": 36.655,
|
1420 |
+
"eval_steps_per_second": 4.582,
|
1421 |
+
"eval_wer": 0.33096884980180064,
|
1422 |
+
"step": 13600
|
1423 |
+
},
|
1424 |
+
{
|
1425 |
+
"epoch": 2.6519550909794813,
|
1426 |
+
"eval_loss": 0.4325660765171051,
|
1427 |
+
"eval_runtime": 154.4812,
|
1428 |
+
"eval_samples_per_second": 36.613,
|
1429 |
+
"eval_steps_per_second": 4.577,
|
1430 |
+
"eval_wer": 0.3343871868530436,
|
1431 |
+
"step": 13700
|
1432 |
+
},
|
1433 |
+
{
|
1434 |
+
"epoch": 2.6713124274099886,
|
1435 |
+
"eval_loss": 0.4263465404510498,
|
1436 |
+
"eval_runtime": 154.0733,
|
1437 |
+
"eval_samples_per_second": 36.71,
|
1438 |
+
"eval_steps_per_second": 4.589,
|
1439 |
+
"eval_wer": 0.32629872735151094,
|
1440 |
+
"step": 13800
|
1441 |
+
},
|
1442 |
+
{
|
1443 |
+
"epoch": 2.6906697638404955,
|
1444 |
+
"eval_loss": 0.42636117339134216,
|
1445 |
+
"eval_runtime": 154.1615,
|
1446 |
+
"eval_samples_per_second": 36.689,
|
1447 |
+
"eval_steps_per_second": 4.586,
|
1448 |
+
"eval_wer": 0.32353838006130536,
|
1449 |
+
"step": 13900
|
1450 |
+
},
|
1451 |
+
{
|
1452 |
+
"epoch": 2.710027100271003,
|
1453 |
+
"grad_norm": 1.1979655027389526,
|
1454 |
+
"learning_rate": 9.25076923076923e-05,
|
1455 |
+
"loss": 0.3592,
|
1456 |
+
"step": 14000
|
1457 |
+
},
|
1458 |
+
{
|
1459 |
+
"epoch": 2.710027100271003,
|
1460 |
+
"eval_loss": 0.4322036802768707,
|
1461 |
+
"eval_runtime": 154.5242,
|
1462 |
+
"eval_samples_per_second": 36.603,
|
1463 |
+
"eval_steps_per_second": 4.575,
|
1464 |
+
"eval_wer": 0.3299738408948661,
|
1465 |
+
"step": 14000
|
1466 |
+
},
|
1467 |
+
{
|
1468 |
+
"epoch": 2.72938443670151,
|
1469 |
+
"eval_loss": 0.4294193983078003,
|
1470 |
+
"eval_runtime": 154.4329,
|
1471 |
+
"eval_samples_per_second": 36.624,
|
1472 |
+
"eval_steps_per_second": 4.578,
|
1473 |
+
"eval_wer": 0.3261542905746979,
|
1474 |
+
"step": 14100
|
1475 |
+
},
|
1476 |
+
{
|
1477 |
+
"epoch": 2.748741773132017,
|
1478 |
+
"eval_loss": 0.43099814653396606,
|
1479 |
+
"eval_runtime": 154.4209,
|
1480 |
+
"eval_samples_per_second": 36.627,
|
1481 |
+
"eval_steps_per_second": 4.578,
|
1482 |
+
"eval_wer": 0.32329765209995026,
|
1483 |
+
"step": 14200
|
1484 |
+
},
|
1485 |
+
{
|
1486 |
+
"epoch": 2.7680991095625243,
|
1487 |
+
"eval_loss": 0.42700281739234924,
|
1488 |
+
"eval_runtime": 155.4008,
|
1489 |
+
"eval_samples_per_second": 36.396,
|
1490 |
+
"eval_steps_per_second": 4.55,
|
1491 |
+
"eval_wer": 0.3268122803357353,
|
1492 |
+
"step": 14300
|
1493 |
+
},
|
1494 |
+
{
|
1495 |
+
"epoch": 2.7874564459930316,
|
1496 |
+
"eval_loss": 0.4209098219871521,
|
1497 |
+
"eval_runtime": 156.5271,
|
1498 |
+
"eval_samples_per_second": 36.134,
|
1499 |
+
"eval_steps_per_second": 4.517,
|
1500 |
+
"eval_wer": 0.3254321066906325,
|
1501 |
+
"step": 14400
|
1502 |
+
},
|
1503 |
+
{
|
1504 |
+
"epoch": 2.8068137824235384,
|
1505 |
+
"grad_norm": 0.6974443793296814,
|
1506 |
+
"learning_rate": 8.48153846153846e-05,
|
1507 |
+
"loss": 0.3459,
|
1508 |
+
"step": 14500
|
1509 |
+
},
|
1510 |
+
{
|
1511 |
+
"epoch": 2.8068137824235384,
|
1512 |
+
"eval_loss": 0.42542555928230286,
|
1513 |
+
"eval_runtime": 157.9392,
|
1514 |
+
"eval_samples_per_second": 35.811,
|
1515 |
+
"eval_steps_per_second": 4.476,
|
1516 |
+
"eval_wer": 0.32729373625844554,
|
1517 |
+
"step": 14500
|
1518 |
+
},
|
1519 |
+
{
|
1520 |
+
"epoch": 2.8261711188540457,
|
1521 |
+
"eval_loss": 0.42783817648887634,
|
1522 |
+
"eval_runtime": 155.0217,
|
1523 |
+
"eval_samples_per_second": 36.485,
|
1524 |
+
"eval_steps_per_second": 4.561,
|
1525 |
+
"eval_wer": 0.3231532153231372,
|
1526 |
+
"step": 14600
|
1527 |
+
},
|
1528 |
+
{
|
1529 |
+
"epoch": 2.845528455284553,
|
1530 |
+
"eval_loss": 0.4212438464164734,
|
1531 |
+
"eval_runtime": 154.853,
|
1532 |
+
"eval_samples_per_second": 36.525,
|
1533 |
+
"eval_steps_per_second": 4.566,
|
1534 |
+
"eval_wer": 0.3215002166551652,
|
1535 |
+
"step": 14700
|
1536 |
+
},
|
1537 |
+
{
|
1538 |
+
"epoch": 2.86488579171506,
|
1539 |
+
"eval_loss": 0.4169256389141083,
|
1540 |
+
"eval_runtime": 154.5142,
|
1541 |
+
"eval_samples_per_second": 36.605,
|
1542 |
+
"eval_steps_per_second": 4.576,
|
1543 |
+
"eval_wer": 0.31928551941069794,
|
1544 |
+
"step": 14800
|
1545 |
+
},
|
1546 |
+
{
|
1547 |
+
"epoch": 2.884243128145567,
|
1548 |
+
"eval_loss": 0.42132049798965454,
|
1549 |
+
"eval_runtime": 154.8091,
|
1550 |
+
"eval_samples_per_second": 36.535,
|
1551 |
+
"eval_steps_per_second": 4.567,
|
1552 |
+
"eval_wer": 0.3195262473720531,
|
1553 |
+
"step": 14900
|
1554 |
+
},
|
1555 |
+
{
|
1556 |
+
"epoch": 2.9036004645760745,
|
1557 |
+
"grad_norm": 1.099702000617981,
|
1558 |
+
"learning_rate": 7.713846153846152e-05,
|
1559 |
+
"loss": 0.3483,
|
1560 |
+
"step": 15000
|
1561 |
+
},
|
1562 |
+
{
|
1563 |
+
"epoch": 2.9036004645760745,
|
1564 |
+
"eval_loss": 0.41696369647979736,
|
1565 |
+
"eval_runtime": 155.3223,
|
1566 |
+
"eval_samples_per_second": 36.415,
|
1567 |
+
"eval_steps_per_second": 4.552,
|
1568 |
+
"eval_wer": 0.31652517212049236,
|
1569 |
+
"step": 15000
|
1570 |
+
},
|
1571 |
+
{
|
1572 |
+
"epoch": 2.9229578010065813,
|
1573 |
+
"eval_loss": 0.41230952739715576,
|
1574 |
+
"eval_runtime": 154.9181,
|
1575 |
+
"eval_samples_per_second": 36.51,
|
1576 |
+
"eval_steps_per_second": 4.564,
|
1577 |
+
"eval_wer": 0.31418208662996905,
|
1578 |
+
"step": 15100
|
1579 |
+
},
|
1580 |
+
{
|
1581 |
+
"epoch": 2.9423151374370886,
|
1582 |
+
"eval_loss": 0.4116990566253662,
|
1583 |
+
"eval_runtime": 154.97,
|
1584 |
+
"eval_samples_per_second": 36.497,
|
1585 |
+
"eval_steps_per_second": 4.562,
|
1586 |
+
"eval_wer": 0.31337966009211854,
|
1587 |
+
"step": 15200
|
1588 |
+
},
|
1589 |
+
{
|
1590 |
+
"epoch": 2.961672473867596,
|
1591 |
+
"eval_loss": 0.410386323928833,
|
1592 |
+
"eval_runtime": 155.0232,
|
1593 |
+
"eval_samples_per_second": 36.485,
|
1594 |
+
"eval_steps_per_second": 4.561,
|
1595 |
+
"eval_wer": 0.31158222464733354,
|
1596 |
+
"step": 15300
|
1597 |
+
},
|
1598 |
+
{
|
1599 |
+
"epoch": 2.9810298102981028,
|
1600 |
+
"eval_loss": 0.41244322061538696,
|
1601 |
+
"eval_runtime": 154.4682,
|
1602 |
+
"eval_samples_per_second": 36.616,
|
1603 |
+
"eval_steps_per_second": 4.577,
|
1604 |
+
"eval_wer": 0.31419813516072603,
|
1605 |
+
"step": 15400
|
1606 |
+
},
|
1607 |
+
{
|
1608 |
+
"epoch": 3.00038714672861,
|
1609 |
+
"grad_norm": 0.725528359413147,
|
1610 |
+
"learning_rate": 6.946153846153845e-05,
|
1611 |
+
"loss": 0.3501,
|
1612 |
+
"step": 15500
|
1613 |
+
},
|
1614 |
+
{
|
1615 |
+
"epoch": 3.00038714672861,
|
1616 |
+
"eval_loss": 0.40684688091278076,
|
1617 |
+
"eval_runtime": 154.5863,
|
1618 |
+
"eval_samples_per_second": 36.588,
|
1619 |
+
"eval_steps_per_second": 4.573,
|
1620 |
+
"eval_wer": 0.31272167033108117,
|
1621 |
+
"step": 15500
|
1622 |
+
},
|
1623 |
+
{
|
1624 |
+
"epoch": 3.0197444831591174,
|
1625 |
+
"eval_loss": 0.4200752079486847,
|
1626 |
+
"eval_runtime": 154.5821,
|
1627 |
+
"eval_samples_per_second": 36.589,
|
1628 |
+
"eval_steps_per_second": 4.574,
|
1629 |
+
"eval_wer": 0.3087416347033429,
|
1630 |
+
"step": 15600
|
1631 |
+
},
|
1632 |
+
{
|
1633 |
+
"epoch": 3.0391018195896247,
|
1634 |
+
"eval_loss": 0.4186869263648987,
|
1635 |
+
"eval_runtime": 154.7417,
|
1636 |
+
"eval_samples_per_second": 36.551,
|
1637 |
+
"eval_steps_per_second": 4.569,
|
1638 |
+
"eval_wer": 0.3137808733610438,
|
1639 |
+
"step": 15700
|
1640 |
+
},
|
1641 |
+
{
|
1642 |
+
"epoch": 3.0584591560201315,
|
1643 |
+
"eval_loss": 0.41133585572242737,
|
1644 |
+
"eval_runtime": 155.21,
|
1645 |
+
"eval_samples_per_second": 36.441,
|
1646 |
+
"eval_steps_per_second": 4.555,
|
1647 |
+
"eval_wer": 0.31092423488629617,
|
1648 |
+
"step": 15800
|
1649 |
+
},
|
1650 |
+
{
|
1651 |
+
"epoch": 3.077816492450639,
|
1652 |
+
"eval_loss": 0.4191639721393585,
|
1653 |
+
"eval_runtime": 155.276,
|
1654 |
+
"eval_samples_per_second": 36.425,
|
1655 |
+
"eval_steps_per_second": 4.553,
|
1656 |
+
"eval_wer": 0.30851695527274475,
|
1657 |
+
"step": 15900
|
1658 |
+
},
|
1659 |
+
{
|
1660 |
+
"epoch": 3.097173828881146,
|
1661 |
+
"grad_norm": 0.5114701390266418,
|
1662 |
+
"learning_rate": 6.176923076923076e-05,
|
1663 |
+
"loss": 0.2754,
|
1664 |
+
"step": 16000
|
1665 |
+
},
|
1666 |
+
{
|
1667 |
+
"epoch": 3.097173828881146,
|
1668 |
+
"eval_loss": 0.4161028265953064,
|
1669 |
+
"eval_runtime": 154.597,
|
1670 |
+
"eval_samples_per_second": 36.585,
|
1671 |
+
"eval_steps_per_second": 4.573,
|
1672 |
+
"eval_wer": 0.30901445972621205,
|
1673 |
+
"step": 16000
|
1674 |
+
},
|
1675 |
+
{
|
1676 |
+
"epoch": 3.116531165311653,
|
1677 |
+
"eval_loss": 0.4183988571166992,
|
1678 |
+
"eval_runtime": 155.0124,
|
1679 |
+
"eval_samples_per_second": 36.487,
|
1680 |
+
"eval_steps_per_second": 4.561,
|
1681 |
+
"eval_wer": 0.307152830158399,
|
1682 |
+
"step": 16100
|
1683 |
+
},
|
1684 |
+
{
|
1685 |
+
"epoch": 3.1358885017421603,
|
1686 |
+
"eval_loss": 0.4186756908893585,
|
1687 |
+
"eval_runtime": 154.8535,
|
1688 |
+
"eval_samples_per_second": 36.525,
|
1689 |
+
"eval_steps_per_second": 4.566,
|
1690 |
+
"eval_wer": 0.3060936271284364,
|
1691 |
+
"step": 16200
|
1692 |
+
},
|
1693 |
+
{
|
1694 |
+
"epoch": 3.1552458381726676,
|
1695 |
+
"eval_loss": 0.4193824827671051,
|
1696 |
+
"eval_runtime": 154.3195,
|
1697 |
+
"eval_samples_per_second": 36.651,
|
1698 |
+
"eval_steps_per_second": 4.581,
|
1699 |
+
"eval_wer": 0.3059652388823803,
|
1700 |
+
"step": 16300
|
1701 |
+
},
|
1702 |
+
{
|
1703 |
+
"epoch": 3.1746031746031744,
|
1704 |
+
"eval_loss": 0.40788766741752625,
|
1705 |
+
"eval_runtime": 154.9673,
|
1706 |
+
"eval_samples_per_second": 36.498,
|
1707 |
+
"eval_steps_per_second": 4.562,
|
1708 |
+
"eval_wer": 0.3038949784147261,
|
1709 |
+
"step": 16400
|
1710 |
+
},
|
1711 |
+
{
|
1712 |
+
"epoch": 3.1939605110336817,
|
1713 |
+
"grad_norm": 0.5594165325164795,
|
1714 |
+
"learning_rate": 5.4076923076923074e-05,
|
1715 |
+
"loss": 0.2802,
|
1716 |
+
"step": 16500
|
1717 |
+
},
|
1718 |
+
{
|
1719 |
+
"epoch": 3.1939605110336817,
|
1720 |
+
"eval_loss": 0.41461309790611267,
|
1721 |
+
"eval_runtime": 154.8662,
|
1722 |
+
"eval_samples_per_second": 36.522,
|
1723 |
+
"eval_steps_per_second": 4.565,
|
1724 |
+
"eval_wer": 0.30424804609138034,
|
1725 |
+
"step": 16500
|
1726 |
+
},
|
1727 |
+
{
|
1728 |
+
"epoch": 3.213317847464189,
|
1729 |
+
"eval_loss": 0.4168522357940674,
|
1730 |
+
"eval_runtime": 155.0374,
|
1731 |
+
"eval_samples_per_second": 36.482,
|
1732 |
+
"eval_steps_per_second": 4.56,
|
1733 |
+
"eval_wer": 0.30116672818603457,
|
1734 |
+
"step": 16600
|
1735 |
+
},
|
1736 |
+
{
|
1737 |
+
"epoch": 3.2326751838946963,
|
1738 |
+
"eval_loss": 0.40926745533943176,
|
1739 |
+
"eval_runtime": 154.7423,
|
1740 |
+
"eval_samples_per_second": 36.551,
|
1741 |
+
"eval_steps_per_second": 4.569,
|
1742 |
+
"eval_wer": 0.3023864165235673,
|
1743 |
+
"step": 16700
|
1744 |
+
},
|
1745 |
+
{
|
1746 |
+
"epoch": 3.252032520325203,
|
1747 |
+
"eval_loss": 0.4115259051322937,
|
1748 |
+
"eval_runtime": 154.7933,
|
1749 |
+
"eval_samples_per_second": 36.539,
|
1750 |
+
"eval_steps_per_second": 4.567,
|
1751 |
+
"eval_wer": 0.3005408354865112,
|
1752 |
+
"step": 16800
|
1753 |
+
},
|
1754 |
+
{
|
1755 |
+
"epoch": 3.2713898567557105,
|
1756 |
+
"eval_loss": 0.40197211503982544,
|
1757 |
+
"eval_runtime": 155.5964,
|
1758 |
+
"eval_samples_per_second": 36.35,
|
1759 |
+
"eval_steps_per_second": 4.544,
|
1760 |
+
"eval_wer": 0.30410360931456726,
|
1761 |
+
"step": 16900
|
1762 |
+
},
|
1763 |
+
{
|
1764 |
+
"epoch": 3.290747193186218,
|
1765 |
+
"grad_norm": 1.4730154275894165,
|
1766 |
+
"learning_rate": 4.6384615384615385e-05,
|
1767 |
+
"loss": 0.2723,
|
1768 |
+
"step": 17000
|
1769 |
+
},
|
1770 |
+
{
|
1771 |
+
"epoch": 3.290747193186218,
|
1772 |
+
"eval_loss": 0.4058869779109955,
|
1773 |
+
"eval_runtime": 155.0898,
|
1774 |
+
"eval_samples_per_second": 36.469,
|
1775 |
+
"eval_steps_per_second": 4.559,
|
1776 |
+
"eval_wer": 0.30442457992970745,
|
1777 |
+
"step": 17000
|
1778 |
+
},
|
1779 |
+
{
|
1780 |
+
"epoch": 3.3101045296167246,
|
1781 |
+
"eval_loss": 0.40676185488700867,
|
1782 |
+
"eval_runtime": 155.0576,
|
1783 |
+
"eval_samples_per_second": 36.477,
|
1784 |
+
"eval_steps_per_second": 4.56,
|
1785 |
+
"eval_wer": 0.3013753590858757,
|
1786 |
+
"step": 17100
|
1787 |
+
},
|
1788 |
+
{
|
1789 |
+
"epoch": 3.329461866047232,
|
1790 |
+
"eval_loss": 0.40653425455093384,
|
1791 |
+
"eval_runtime": 155.8377,
|
1792 |
+
"eval_samples_per_second": 36.294,
|
1793 |
+
"eval_steps_per_second": 4.537,
|
1794 |
+
"eval_wer": 0.30878978029561394,
|
1795 |
+
"step": 17200
|
1796 |
+
},
|
1797 |
+
{
|
1798 |
+
"epoch": 3.3488192024777392,
|
1799 |
+
"eval_loss": 0.4082197844982147,
|
1800 |
+
"eval_runtime": 155.7924,
|
1801 |
+
"eval_samples_per_second": 36.305,
|
1802 |
+
"eval_steps_per_second": 4.538,
|
1803 |
+
"eval_wer": 0.3010543884707355,
|
1804 |
+
"step": 17300
|
1805 |
+
},
|
1806 |
+
{
|
1807 |
+
"epoch": 3.368176538908246,
|
1808 |
+
"eval_loss": 0.4083554446697235,
|
1809 |
+
"eval_runtime": 155.6775,
|
1810 |
+
"eval_samples_per_second": 36.332,
|
1811 |
+
"eval_steps_per_second": 4.541,
|
1812 |
+
"eval_wer": 0.3007494663863523,
|
1813 |
+
"step": 17400
|
1814 |
+
},
|
1815 |
+
{
|
1816 |
+
"epoch": 3.3875338753387534,
|
1817 |
+
"grad_norm": 0.5211097598075867,
|
1818 |
+
"learning_rate": 3.87076923076923e-05,
|
1819 |
+
"loss": 0.2557,
|
1820 |
+
"step": 17500
|
1821 |
+
},
|
1822 |
+
{
|
1823 |
+
"epoch": 3.3875338753387534,
|
1824 |
+
"eval_loss": 0.4009736180305481,
|
1825 |
+
"eval_runtime": 155.105,
|
1826 |
+
"eval_samples_per_second": 36.466,
|
1827 |
+
"eval_steps_per_second": 4.558,
|
1828 |
+
"eval_wer": 0.29924090449519347,
|
1829 |
+
"step": 17500
|
1830 |
+
},
|
1831 |
+
{
|
1832 |
+
"epoch": 3.4068912117692607,
|
1833 |
+
"eval_loss": 0.4061805009841919,
|
1834 |
+
"eval_runtime": 154.8792,
|
1835 |
+
"eval_samples_per_second": 36.519,
|
1836 |
+
"eval_steps_per_second": 4.565,
|
1837 |
+
"eval_wer": 0.2999951854407729,
|
1838 |
+
"step": 17600
|
1839 |
+
},
|
1840 |
+
{
|
1841 |
+
"epoch": 3.4262485481997675,
|
1842 |
+
"eval_loss": 0.40264037251472473,
|
1843 |
+
"eval_runtime": 155.9957,
|
1844 |
+
"eval_samples_per_second": 36.257,
|
1845 |
+
"eval_steps_per_second": 4.532,
|
1846 |
+
"eval_wer": 0.2980533132191748,
|
1847 |
+
"step": 17700
|
1848 |
+
},
|
1849 |
+
{
|
1850 |
+
"epoch": 3.445605884630275,
|
1851 |
+
"eval_loss": 0.40035372972488403,
|
1852 |
+
"eval_runtime": 155.0928,
|
1853 |
+
"eval_samples_per_second": 36.468,
|
1854 |
+
"eval_steps_per_second": 4.559,
|
1855 |
+
"eval_wer": 0.29893598241081026,
|
1856 |
+
"step": 17800
|
1857 |
+
},
|
1858 |
+
{
|
1859 |
+
"epoch": 3.464963221060782,
|
1860 |
+
"eval_loss": 0.40443336963653564,
|
1861 |
+
"eval_runtime": 154.9305,
|
1862 |
+
"eval_samples_per_second": 36.507,
|
1863 |
+
"eval_steps_per_second": 4.563,
|
1864 |
+
"eval_wer": 0.29906437065686636,
|
1865 |
+
"step": 17900
|
1866 |
+
},
|
1867 |
+
{
|
1868 |
+
"epoch": 3.484320557491289,
|
1869 |
+
"grad_norm": 0.7458967566490173,
|
1870 |
+
"learning_rate": 3.101538461538461e-05,
|
1871 |
+
"loss": 0.2578,
|
1872 |
+
"step": 18000
|
1873 |
+
},
|
1874 |
+
{
|
1875 |
+
"epoch": 3.484320557491289,
|
1876 |
+
"eval_loss": 0.4003549814224243,
|
1877 |
+
"eval_runtime": 155.7394,
|
1878 |
+
"eval_samples_per_second": 36.317,
|
1879 |
+
"eval_steps_per_second": 4.54,
|
1880 |
+
"eval_wer": 0.29660894545104394,
|
1881 |
+
"step": 18000
|
1882 |
+
},
|
1883 |
+
{
|
1884 |
+
"epoch": 3.5036778939217963,
|
1885 |
+
"eval_loss": 0.40592488646507263,
|
1886 |
+
"eval_runtime": 159.2644,
|
1887 |
+
"eval_samples_per_second": 35.513,
|
1888 |
+
"eval_steps_per_second": 4.439,
|
1889 |
+
"eval_wer": 0.29449053939111874,
|
1890 |
+
"step": 18100
|
1891 |
+
},
|
1892 |
+
{
|
1893 |
+
"epoch": 3.5230352303523036,
|
1894 |
+
"eval_loss": 0.4014962613582611,
|
1895 |
+
"eval_runtime": 155.6654,
|
1896 |
+
"eval_samples_per_second": 36.334,
|
1897 |
+
"eval_steps_per_second": 4.542,
|
1898 |
+
"eval_wer": 0.29632007189741777,
|
1899 |
+
"step": 18200
|
1900 |
+
},
|
1901 |
+
{
|
1902 |
+
"epoch": 3.5423925667828104,
|
1903 |
+
"eval_loss": 0.396659791469574,
|
1904 |
+
"eval_runtime": 156.1536,
|
1905 |
+
"eval_samples_per_second": 36.221,
|
1906 |
+
"eval_steps_per_second": 4.528,
|
1907 |
+
"eval_wer": 0.29585466450546455,
|
1908 |
+
"step": 18300
|
1909 |
+
},
|
1910 |
+
{
|
1911 |
+
"epoch": 3.5617499032133177,
|
1912 |
+
"eval_loss": 0.4001907706260681,
|
1913 |
+
"eval_runtime": 155.7578,
|
1914 |
+
"eval_samples_per_second": 36.313,
|
1915 |
+
"eval_steps_per_second": 4.539,
|
1916 |
+
"eval_wer": 0.29412142318370754,
|
1917 |
+
"step": 18400
|
1918 |
+
},
|
1919 |
+
{
|
1920 |
+
"epoch": 3.581107239643825,
|
1921 |
+
"grad_norm": 0.6122294664382935,
|
1922 |
+
"learning_rate": 2.3338461538461535e-05,
|
1923 |
+
"loss": 0.2508,
|
1924 |
+
"step": 18500
|
1925 |
+
},
|
1926 |
+
{
|
1927 |
+
"epoch": 3.581107239643825,
|
1928 |
+
"eval_loss": 0.39826107025146484,
|
1929 |
+
"eval_runtime": 155.467,
|
1930 |
+
"eval_samples_per_second": 36.381,
|
1931 |
+
"eval_steps_per_second": 4.548,
|
1932 |
+
"eval_wer": 0.2945547335141468,
|
1933 |
+
"step": 18500
|
1934 |
+
},
|
1935 |
+
{
|
1936 |
+
"epoch": 3.600464576074332,
|
1937 |
+
"eval_loss": 0.3958674967288971,
|
1938 |
+
"eval_runtime": 155.8242,
|
1939 |
+
"eval_samples_per_second": 36.297,
|
1940 |
+
"eval_steps_per_second": 4.537,
|
1941 |
+
"eval_wer": 0.29365601579175427,
|
1942 |
+
"step": 18600
|
1943 |
+
},
|
1944 |
+
{
|
1945 |
+
"epoch": 3.619821912504839,
|
1946 |
+
"eval_loss": 0.3970955014228821,
|
1947 |
+
"eval_runtime": 155.4329,
|
1948 |
+
"eval_samples_per_second": 36.389,
|
1949 |
+
"eval_steps_per_second": 4.549,
|
1950 |
+
"eval_wer": 0.2942016658374926,
|
1951 |
+
"step": 18700
|
1952 |
+
},
|
1953 |
+
{
|
1954 |
+
"epoch": 3.6391792489353465,
|
1955 |
+
"eval_loss": 0.3906669616699219,
|
1956 |
+
"eval_runtime": 155.4929,
|
1957 |
+
"eval_samples_per_second": 36.375,
|
1958 |
+
"eval_steps_per_second": 4.547,
|
1959 |
+
"eval_wer": 0.2923239877389225,
|
1960 |
+
"step": 18800
|
1961 |
+
},
|
1962 |
+
{
|
1963 |
+
"epoch": 3.658536585365854,
|
1964 |
+
"eval_loss": 0.39506247639656067,
|
1965 |
+
"eval_runtime": 155.5246,
|
1966 |
+
"eval_samples_per_second": 36.367,
|
1967 |
+
"eval_steps_per_second": 4.546,
|
1968 |
+
"eval_wer": 0.2903981640480814,
|
1969 |
+
"step": 18900
|
1970 |
+
},
|
1971 |
+
{
|
1972 |
+
"epoch": 3.6778939217963607,
|
1973 |
+
"grad_norm": 0.33715635538101196,
|
1974 |
+
"learning_rate": 1.5646153846153846e-05,
|
1975 |
+
"loss": 0.2659,
|
1976 |
+
"step": 19000
|
1977 |
+
},
|
1978 |
+
{
|
1979 |
+
"epoch": 3.6778939217963607,
|
1980 |
+
"eval_loss": 0.3892674744129181,
|
1981 |
+
"eval_runtime": 155.5533,
|
1982 |
+
"eval_samples_per_second": 36.361,
|
1983 |
+
"eval_steps_per_second": 4.545,
|
1984 |
+
"eval_wer": 0.29309431721525897,
|
1985 |
+
"step": 19000
|
1986 |
+
},
|
1987 |
+
{
|
1988 |
+
"epoch": 3.697251258226868,
|
1989 |
+
"eval_loss": 0.39077267050743103,
|
1990 |
+
"eval_runtime": 155.448,
|
1991 |
+
"eval_samples_per_second": 36.385,
|
1992 |
+
"eval_steps_per_second": 4.548,
|
1993 |
+
"eval_wer": 0.2900771934329412,
|
1994 |
+
"step": 19100
|
1995 |
+
},
|
1996 |
+
{
|
1997 |
+
"epoch": 3.7166085946573753,
|
1998 |
+
"eval_loss": 0.39407432079315186,
|
1999 |
+
"eval_runtime": 155.4696,
|
2000 |
+
"eval_samples_per_second": 36.38,
|
2001 |
+
"eval_steps_per_second": 4.548,
|
2002 |
+
"eval_wer": 0.2884241947649693,
|
2003 |
+
"step": 19200
|
2004 |
+
},
|
2005 |
+
{
|
2006 |
+
"epoch": 3.7359659310878826,
|
2007 |
+
"eval_loss": 0.3924821615219116,
|
2008 |
+
"eval_runtime": 155.4791,
|
2009 |
+
"eval_samples_per_second": 36.378,
|
2010 |
+
"eval_steps_per_second": 4.547,
|
2011 |
+
"eval_wer": 0.2890019418722216,
|
2012 |
+
"step": 19300
|
2013 |
+
},
|
2014 |
+
{
|
2015 |
+
"epoch": 3.7553232675183894,
|
2016 |
+
"eval_loss": 0.3916691243648529,
|
2017 |
+
"eval_runtime": 155.9516,
|
2018 |
+
"eval_samples_per_second": 36.268,
|
2019 |
+
"eval_steps_per_second": 4.533,
|
2020 |
+
"eval_wer": 0.2892908154258478,
|
2021 |
+
"step": 19400
|
2022 |
+
},
|
2023 |
+
{
|
2024 |
+
"epoch": 3.7746806039488967,
|
2025 |
+
"grad_norm": 0.4647356867790222,
|
2026 |
+
"learning_rate": 7.953846153846153e-06,
|
2027 |
+
"loss": 0.2488,
|
2028 |
+
"step": 19500
|
2029 |
+
},
|
2030 |
+
{
|
2031 |
+
"epoch": 3.7746806039488967,
|
2032 |
+
"eval_loss": 0.39043277502059937,
|
2033 |
+
"eval_runtime": 155.3552,
|
2034 |
+
"eval_samples_per_second": 36.407,
|
2035 |
+
"eval_steps_per_second": 4.551,
|
2036 |
+
"eval_wer": 0.2884562918264833,
|
2037 |
+
"step": 19500
|
2038 |
+
},
|
2039 |
+
{
|
2040 |
+
"epoch": 3.794037940379404,
|
2041 |
+
"eval_loss": 0.39014604687690735,
|
2042 |
+
"eval_runtime": 155.2137,
|
2043 |
+
"eval_samples_per_second": 36.44,
|
2044 |
+
"eval_steps_per_second": 4.555,
|
2045 |
+
"eval_wer": 0.2887933109723805,
|
2046 |
+
"step": 19600
|
2047 |
+
},
|
2048 |
+
{
|
2049 |
+
"epoch": 3.813395276809911,
|
2050 |
+
"eval_loss": 0.3883425295352936,
|
2051 |
+
"eval_runtime": 155.5369,
|
2052 |
+
"eval_samples_per_second": 36.364,
|
2053 |
+
"eval_steps_per_second": 4.546,
|
2054 |
+
"eval_wer": 0.28922662130281973,
|
2055 |
+
"step": 19700
|
2056 |
+
},
|
2057 |
+
{
|
2058 |
+
"epoch": 3.832752613240418,
|
2059 |
+
"eval_loss": 0.38913780450820923,
|
2060 |
+
"eval_runtime": 155.8958,
|
2061 |
+
"eval_samples_per_second": 36.281,
|
2062 |
+
"eval_steps_per_second": 4.535,
|
2063 |
+
"eval_wer": 0.28903403893373564,
|
2064 |
+
"step": 19800
|
2065 |
+
},
|
2066 |
+
{
|
2067 |
+
"epoch": 3.8521099496709255,
|
2068 |
+
"eval_loss": 0.3888201415538788,
|
2069 |
+
"eval_runtime": 155.6372,
|
2070 |
+
"eval_samples_per_second": 36.341,
|
2071 |
+
"eval_steps_per_second": 4.543,
|
2072 |
+
"eval_wer": 0.2888254080338945,
|
2073 |
+
"step": 19900
|
2074 |
+
},
|
2075 |
+
{
|
2076 |
+
"epoch": 3.8714672861014323,
|
2077 |
+
"grad_norm": 0.3741956055164337,
|
2078 |
+
"learning_rate": 2.615384615384615e-07,
|
2079 |
+
"loss": 0.2602,
|
2080 |
+
"step": 20000
|
2081 |
+
},
|
2082 |
+
{
|
2083 |
+
"epoch": 3.8714672861014323,
|
2084 |
+
"eval_loss": 0.38884833455085754,
|
2085 |
+
"eval_runtime": 155.0772,
|
2086 |
+
"eval_samples_per_second": 36.472,
|
2087 |
+
"eval_steps_per_second": 4.559,
|
2088 |
+
"eval_wer": 0.2884883888879973,
|
2089 |
+
"step": 20000
|
2090 |
+
},
|
2091 |
+
{
|
2092 |
+
"epoch": 3.8714672861014323,
|
2093 |
+
"step": 20000,
|
2094 |
+
"total_flos": 2.249387574100498e+19,
|
2095 |
+
"train_loss": 0.15996522521972656,
|
2096 |
+
"train_runtime": 19346.8732,
|
2097 |
+
"train_samples_per_second": 8.27,
|
2098 |
+
"train_steps_per_second": 1.034
|
2099 |
+
}
|
2100 |
+
],
|
2101 |
+
"logging_steps": 500,
|
2102 |
+
"max_steps": 20000,
|
2103 |
+
"num_input_tokens_seen": 0,
|
2104 |
+
"num_train_epochs": 4,
|
2105 |
+
"save_steps": 400,
|
2106 |
+
"total_flos": 2.249387574100498e+19,
|
2107 |
+
"train_batch_size": 8,
|
2108 |
+
"trial_name": null,
|
2109 |
+
"trial_params": null
|
2110 |
+
}
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bd8f3d0ded2642d5d200c2ed12ad9bf454ae146a69f8cb7e882c46c9847874d8
|
3 |
+
size 5048
|
vocab.json
CHANGED
@@ -1 +1,46 @@
|
|
1 |
-
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"'": 1,
|
3 |
+
"-": 2,
|
4 |
+
"[PAD]": 43,
|
5 |
+
"[UNK]": 42,
|
6 |
+
"a": 3,
|
7 |
+
"b": 4,
|
8 |
+
"c": 5,
|
9 |
+
"d": 6,
|
10 |
+
"e": 7,
|
11 |
+
"f": 8,
|
12 |
+
"g": 9,
|
13 |
+
"h": 10,
|
14 |
+
"i": 11,
|
15 |
+
"j": 12,
|
16 |
+
"k": 13,
|
17 |
+
"l": 14,
|
18 |
+
"m": 15,
|
19 |
+
"n": 16,
|
20 |
+
"o": 17,
|
21 |
+
"p": 18,
|
22 |
+
"q": 19,
|
23 |
+
"r": 20,
|
24 |
+
"s": 21,
|
25 |
+
"t": 22,
|
26 |
+
"u": 23,
|
27 |
+
"v": 24,
|
28 |
+
"w": 25,
|
29 |
+
"x": 26,
|
30 |
+
"y": 27,
|
31 |
+
"z": 28,
|
32 |
+
"|": 0,
|
33 |
+
"à": 29,
|
34 |
+
"â": 30,
|
35 |
+
"ä": 31,
|
36 |
+
"ê": 32,
|
37 |
+
"ë": 33,
|
38 |
+
"î": 34,
|
39 |
+
"ï": 35,
|
40 |
+
"ò": 36,
|
41 |
+
"ô": 37,
|
42 |
+
"û": 38,
|
43 |
+
"ÿ": 39,
|
44 |
+
"ŵ": 40,
|
45 |
+
"ŷ": 41
|
46 |
+
}
|