model upd
Browse files- all_results.json +12 -12
- config.json +1 -1
- eval_results.json +7 -7
- log_mozilla-foundation_common_voice_8_0_uk_test_predictions.txt +0 -0
- mozilla-foundation_common_voice_8_0_uk_test_eval_results.txt +2 -2
- pytorch_model.bin +1 -1
- run.sh +7 -6
- run_speech_recognition_ctc.py +3 -3
- runs/Feb05_17-03-38_job-680ae191-b2c7-4b97-adaf-cb186b6c96a6/1644080998.635567/events.out.tfevents.1644080998.job-680ae191-b2c7-4b97-adaf-cb186b6c96a6.1687777.1 +3 -0
- runs/Feb05_17-03-38_job-680ae191-b2c7-4b97-adaf-cb186b6c96a6/events.out.tfevents.1644080998.job-680ae191-b2c7-4b97-adaf-cb186b6c96a6.1687777.0 +3 -0
- runs/Feb05_17-03-38_job-680ae191-b2c7-4b97-adaf-cb186b6c96a6/events.out.tfevents.1644214441.job-680ae191-b2c7-4b97-adaf-cb186b6c96a6.1687777.2 +3 -0
- special_tokens_map.json +1 -1
- speech-recognition-community-v2_dev_data_uk_validation_eval_results.txt +2 -0
- train_results.json +6 -6
- trainer_state.json +614 -454
- training_args.bin +1 -1
all_results.json
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_cer": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
"eval_samples": 5802,
|
| 7 |
-
"eval_samples_per_second": 29.
|
| 8 |
-
"eval_steps_per_second": 0.
|
| 9 |
-
"eval_wer": 0.
|
| 10 |
-
"train_loss":
|
| 11 |
-
"train_runtime":
|
| 12 |
-
"train_samples":
|
| 13 |
-
"train_samples_per_second":
|
| 14 |
-
"train_steps_per_second": 0.
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 38.59,
|
| 3 |
+
"eval_cer": 0.032260099136267845,
|
| 4 |
+
"eval_loss": 0.10924588888883591,
|
| 5 |
+
"eval_runtime": 200.0507,
|
| 6 |
"eval_samples": 5802,
|
| 7 |
+
"eval_samples_per_second": 29.003,
|
| 8 |
+
"eval_steps_per_second": 0.455,
|
| 9 |
+
"eval_wer": 0.17520883477275945,
|
| 10 |
+
"train_loss": 1.0786900800069172,
|
| 11 |
+
"train_runtime": 133237.4383,
|
| 12 |
+
"train_samples": 39803,
|
| 13 |
+
"train_samples_per_second": 11.528,
|
| 14 |
+
"train_steps_per_second": 0.09
|
| 15 |
}
|
config.json
CHANGED
|
@@ -64,7 +64,7 @@
|
|
| 64 |
"mask_feature_prob": 0.25,
|
| 65 |
"mask_time_length": 10,
|
| 66 |
"mask_time_min_masks": 2,
|
| 67 |
-
"mask_time_prob": 0.
|
| 68 |
"model_type": "wav2vec2",
|
| 69 |
"num_adapter_layers": 3,
|
| 70 |
"num_attention_heads": 16,
|
|
|
|
| 64 |
"mask_feature_prob": 0.25,
|
| 65 |
"mask_time_length": 10,
|
| 66 |
"mask_time_min_masks": 2,
|
| 67 |
+
"mask_time_prob": 0.7,
|
| 68 |
"model_type": "wav2vec2",
|
| 69 |
"num_adapter_layers": 3,
|
| 70 |
"num_attention_heads": 16,
|
eval_results.json
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_cer": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
"eval_samples": 5802,
|
| 7 |
-
"eval_samples_per_second": 29.
|
| 8 |
-
"eval_steps_per_second": 0.
|
| 9 |
-
"eval_wer": 0.
|
| 10 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 38.59,
|
| 3 |
+
"eval_cer": 0.032260099136267845,
|
| 4 |
+
"eval_loss": 0.10924588888883591,
|
| 5 |
+
"eval_runtime": 200.0507,
|
| 6 |
"eval_samples": 5802,
|
| 7 |
+
"eval_samples_per_second": 29.003,
|
| 8 |
+
"eval_steps_per_second": 0.455,
|
| 9 |
+
"eval_wer": 0.17520883477275945
|
| 10 |
}
|
log_mozilla-foundation_common_voice_8_0_uk_test_predictions.txt
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
mozilla-foundation_common_voice_8_0_uk_test_eval_results.txt
CHANGED
|
@@ -1,2 +1,2 @@
|
|
| 1 |
-
WER: 0.
|
| 2 |
-
CER: 0.
|
|
|
|
| 1 |
+
WER: 0.10406342913776016
|
| 2 |
+
CER: 0.020387492208601702
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3850512561
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0899f4d43292dcdf2428937a08f0be20d919e37589599d1f33fbd667896e6296
|
| 3 |
size 3850512561
|
run.sh
CHANGED
|
@@ -1,9 +1,10 @@
|
|
| 1 |
python run_speech_recognition_ctc.py \
|
| 2 |
-
--dataset_name /workspace/data/uk/
|
| 3 |
--train_split_name train \
|
| 4 |
--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
|
| 5 |
-
--output_dir
|
| 6 |
-
--
|
|
|
|
| 7 |
--per_device_train_batch_size="16" \
|
| 8 |
--per_device_eval_batch_size="64" \
|
| 9 |
--gradient_accumulation_steps="8" \
|
|
@@ -18,12 +19,12 @@ python run_speech_recognition_ctc.py \
|
|
| 18 |
--save_steps="500" \
|
| 19 |
--eval_steps="500" \
|
| 20 |
--logging_steps="100" \
|
| 21 |
-
--save_total_limit
|
| 22 |
--freeze_feature_encoder \
|
| 23 |
--layerdrop="0.1" \
|
| 24 |
--activation_dropout="0.1" \
|
| 25 |
--feat_proj_dropout="0.0" \
|
| 26 |
-
--mask_time_prob="0.
|
| 27 |
--mask_time_length="10" \
|
| 28 |
--mask_feature_prob="0.25" \
|
| 29 |
--mask_feature_length="64" \
|
|
@@ -34,6 +35,6 @@ python run_speech_recognition_ctc.py \
|
|
| 34 |
--do_train --do_eval \
|
| 35 |
--load_best_model_at_end \
|
| 36 |
--report_to all \
|
| 37 |
-
--run_name="xlsr-uk-
|
| 38 |
--wandb_project="xlsr-uk" \
|
| 39 |
--bnb --tristage_sched
|
|
|
|
| 1 |
python run_speech_recognition_ctc.py \
|
| 2 |
+
--dataset_name /workspace/data/uk/composed_dataset/ \
|
| 3 |
--train_split_name train \
|
| 4 |
--model_name_or_path="facebook/wav2vec2-xls-r-1b" \
|
| 5 |
+
--output_dir ./ \
|
| 6 |
+
--overwrite_output_dir \
|
| 7 |
+
--max_steps 12000 \
|
| 8 |
--per_device_train_batch_size="16" \
|
| 9 |
--per_device_eval_batch_size="64" \
|
| 10 |
--gradient_accumulation_steps="8" \
|
|
|
|
| 19 |
--save_steps="500" \
|
| 20 |
--eval_steps="500" \
|
| 21 |
--logging_steps="100" \
|
| 22 |
+
--save_total_limit 10 \
|
| 23 |
--freeze_feature_encoder \
|
| 24 |
--layerdrop="0.1" \
|
| 25 |
--activation_dropout="0.1" \
|
| 26 |
--feat_proj_dropout="0.0" \
|
| 27 |
+
--mask_time_prob="0.7" \
|
| 28 |
--mask_time_length="10" \
|
| 29 |
--mask_feature_prob="0.25" \
|
| 30 |
--mask_feature_length="64" \
|
|
|
|
| 35 |
--do_train --do_eval \
|
| 36 |
--load_best_model_at_end \
|
| 37 |
--report_to all \
|
| 38 |
+
--run_name="xlsr-uk-1b-1" \
|
| 39 |
--wandb_project="xlsr-uk" \
|
| 40 |
--bnb --tristage_sched
|
run_speech_recognition_ctc.py
CHANGED
|
@@ -438,7 +438,7 @@ def main():
|
|
| 438 |
raw_datasets = DatasetDict()
|
| 439 |
|
| 440 |
if training_args.do_train:
|
| 441 |
-
if
|
| 442 |
raw_datasets["train"] = load_from_disk(f"{data_args.dataset_name}/{data_args.train_split_name}")
|
| 443 |
else:
|
| 444 |
raw_datasets["train"] = load_dataset(
|
|
@@ -466,7 +466,7 @@ def main():
|
|
| 466 |
raw_datasets["train"] = raw_datasets["train"].select(range(data_args.max_train_samples))
|
| 467 |
|
| 468 |
if training_args.do_eval:
|
| 469 |
-
if
|
| 470 |
raw_datasets["eval"] = load_from_disk(f"{data_args.dataset_name}/{data_args.eval_split_name}")
|
| 471 |
else:
|
| 472 |
raw_datasets["eval"] = load_dataset(
|
|
@@ -744,7 +744,7 @@ def main():
|
|
| 744 |
eps=training_args.adam_epsilon,
|
| 745 |
)
|
| 746 |
if extra_args.tristage_sched:
|
| 747 |
-
scheduler = get_tri_stage_schedule(optimizer, training_args.max_steps)
|
| 748 |
else:
|
| 749 |
scheduler = None
|
| 750 |
optimizers = (optimizer, scheduler)
|
|
|
|
| 438 |
raw_datasets = DatasetDict()
|
| 439 |
|
| 440 |
if training_args.do_train:
|
| 441 |
+
if os.path.isdir(data_args.dataset_name):
|
| 442 |
raw_datasets["train"] = load_from_disk(f"{data_args.dataset_name}/{data_args.train_split_name}")
|
| 443 |
else:
|
| 444 |
raw_datasets["train"] = load_dataset(
|
|
|
|
| 466 |
raw_datasets["train"] = raw_datasets["train"].select(range(data_args.max_train_samples))
|
| 467 |
|
| 468 |
if training_args.do_eval:
|
| 469 |
+
if os.path.isdir(data_args.dataset_name):
|
| 470 |
raw_datasets["eval"] = load_from_disk(f"{data_args.dataset_name}/{data_args.eval_split_name}")
|
| 471 |
else:
|
| 472 |
raw_datasets["eval"] = load_dataset(
|
|
|
|
| 744 |
eps=training_args.adam_epsilon,
|
| 745 |
)
|
| 746 |
if extra_args.tristage_sched:
|
| 747 |
+
scheduler = get_tri_stage_schedule(optimizer, training_args.max_steps, ratios=[0.1,0.3,0.6])
|
| 748 |
else:
|
| 749 |
scheduler = None
|
| 750 |
optimizers = (optimizer, scheduler)
|
runs/Feb05_17-03-38_job-680ae191-b2c7-4b97-adaf-cb186b6c96a6/1644080998.635567/events.out.tfevents.1644080998.job-680ae191-b2c7-4b97-adaf-cb186b6c96a6.1687777.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:14269093cd8a901cf553b343358bb3a6e11d7e016a5e3089a4fbef9e16f15342
|
| 3 |
+
size 4769
|
runs/Feb05_17-03-38_job-680ae191-b2c7-4b97-adaf-cb186b6c96a6/events.out.tfevents.1644080998.job-680ae191-b2c7-4b97-adaf-cb186b6c96a6.1687777.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5aebc5b8ec89626c05af71d1ad757ef5ae42fef6c33928134292765606f21235
|
| 3 |
+
size 32651
|
runs/Feb05_17-03-38_job-680ae191-b2c7-4b97-adaf-cb186b6c96a6/events.out.tfevents.1644214441.job-680ae191-b2c7-4b97-adaf-cb186b6c96a6.1687777.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e0e5e8b6387cfca950c01f1519bcfc4e021fb5b5c49e87204947a35d11ba276b
|
| 3 |
+
size 405
|
special_tokens_map.json
CHANGED
|
@@ -1 +1 @@
|
|
| 1 |
-
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}
|
|
|
|
| 1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}
|
speech-recognition-community-v2_dev_data_uk_validation_eval_results.txt
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
WER: 0.39531368102796677
|
| 2 |
+
CER: 0.20685439144902693
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
-
"train_samples":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 38.59,
|
| 3 |
+
"train_loss": 1.0786900800069172,
|
| 4 |
+
"train_runtime": 133237.4383,
|
| 5 |
+
"train_samples": 39803,
|
| 6 |
+
"train_samples_per_second": 11.528,
|
| 7 |
+
"train_steps_per_second": 0.09
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,825 +1,985 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "./checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
-
"epoch": 0.
|
| 12 |
-
"learning_rate":
|
| 13 |
-
"loss":
|
| 14 |
"step": 100
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch":
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss": 3.
|
| 20 |
"step": 200
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch":
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss":
|
| 26 |
"step": 300
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch":
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss":
|
| 32 |
"step": 400
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch":
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss": 1.
|
| 38 |
"step": 500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch":
|
| 42 |
-
"eval_cer": 0.
|
| 43 |
-
"eval_loss": 0.
|
| 44 |
-
"eval_runtime":
|
| 45 |
-
"eval_samples_per_second": 29.
|
| 46 |
-
"eval_steps_per_second": 0.
|
| 47 |
-
"eval_wer": 0.
|
| 48 |
"step": 500
|
| 49 |
},
|
| 50 |
{
|
| 51 |
-
"epoch":
|
| 52 |
-
"learning_rate":
|
| 53 |
-
"loss": 1.
|
| 54 |
"step": 600
|
| 55 |
},
|
| 56 |
{
|
| 57 |
-
"epoch":
|
| 58 |
-
"learning_rate":
|
| 59 |
-
"loss": 1.
|
| 60 |
"step": 700
|
| 61 |
},
|
| 62 |
{
|
| 63 |
-
"epoch":
|
| 64 |
-
"learning_rate":
|
| 65 |
-
"loss": 1.
|
| 66 |
"step": 800
|
| 67 |
},
|
| 68 |
{
|
| 69 |
-
"epoch":
|
| 70 |
-
"learning_rate":
|
| 71 |
-
"loss": 1.
|
| 72 |
"step": 900
|
| 73 |
},
|
| 74 |
{
|
| 75 |
-
"epoch":
|
| 76 |
-
"learning_rate":
|
| 77 |
-
"loss":
|
| 78 |
"step": 1000
|
| 79 |
},
|
| 80 |
{
|
| 81 |
-
"epoch":
|
| 82 |
-
"eval_cer": 0.
|
| 83 |
-
"eval_loss": 0.
|
| 84 |
-
"eval_runtime":
|
| 85 |
-
"eval_samples_per_second": 29.
|
| 86 |
-
"eval_steps_per_second": 0.
|
| 87 |
-
"eval_wer": 0.
|
| 88 |
"step": 1000
|
| 89 |
},
|
| 90 |
{
|
| 91 |
-
"epoch":
|
| 92 |
-
"learning_rate":
|
| 93 |
-
"loss":
|
| 94 |
"step": 1100
|
| 95 |
},
|
| 96 |
{
|
| 97 |
-
"epoch":
|
| 98 |
-
"learning_rate":
|
| 99 |
-
"loss":
|
| 100 |
"step": 1200
|
| 101 |
},
|
| 102 |
{
|
| 103 |
-
"epoch":
|
| 104 |
-
"learning_rate":
|
| 105 |
-
"loss":
|
| 106 |
"step": 1300
|
| 107 |
},
|
| 108 |
{
|
| 109 |
-
"epoch":
|
| 110 |
-
"learning_rate":
|
| 111 |
-
"loss":
|
| 112 |
"step": 1400
|
| 113 |
},
|
| 114 |
{
|
| 115 |
-
"epoch":
|
| 116 |
-
"learning_rate":
|
| 117 |
-
"loss":
|
| 118 |
"step": 1500
|
| 119 |
},
|
| 120 |
{
|
| 121 |
-
"epoch":
|
| 122 |
-
"eval_cer": 0.
|
| 123 |
-
"eval_loss": 0.
|
| 124 |
-
"eval_runtime":
|
| 125 |
-
"eval_samples_per_second": 29.
|
| 126 |
-
"eval_steps_per_second": 0.
|
| 127 |
-
"eval_wer": 0.
|
| 128 |
"step": 1500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
-
"epoch":
|
| 132 |
-
"learning_rate":
|
| 133 |
-
"loss":
|
| 134 |
"step": 1600
|
| 135 |
},
|
| 136 |
{
|
| 137 |
-
"epoch":
|
| 138 |
-
"learning_rate":
|
| 139 |
-
"loss":
|
| 140 |
"step": 1700
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch":
|
| 144 |
-
"learning_rate":
|
| 145 |
-
"loss":
|
| 146 |
"step": 1800
|
| 147 |
},
|
| 148 |
{
|
| 149 |
-
"epoch":
|
| 150 |
-
"learning_rate":
|
| 151 |
-
"loss":
|
| 152 |
"step": 1900
|
| 153 |
},
|
| 154 |
{
|
| 155 |
-
"epoch":
|
| 156 |
-
"learning_rate":
|
| 157 |
-
"loss":
|
| 158 |
"step": 2000
|
| 159 |
},
|
| 160 |
{
|
| 161 |
-
"epoch":
|
| 162 |
-
"eval_cer": 0.
|
| 163 |
-
"eval_loss": 0.
|
| 164 |
-
"eval_runtime":
|
| 165 |
-
"eval_samples_per_second": 29.
|
| 166 |
-
"eval_steps_per_second": 0.
|
| 167 |
-
"eval_wer": 0.
|
| 168 |
"step": 2000
|
| 169 |
},
|
| 170 |
{
|
| 171 |
-
"epoch":
|
| 172 |
-
"learning_rate":
|
| 173 |
-
"loss":
|
| 174 |
"step": 2100
|
| 175 |
},
|
| 176 |
{
|
| 177 |
-
"epoch":
|
| 178 |
-
"learning_rate":
|
| 179 |
-
"loss":
|
| 180 |
"step": 2200
|
| 181 |
},
|
| 182 |
{
|
| 183 |
-
"epoch":
|
| 184 |
-
"learning_rate":
|
| 185 |
-
"loss":
|
| 186 |
"step": 2300
|
| 187 |
},
|
| 188 |
{
|
| 189 |
-
"epoch":
|
| 190 |
-
"learning_rate":
|
| 191 |
-
"loss":
|
| 192 |
"step": 2400
|
| 193 |
},
|
| 194 |
{
|
| 195 |
-
"epoch":
|
| 196 |
-
"learning_rate":
|
| 197 |
-
"loss":
|
| 198 |
"step": 2500
|
| 199 |
},
|
| 200 |
{
|
| 201 |
-
"epoch":
|
| 202 |
-
"eval_cer": 0.
|
| 203 |
-
"eval_loss": 0.
|
| 204 |
-
"eval_runtime": 196.
|
| 205 |
-
"eval_samples_per_second": 29.
|
| 206 |
-
"eval_steps_per_second": 0.
|
| 207 |
-
"eval_wer": 0.
|
| 208 |
"step": 2500
|
| 209 |
},
|
| 210 |
{
|
| 211 |
-
"epoch":
|
| 212 |
-
"learning_rate":
|
| 213 |
-
"loss":
|
| 214 |
"step": 2600
|
| 215 |
},
|
| 216 |
{
|
| 217 |
-
"epoch":
|
| 218 |
-
"learning_rate":
|
| 219 |
-
"loss":
|
| 220 |
"step": 2700
|
| 221 |
},
|
| 222 |
{
|
| 223 |
-
"epoch":
|
| 224 |
-
"learning_rate":
|
| 225 |
-
"loss":
|
| 226 |
"step": 2800
|
| 227 |
},
|
| 228 |
{
|
| 229 |
-
"epoch":
|
| 230 |
-
"learning_rate":
|
| 231 |
-
"loss":
|
| 232 |
"step": 2900
|
| 233 |
},
|
| 234 |
{
|
| 235 |
-
"epoch":
|
| 236 |
-
"learning_rate":
|
| 237 |
-
"loss":
|
| 238 |
"step": 3000
|
| 239 |
},
|
| 240 |
{
|
| 241 |
-
"epoch":
|
| 242 |
-
"eval_cer": 0.
|
| 243 |
-
"eval_loss": 0.
|
| 244 |
-
"eval_runtime":
|
| 245 |
-
"eval_samples_per_second":
|
| 246 |
-
"eval_steps_per_second": 0.
|
| 247 |
-
"eval_wer": 0.
|
| 248 |
"step": 3000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
-
"epoch":
|
| 252 |
-
"learning_rate":
|
| 253 |
-
"loss":
|
| 254 |
"step": 3100
|
| 255 |
},
|
| 256 |
{
|
| 257 |
-
"epoch":
|
| 258 |
-
"learning_rate":
|
| 259 |
-
"loss":
|
| 260 |
"step": 3200
|
| 261 |
},
|
| 262 |
{
|
| 263 |
-
"epoch":
|
| 264 |
-
"learning_rate":
|
| 265 |
-
"loss":
|
| 266 |
"step": 3300
|
| 267 |
},
|
| 268 |
{
|
| 269 |
-
"epoch":
|
| 270 |
-
"learning_rate":
|
| 271 |
-
"loss":
|
| 272 |
"step": 3400
|
| 273 |
},
|
| 274 |
{
|
| 275 |
-
"epoch":
|
| 276 |
-
"learning_rate":
|
| 277 |
-
"loss":
|
| 278 |
"step": 3500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
-
"epoch":
|
| 282 |
-
"eval_cer": 0.
|
| 283 |
-
"eval_loss": 0.
|
| 284 |
-
"eval_runtime": 196.
|
| 285 |
-
"eval_samples_per_second": 29.
|
| 286 |
-
"eval_steps_per_second": 0.
|
| 287 |
-
"eval_wer": 0.
|
| 288 |
"step": 3500
|
| 289 |
},
|
| 290 |
{
|
| 291 |
-
"epoch":
|
| 292 |
-
"learning_rate":
|
| 293 |
-
"loss":
|
| 294 |
"step": 3600
|
| 295 |
},
|
| 296 |
{
|
| 297 |
-
"epoch":
|
| 298 |
-
"learning_rate":
|
| 299 |
-
"loss":
|
| 300 |
"step": 3700
|
| 301 |
},
|
| 302 |
{
|
| 303 |
-
"epoch":
|
| 304 |
-
"learning_rate":
|
| 305 |
-
"loss":
|
| 306 |
"step": 3800
|
| 307 |
},
|
| 308 |
{
|
| 309 |
-
"epoch":
|
| 310 |
-
"learning_rate":
|
| 311 |
-
"loss":
|
| 312 |
"step": 3900
|
| 313 |
},
|
| 314 |
{
|
| 315 |
-
"epoch":
|
| 316 |
-
"learning_rate":
|
| 317 |
-
"loss":
|
| 318 |
"step": 4000
|
| 319 |
},
|
| 320 |
{
|
| 321 |
-
"epoch":
|
| 322 |
-
"eval_cer": 0.
|
| 323 |
-
"eval_loss": 0.
|
| 324 |
-
"eval_runtime":
|
| 325 |
-
"eval_samples_per_second": 29.
|
| 326 |
-
"eval_steps_per_second": 0.
|
| 327 |
-
"eval_wer": 0.
|
| 328 |
"step": 4000
|
| 329 |
},
|
| 330 |
{
|
| 331 |
-
"epoch":
|
| 332 |
-
"learning_rate":
|
| 333 |
-
"loss":
|
| 334 |
"step": 4100
|
| 335 |
},
|
| 336 |
{
|
| 337 |
-
"epoch":
|
| 338 |
-
"learning_rate":
|
| 339 |
-
"loss":
|
| 340 |
"step": 4200
|
| 341 |
},
|
| 342 |
{
|
| 343 |
-
"epoch":
|
| 344 |
-
"learning_rate":
|
| 345 |
-
"loss":
|
| 346 |
"step": 4300
|
| 347 |
},
|
| 348 |
{
|
| 349 |
-
"epoch":
|
| 350 |
-
"learning_rate":
|
| 351 |
-
"loss":
|
| 352 |
"step": 4400
|
| 353 |
},
|
| 354 |
{
|
| 355 |
-
"epoch":
|
| 356 |
-
"learning_rate":
|
| 357 |
-
"loss":
|
| 358 |
"step": 4500
|
| 359 |
},
|
| 360 |
{
|
| 361 |
-
"epoch":
|
| 362 |
-
"eval_cer": 0.
|
| 363 |
-
"eval_loss": 0.
|
| 364 |
-
"eval_runtime":
|
| 365 |
-
"eval_samples_per_second": 29.
|
| 366 |
-
"eval_steps_per_second": 0.
|
| 367 |
-
"eval_wer": 0.
|
| 368 |
"step": 4500
|
| 369 |
},
|
| 370 |
{
|
| 371 |
-
"epoch":
|
| 372 |
-
"learning_rate":
|
| 373 |
-
"loss":
|
| 374 |
"step": 4600
|
| 375 |
},
|
| 376 |
{
|
| 377 |
-
"epoch":
|
| 378 |
-
"learning_rate":
|
| 379 |
-
"loss":
|
| 380 |
"step": 4700
|
| 381 |
},
|
| 382 |
{
|
| 383 |
-
"epoch":
|
| 384 |
-
"learning_rate":
|
| 385 |
-
"loss":
|
| 386 |
"step": 4800
|
| 387 |
},
|
| 388 |
{
|
| 389 |
-
"epoch":
|
| 390 |
-
"learning_rate":
|
| 391 |
-
"loss":
|
| 392 |
"step": 4900
|
| 393 |
},
|
| 394 |
{
|
| 395 |
-
"epoch":
|
| 396 |
-
"learning_rate":
|
| 397 |
-
"loss":
|
| 398 |
"step": 5000
|
| 399 |
},
|
| 400 |
{
|
| 401 |
-
"epoch":
|
| 402 |
-
"eval_cer": 0.
|
| 403 |
-
"eval_loss": 0.
|
| 404 |
-
"eval_runtime":
|
| 405 |
-
"eval_samples_per_second":
|
| 406 |
-
"eval_steps_per_second": 0.
|
| 407 |
-
"eval_wer": 0.
|
| 408 |
"step": 5000
|
| 409 |
},
|
| 410 |
{
|
| 411 |
-
"epoch":
|
| 412 |
-
"learning_rate":
|
| 413 |
-
"loss":
|
| 414 |
"step": 5100
|
| 415 |
},
|
| 416 |
{
|
| 417 |
-
"epoch":
|
| 418 |
-
"learning_rate":
|
| 419 |
-
"loss": 0.
|
| 420 |
"step": 5200
|
| 421 |
},
|
| 422 |
{
|
| 423 |
-
"epoch":
|
| 424 |
-
"learning_rate":
|
| 425 |
-
"loss":
|
| 426 |
"step": 5300
|
| 427 |
},
|
| 428 |
{
|
| 429 |
-
"epoch":
|
| 430 |
-
"learning_rate":
|
| 431 |
-
"loss": 0.
|
| 432 |
"step": 5400
|
| 433 |
},
|
| 434 |
{
|
| 435 |
-
"epoch":
|
| 436 |
-
"learning_rate":
|
| 437 |
-
"loss": 0.
|
| 438 |
"step": 5500
|
| 439 |
},
|
| 440 |
{
|
| 441 |
-
"epoch":
|
| 442 |
-
"eval_cer": 0.
|
| 443 |
-
"eval_loss": 0.
|
| 444 |
-
"eval_runtime":
|
| 445 |
-
"eval_samples_per_second":
|
| 446 |
-
"eval_steps_per_second": 0.
|
| 447 |
-
"eval_wer": 0.
|
| 448 |
"step": 5500
|
| 449 |
},
|
| 450 |
{
|
| 451 |
-
"epoch":
|
| 452 |
-
"learning_rate":
|
| 453 |
-
"loss": 0.
|
| 454 |
"step": 5600
|
| 455 |
},
|
| 456 |
{
|
| 457 |
-
"epoch":
|
| 458 |
-
"learning_rate":
|
| 459 |
-
"loss": 0.
|
| 460 |
"step": 5700
|
| 461 |
},
|
| 462 |
{
|
| 463 |
-
"epoch":
|
| 464 |
-
"learning_rate":
|
| 465 |
-
"loss": 0.
|
| 466 |
"step": 5800
|
| 467 |
},
|
| 468 |
{
|
| 469 |
-
"epoch":
|
| 470 |
-
"learning_rate":
|
| 471 |
-
"loss": 0.
|
| 472 |
"step": 5900
|
| 473 |
},
|
| 474 |
{
|
| 475 |
-
"epoch":
|
| 476 |
-
"learning_rate":
|
| 477 |
-
"loss": 0.
|
| 478 |
"step": 6000
|
| 479 |
},
|
| 480 |
{
|
| 481 |
-
"epoch":
|
| 482 |
-
"eval_cer": 0.
|
| 483 |
-
"eval_loss": 0.
|
| 484 |
-
"eval_runtime":
|
| 485 |
-
"eval_samples_per_second":
|
| 486 |
-
"eval_steps_per_second": 0.
|
| 487 |
-
"eval_wer": 0.
|
| 488 |
"step": 6000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
-
"epoch":
|
| 492 |
-
"learning_rate":
|
| 493 |
-
"loss": 0.
|
| 494 |
"step": 6100
|
| 495 |
},
|
| 496 |
{
|
| 497 |
-
"epoch":
|
| 498 |
-
"learning_rate":
|
| 499 |
-
"loss": 0.
|
| 500 |
"step": 6200
|
| 501 |
},
|
| 502 |
{
|
| 503 |
-
"epoch":
|
| 504 |
-
"learning_rate":
|
| 505 |
-
"loss": 0.
|
| 506 |
"step": 6300
|
| 507 |
},
|
| 508 |
{
|
| 509 |
-
"epoch":
|
| 510 |
-
"learning_rate":
|
| 511 |
-
"loss": 0.
|
| 512 |
"step": 6400
|
| 513 |
},
|
| 514 |
{
|
| 515 |
-
"epoch":
|
| 516 |
-
"learning_rate":
|
| 517 |
-
"loss": 0.
|
| 518 |
"step": 6500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
-
"epoch":
|
| 522 |
-
"eval_cer": 0.
|
| 523 |
-
"eval_loss": 0.
|
| 524 |
-
"eval_runtime":
|
| 525 |
-
"eval_samples_per_second":
|
| 526 |
-
"eval_steps_per_second": 0.
|
| 527 |
-
"eval_wer": 0.
|
| 528 |
"step": 6500
|
| 529 |
},
|
| 530 |
{
|
| 531 |
-
"epoch":
|
| 532 |
-
"learning_rate":
|
| 533 |
-
"loss": 0.
|
| 534 |
"step": 6600
|
| 535 |
},
|
| 536 |
{
|
| 537 |
-
"epoch":
|
| 538 |
-
"learning_rate":
|
| 539 |
-
"loss": 0.
|
| 540 |
"step": 6700
|
| 541 |
},
|
| 542 |
{
|
| 543 |
-
"epoch":
|
| 544 |
-
"learning_rate":
|
| 545 |
-
"loss": 0.
|
| 546 |
"step": 6800
|
| 547 |
},
|
| 548 |
{
|
| 549 |
-
"epoch":
|
| 550 |
-
"learning_rate":
|
| 551 |
-
"loss": 0.
|
| 552 |
"step": 6900
|
| 553 |
},
|
| 554 |
{
|
| 555 |
-
"epoch":
|
| 556 |
-
"learning_rate":
|
| 557 |
-
"loss": 0.
|
| 558 |
"step": 7000
|
| 559 |
},
|
| 560 |
{
|
| 561 |
-
"epoch":
|
| 562 |
-
"eval_cer": 0.
|
| 563 |
-
"eval_loss": 0.
|
| 564 |
-
"eval_runtime":
|
| 565 |
-
"eval_samples_per_second": 29.
|
| 566 |
-
"eval_steps_per_second": 0.
|
| 567 |
-
"eval_wer": 0.
|
| 568 |
"step": 7000
|
| 569 |
},
|
| 570 |
{
|
| 571 |
-
"epoch":
|
| 572 |
-
"learning_rate":
|
| 573 |
-
"loss": 0.
|
| 574 |
"step": 7100
|
| 575 |
},
|
| 576 |
{
|
| 577 |
-
"epoch":
|
| 578 |
-
"learning_rate":
|
| 579 |
-
"loss": 0.
|
| 580 |
"step": 7200
|
| 581 |
},
|
| 582 |
{
|
| 583 |
-
"epoch":
|
| 584 |
-
"learning_rate":
|
| 585 |
-
"loss": 0.
|
| 586 |
"step": 7300
|
| 587 |
},
|
| 588 |
{
|
| 589 |
-
"epoch":
|
| 590 |
-
"learning_rate":
|
| 591 |
-
"loss": 0.
|
| 592 |
"step": 7400
|
| 593 |
},
|
| 594 |
{
|
| 595 |
-
"epoch":
|
| 596 |
-
"learning_rate":
|
| 597 |
-
"loss": 0.
|
| 598 |
"step": 7500
|
| 599 |
},
|
| 600 |
{
|
| 601 |
-
"epoch":
|
| 602 |
-
"eval_cer": 0.
|
| 603 |
-
"eval_loss": 0.
|
| 604 |
-
"eval_runtime":
|
| 605 |
-
"eval_samples_per_second":
|
| 606 |
-
"eval_steps_per_second": 0.
|
| 607 |
-
"eval_wer": 0.
|
| 608 |
"step": 7500
|
| 609 |
},
|
| 610 |
{
|
| 611 |
-
"epoch":
|
| 612 |
-
"learning_rate":
|
| 613 |
-
"loss": 0.
|
| 614 |
"step": 7600
|
| 615 |
},
|
| 616 |
{
|
| 617 |
-
"epoch":
|
| 618 |
-
"learning_rate": 3.
|
| 619 |
-
"loss": 0.
|
| 620 |
"step": 7700
|
| 621 |
},
|
| 622 |
{
|
| 623 |
-
"epoch":
|
| 624 |
-
"learning_rate": 3.
|
| 625 |
-
"loss": 0.
|
| 626 |
"step": 7800
|
| 627 |
},
|
| 628 |
{
|
| 629 |
-
"epoch":
|
| 630 |
-
"learning_rate":
|
| 631 |
-
"loss": 0.
|
| 632 |
"step": 7900
|
| 633 |
},
|
| 634 |
{
|
| 635 |
-
"epoch":
|
| 636 |
-
"learning_rate":
|
| 637 |
-
"loss": 0.
|
| 638 |
"step": 8000
|
| 639 |
},
|
| 640 |
{
|
| 641 |
-
"epoch":
|
| 642 |
-
"eval_cer": 0.
|
| 643 |
-
"eval_loss": 0.
|
| 644 |
-
"eval_runtime": 197.
|
| 645 |
-
"eval_samples_per_second": 29.
|
| 646 |
-
"eval_steps_per_second": 0.
|
| 647 |
-
"eval_wer": 0.
|
| 648 |
"step": 8000
|
| 649 |
},
|
| 650 |
{
|
| 651 |
-
"epoch":
|
| 652 |
-
"learning_rate":
|
| 653 |
-
"loss": 0.
|
| 654 |
"step": 8100
|
| 655 |
},
|
| 656 |
{
|
| 657 |
-
"epoch":
|
| 658 |
-
"learning_rate":
|
| 659 |
-
"loss": 0.
|
| 660 |
"step": 8200
|
| 661 |
},
|
| 662 |
{
|
| 663 |
-
"epoch":
|
| 664 |
-
"learning_rate": 2.
|
| 665 |
-
"loss": 0.
|
| 666 |
"step": 8300
|
| 667 |
},
|
| 668 |
{
|
| 669 |
-
"epoch":
|
| 670 |
-
"learning_rate": 2.
|
| 671 |
-
"loss": 0.
|
| 672 |
"step": 8400
|
| 673 |
},
|
| 674 |
{
|
| 675 |
-
"epoch":
|
| 676 |
-
"learning_rate": 2.
|
| 677 |
-
"loss": 0.
|
| 678 |
"step": 8500
|
| 679 |
},
|
| 680 |
{
|
| 681 |
-
"epoch":
|
| 682 |
-
"eval_cer": 0.
|
| 683 |
-
"eval_loss": 0.
|
| 684 |
-
"eval_runtime":
|
| 685 |
-
"eval_samples_per_second": 29.
|
| 686 |
-
"eval_steps_per_second": 0.
|
| 687 |
-
"eval_wer": 0.
|
| 688 |
"step": 8500
|
| 689 |
},
|
| 690 |
{
|
| 691 |
-
"epoch":
|
| 692 |
-
"learning_rate": 2.
|
| 693 |
-
"loss": 0.
|
| 694 |
"step": 8600
|
| 695 |
},
|
| 696 |
{
|
| 697 |
-
"epoch":
|
| 698 |
-
"learning_rate": 2.
|
| 699 |
-
"loss": 0.
|
| 700 |
"step": 8700
|
| 701 |
},
|
| 702 |
{
|
| 703 |
-
"epoch":
|
| 704 |
-
"learning_rate": 2.
|
| 705 |
-
"loss": 0.
|
| 706 |
"step": 8800
|
| 707 |
},
|
| 708 |
{
|
| 709 |
-
"epoch":
|
| 710 |
-
"learning_rate": 2.
|
| 711 |
-
"loss": 0.
|
| 712 |
"step": 8900
|
| 713 |
},
|
| 714 |
{
|
| 715 |
-
"epoch":
|
| 716 |
-
"learning_rate":
|
| 717 |
-
"loss": 0.
|
| 718 |
"step": 9000
|
| 719 |
},
|
| 720 |
{
|
| 721 |
-
"epoch":
|
| 722 |
-
"eval_cer": 0.
|
| 723 |
-
"eval_loss": 0.
|
| 724 |
-
"eval_runtime":
|
| 725 |
-
"eval_samples_per_second": 29.
|
| 726 |
-
"eval_steps_per_second": 0.
|
| 727 |
-
"eval_wer": 0.
|
| 728 |
"step": 9000
|
| 729 |
},
|
| 730 |
{
|
| 731 |
-
"epoch":
|
| 732 |
-
"learning_rate":
|
| 733 |
-
"loss": 0.
|
| 734 |
"step": 9100
|
| 735 |
},
|
| 736 |
{
|
| 737 |
-
"epoch":
|
| 738 |
-
"learning_rate":
|
| 739 |
-
"loss": 0.
|
| 740 |
"step": 9200
|
| 741 |
},
|
| 742 |
{
|
| 743 |
-
"epoch":
|
| 744 |
-
"learning_rate":
|
| 745 |
-
"loss": 0.
|
| 746 |
"step": 9300
|
| 747 |
},
|
| 748 |
{
|
| 749 |
-
"epoch":
|
| 750 |
-
"learning_rate": 1.
|
| 751 |
-
"loss": 0.
|
| 752 |
"step": 9400
|
| 753 |
},
|
| 754 |
{
|
| 755 |
-
"epoch":
|
| 756 |
-
"learning_rate": 1.
|
| 757 |
-
"loss": 0.
|
| 758 |
"step": 9500
|
| 759 |
},
|
| 760 |
{
|
| 761 |
-
"epoch":
|
| 762 |
-
"eval_cer": 0.
|
| 763 |
-
"eval_loss": 0.
|
| 764 |
-
"eval_runtime":
|
| 765 |
-
"eval_samples_per_second":
|
| 766 |
-
"eval_steps_per_second": 0.
|
| 767 |
-
"eval_wer": 0.
|
| 768 |
"step": 9500
|
| 769 |
},
|
| 770 |
{
|
| 771 |
-
"epoch":
|
| 772 |
-
"learning_rate": 1.
|
| 773 |
-
"loss": 0.
|
| 774 |
"step": 9600
|
| 775 |
},
|
| 776 |
{
|
| 777 |
-
"epoch":
|
| 778 |
-
"learning_rate":
|
| 779 |
-
"loss": 0.
|
| 780 |
"step": 9700
|
| 781 |
},
|
| 782 |
{
|
| 783 |
-
"epoch":
|
| 784 |
-
"learning_rate":
|
| 785 |
-
"loss": 0.
|
| 786 |
"step": 9800
|
| 787 |
},
|
| 788 |
{
|
| 789 |
-
"epoch":
|
| 790 |
-
"learning_rate":
|
| 791 |
-
"loss": 0.
|
| 792 |
"step": 9900
|
| 793 |
},
|
| 794 |
{
|
| 795 |
-
"epoch":
|
| 796 |
-
"learning_rate":
|
| 797 |
-
"loss": 0.
|
| 798 |
"step": 10000
|
| 799 |
},
|
| 800 |
{
|
| 801 |
-
"epoch":
|
| 802 |
-
"eval_cer": 0.
|
| 803 |
-
"eval_loss": 0.
|
| 804 |
-
"eval_runtime":
|
| 805 |
-
"eval_samples_per_second": 29.
|
| 806 |
-
"eval_steps_per_second": 0.
|
| 807 |
-
"eval_wer": 0.
|
| 808 |
"step": 10000
|
| 809 |
},
|
| 810 |
{
|
| 811 |
-
"epoch":
|
| 812 |
-
"
|
| 813 |
-
"
|
| 814 |
-
"
|
| 815 |
-
|
| 816 |
-
|
| 817 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 818 |
}
|
| 819 |
],
|
| 820 |
-
"max_steps":
|
| 821 |
-
"num_train_epochs":
|
| 822 |
-
"total_flos":
|
| 823 |
"trial_name": null,
|
| 824 |
"trial_params": null
|
| 825 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.10924588888883591,
|
| 3 |
+
"best_model_checkpoint": "./checkpoint-12000",
|
| 4 |
+
"epoch": 38.58520900321543,
|
| 5 |
+
"global_step": 12000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
+
"epoch": 0.32,
|
| 12 |
+
"learning_rate": 4.50125e-06,
|
| 13 |
+
"loss": 8.0884,
|
| 14 |
"step": 100
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.64,
|
| 18 |
+
"learning_rate": 8.62625e-06,
|
| 19 |
+
"loss": 3.2246,
|
| 20 |
"step": 200
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.96,
|
| 24 |
+
"learning_rate": 1.2751250000000001e-05,
|
| 25 |
+
"loss": 3.1607,
|
| 26 |
"step": 300
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 1.29,
|
| 30 |
+
"learning_rate": 1.6876250000000003e-05,
|
| 31 |
+
"loss": 2.3964,
|
| 32 |
"step": 400
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 1.61,
|
| 36 |
+
"learning_rate": 2.100125e-05,
|
| 37 |
+
"loss": 1.7005,
|
| 38 |
"step": 500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 1.61,
|
| 42 |
+
"eval_cer": 0.11636638864978778,
|
| 43 |
+
"eval_loss": 0.4082379639148712,
|
| 44 |
+
"eval_runtime": 199.3377,
|
| 45 |
+
"eval_samples_per_second": 29.106,
|
| 46 |
+
"eval_steps_per_second": 0.457,
|
| 47 |
+
"eval_wer": 0.5583793477747888,
|
| 48 |
"step": 500
|
| 49 |
},
|
| 50 |
{
|
| 51 |
+
"epoch": 1.93,
|
| 52 |
+
"learning_rate": 2.512625e-05,
|
| 53 |
+
"loss": 1.4874,
|
| 54 |
"step": 600
|
| 55 |
},
|
| 56 |
{
|
| 57 |
+
"epoch": 2.25,
|
| 58 |
+
"learning_rate": 2.9251250000000002e-05,
|
| 59 |
+
"loss": 1.3431,
|
| 60 |
"step": 700
|
| 61 |
},
|
| 62 |
{
|
| 63 |
+
"epoch": 2.57,
|
| 64 |
+
"learning_rate": 3.3376250000000004e-05,
|
| 65 |
+
"loss": 1.2316,
|
| 66 |
"step": 800
|
| 67 |
},
|
| 68 |
{
|
| 69 |
+
"epoch": 2.89,
|
| 70 |
+
"learning_rate": 3.750125e-05,
|
| 71 |
+
"loss": 1.187,
|
| 72 |
"step": 900
|
| 73 |
},
|
| 74 |
{
|
| 75 |
+
"epoch": 3.22,
|
| 76 |
+
"learning_rate": 4.162625e-05,
|
| 77 |
+
"loss": 1.1555,
|
| 78 |
"step": 1000
|
| 79 |
},
|
| 80 |
{
|
| 81 |
+
"epoch": 3.22,
|
| 82 |
+
"eval_cer": 0.05566026535276483,
|
| 83 |
+
"eval_loss": 0.2020130306482315,
|
| 84 |
+
"eval_runtime": 199.2116,
|
| 85 |
+
"eval_samples_per_second": 29.125,
|
| 86 |
+
"eval_steps_per_second": 0.457,
|
| 87 |
+
"eval_wer": 0.29534192269573833,
|
| 88 |
"step": 1000
|
| 89 |
},
|
| 90 |
{
|
| 91 |
+
"epoch": 3.54,
|
| 92 |
+
"learning_rate": 4.575125e-05,
|
| 93 |
+
"loss": 1.1286,
|
| 94 |
"step": 1100
|
| 95 |
},
|
| 96 |
{
|
| 97 |
+
"epoch": 3.86,
|
| 98 |
+
"learning_rate": 4.9876250000000005e-05,
|
| 99 |
+
"loss": 1.1143,
|
| 100 |
"step": 1200
|
| 101 |
},
|
| 102 |
{
|
| 103 |
+
"epoch": 4.18,
|
| 104 |
+
"learning_rate": 5e-05,
|
| 105 |
+
"loss": 1.1067,
|
| 106 |
"step": 1300
|
| 107 |
},
|
| 108 |
{
|
| 109 |
+
"epoch": 4.5,
|
| 110 |
+
"learning_rate": 5e-05,
|
| 111 |
+
"loss": 1.0992,
|
| 112 |
"step": 1400
|
| 113 |
},
|
| 114 |
{
|
| 115 |
+
"epoch": 4.82,
|
| 116 |
+
"learning_rate": 5e-05,
|
| 117 |
+
"loss": 1.0927,
|
| 118 |
"step": 1500
|
| 119 |
},
|
| 120 |
{
|
| 121 |
+
"epoch": 4.82,
|
| 122 |
+
"eval_cer": 0.04799130331542548,
|
| 123 |
+
"eval_loss": 0.1707664430141449,
|
| 124 |
+
"eval_runtime": 197.8453,
|
| 125 |
+
"eval_samples_per_second": 29.326,
|
| 126 |
+
"eval_steps_per_second": 0.46,
|
| 127 |
+
"eval_wer": 0.25843598093350323,
|
| 128 |
"step": 1500
|
| 129 |
},
|
| 130 |
{
|
| 131 |
+
"epoch": 5.14,
|
| 132 |
+
"learning_rate": 5e-05,
|
| 133 |
+
"loss": 1.0907,
|
| 134 |
"step": 1600
|
| 135 |
},
|
| 136 |
{
|
| 137 |
+
"epoch": 5.47,
|
| 138 |
+
"learning_rate": 5e-05,
|
| 139 |
+
"loss": 1.0765,
|
| 140 |
"step": 1700
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 5.79,
|
| 144 |
+
"learning_rate": 5e-05,
|
| 145 |
+
"loss": 1.0693,
|
| 146 |
"step": 1800
|
| 147 |
},
|
| 148 |
{
|
| 149 |
+
"epoch": 6.11,
|
| 150 |
+
"learning_rate": 5e-05,
|
| 151 |
+
"loss": 1.0547,
|
| 152 |
"step": 1900
|
| 153 |
},
|
| 154 |
{
|
| 155 |
+
"epoch": 6.43,
|
| 156 |
+
"learning_rate": 5e-05,
|
| 157 |
+
"loss": 1.0707,
|
| 158 |
"step": 2000
|
| 159 |
},
|
| 160 |
{
|
| 161 |
+
"epoch": 6.43,
|
| 162 |
+
"eval_cer": 0.04497120892820041,
|
| 163 |
+
"eval_loss": 0.15630319714546204,
|
| 164 |
+
"eval_runtime": 197.159,
|
| 165 |
+
"eval_samples_per_second": 29.428,
|
| 166 |
+
"eval_steps_per_second": 0.462,
|
| 167 |
+
"eval_wer": 0.24054934163952996,
|
| 168 |
"step": 2000
|
| 169 |
},
|
| 170 |
{
|
| 171 |
+
"epoch": 6.75,
|
| 172 |
+
"learning_rate": 5e-05,
|
| 173 |
+
"loss": 1.0647,
|
| 174 |
"step": 2100
|
| 175 |
},
|
| 176 |
{
|
| 177 |
+
"epoch": 7.07,
|
| 178 |
+
"learning_rate": 5e-05,
|
| 179 |
+
"loss": 1.054,
|
| 180 |
"step": 2200
|
| 181 |
},
|
| 182 |
{
|
| 183 |
+
"epoch": 7.4,
|
| 184 |
+
"learning_rate": 5e-05,
|
| 185 |
+
"loss": 1.0478,
|
| 186 |
"step": 2300
|
| 187 |
},
|
| 188 |
{
|
| 189 |
+
"epoch": 7.72,
|
| 190 |
+
"learning_rate": 5e-05,
|
| 191 |
+
"loss": 1.0611,
|
| 192 |
"step": 2400
|
| 193 |
},
|
| 194 |
{
|
| 195 |
+
"epoch": 8.04,
|
| 196 |
+
"learning_rate": 5e-05,
|
| 197 |
+
"loss": 1.0728,
|
| 198 |
"step": 2500
|
| 199 |
},
|
| 200 |
{
|
| 201 |
+
"epoch": 8.04,
|
| 202 |
+
"eval_cer": 0.04629574663856816,
|
| 203 |
+
"eval_loss": 0.16203930974006653,
|
| 204 |
+
"eval_runtime": 196.804,
|
| 205 |
+
"eval_samples_per_second": 29.481,
|
| 206 |
+
"eval_steps_per_second": 0.462,
|
| 207 |
+
"eval_wer": 0.2442304969559677,
|
| 208 |
"step": 2500
|
| 209 |
},
|
| 210 |
{
|
| 211 |
+
"epoch": 8.36,
|
| 212 |
+
"learning_rate": 5e-05,
|
| 213 |
+
"loss": 1.0563,
|
| 214 |
"step": 2600
|
| 215 |
},
|
| 216 |
{
|
| 217 |
+
"epoch": 8.68,
|
| 218 |
+
"learning_rate": 5e-05,
|
| 219 |
+
"loss": 1.0404,
|
| 220 |
"step": 2700
|
| 221 |
},
|
| 222 |
{
|
| 223 |
+
"epoch": 9.0,
|
| 224 |
+
"learning_rate": 5e-05,
|
| 225 |
+
"loss": 1.0791,
|
| 226 |
"step": 2800
|
| 227 |
},
|
| 228 |
{
|
| 229 |
+
"epoch": 9.32,
|
| 230 |
+
"learning_rate": 5e-05,
|
| 231 |
+
"loss": 1.0535,
|
| 232 |
"step": 2900
|
| 233 |
},
|
| 234 |
{
|
| 235 |
+
"epoch": 9.65,
|
| 236 |
+
"learning_rate": 5e-05,
|
| 237 |
+
"loss": 1.0268,
|
| 238 |
"step": 3000
|
| 239 |
},
|
| 240 |
{
|
| 241 |
+
"epoch": 9.65,
|
| 242 |
+
"eval_cer": 0.04575776913715829,
|
| 243 |
+
"eval_loss": 0.15875375270843506,
|
| 244 |
+
"eval_runtime": 201.4769,
|
| 245 |
+
"eval_samples_per_second": 28.797,
|
| 246 |
+
"eval_steps_per_second": 0.452,
|
| 247 |
+
"eval_wer": 0.2377884751522016,
|
| 248 |
"step": 3000
|
| 249 |
},
|
| 250 |
{
|
| 251 |
+
"epoch": 9.97,
|
| 252 |
+
"learning_rate": 5e-05,
|
| 253 |
+
"loss": 1.0322,
|
| 254 |
"step": 3100
|
| 255 |
},
|
| 256 |
{
|
| 257 |
+
"epoch": 10.29,
|
| 258 |
+
"learning_rate": 5e-05,
|
| 259 |
+
"loss": 1.0208,
|
| 260 |
"step": 3200
|
| 261 |
},
|
| 262 |
{
|
| 263 |
+
"epoch": 10.61,
|
| 264 |
+
"learning_rate": 5e-05,
|
| 265 |
+
"loss": 1.0172,
|
| 266 |
"step": 3300
|
| 267 |
},
|
| 268 |
{
|
| 269 |
+
"epoch": 10.93,
|
| 270 |
+
"learning_rate": 5e-05,
|
| 271 |
+
"loss": 1.019,
|
| 272 |
"step": 3400
|
| 273 |
},
|
| 274 |
{
|
| 275 |
+
"epoch": 11.25,
|
| 276 |
+
"learning_rate": 5e-05,
|
| 277 |
+
"loss": 1.0328,
|
| 278 |
"step": 3500
|
| 279 |
},
|
| 280 |
{
|
| 281 |
+
"epoch": 11.25,
|
| 282 |
+
"eval_cer": 0.04419206909857232,
|
| 283 |
+
"eval_loss": 0.14661966264247894,
|
| 284 |
+
"eval_runtime": 196.9894,
|
| 285 |
+
"eval_samples_per_second": 29.453,
|
| 286 |
+
"eval_steps_per_second": 0.462,
|
| 287 |
+
"eval_wer": 0.23516919156165936,
|
| 288 |
"step": 3500
|
| 289 |
},
|
| 290 |
{
|
| 291 |
+
"epoch": 11.58,
|
| 292 |
+
"learning_rate": 5e-05,
|
| 293 |
+
"loss": 1.0153,
|
| 294 |
"step": 3600
|
| 295 |
},
|
| 296 |
{
|
| 297 |
+
"epoch": 11.9,
|
| 298 |
+
"learning_rate": 5e-05,
|
| 299 |
+
"loss": 1.0206,
|
| 300 |
"step": 3700
|
| 301 |
},
|
| 302 |
{
|
| 303 |
+
"epoch": 12.22,
|
| 304 |
+
"learning_rate": 5e-05,
|
| 305 |
+
"loss": 1.0168,
|
| 306 |
"step": 3800
|
| 307 |
},
|
| 308 |
{
|
| 309 |
+
"epoch": 12.54,
|
| 310 |
+
"learning_rate": 5e-05,
|
| 311 |
+
"loss": 1.0269,
|
| 312 |
"step": 3900
|
| 313 |
},
|
| 314 |
{
|
| 315 |
+
"epoch": 12.86,
|
| 316 |
+
"learning_rate": 5e-05,
|
| 317 |
+
"loss": 1.0249,
|
| 318 |
"step": 4000
|
| 319 |
},
|
| 320 |
{
|
| 321 |
+
"epoch": 12.86,
|
| 322 |
+
"eval_cer": 0.04486361342791843,
|
| 323 |
+
"eval_loss": 0.15519459545612335,
|
| 324 |
+
"eval_runtime": 197.5966,
|
| 325 |
+
"eval_samples_per_second": 29.363,
|
| 326 |
+
"eval_steps_per_second": 0.461,
|
| 327 |
+
"eval_wer": 0.23413091698522817,
|
| 328 |
"step": 4000
|
| 329 |
},
|
| 330 |
{
|
| 331 |
+
"epoch": 13.18,
|
| 332 |
+
"learning_rate": 5e-05,
|
| 333 |
+
"loss": 1.022,
|
| 334 |
"step": 4100
|
| 335 |
},
|
| 336 |
{
|
| 337 |
+
"epoch": 13.5,
|
| 338 |
+
"learning_rate": 5e-05,
|
| 339 |
+
"loss": 1.0219,
|
| 340 |
"step": 4200
|
| 341 |
},
|
| 342 |
{
|
| 343 |
+
"epoch": 13.83,
|
| 344 |
+
"learning_rate": 5e-05,
|
| 345 |
+
"loss": 1.0203,
|
| 346 |
"step": 4300
|
| 347 |
},
|
| 348 |
{
|
| 349 |
+
"epoch": 14.15,
|
| 350 |
+
"learning_rate": 5e-05,
|
| 351 |
+
"loss": 1.0171,
|
| 352 |
"step": 4400
|
| 353 |
},
|
| 354 |
{
|
| 355 |
+
"epoch": 14.47,
|
| 356 |
+
"learning_rate": 5e-05,
|
| 357 |
+
"loss": 1.016,
|
| 358 |
"step": 4500
|
| 359 |
},
|
| 360 |
{
|
| 361 |
+
"epoch": 14.47,
|
| 362 |
+
"eval_cer": 0.047286367279095305,
|
| 363 |
+
"eval_loss": 0.16016805171966553,
|
| 364 |
+
"eval_runtime": 197.4133,
|
| 365 |
+
"eval_samples_per_second": 29.39,
|
| 366 |
+
"eval_steps_per_second": 0.461,
|
| 367 |
+
"eval_wer": 0.2435461796215017,
|
| 368 |
"step": 4500
|
| 369 |
},
|
| 370 |
{
|
| 371 |
+
"epoch": 14.79,
|
| 372 |
+
"learning_rate": 5e-05,
|
| 373 |
+
"loss": 1.0233,
|
| 374 |
"step": 4600
|
| 375 |
},
|
| 376 |
{
|
| 377 |
+
"epoch": 15.11,
|
| 378 |
+
"learning_rate": 5e-05,
|
| 379 |
+
"loss": 1.0139,
|
| 380 |
"step": 4700
|
| 381 |
},
|
| 382 |
{
|
| 383 |
+
"epoch": 15.43,
|
| 384 |
+
"learning_rate": 5e-05,
|
| 385 |
+
"loss": 1.0252,
|
| 386 |
"step": 4800
|
| 387 |
},
|
| 388 |
{
|
| 389 |
+
"epoch": 15.76,
|
| 390 |
+
"learning_rate": 4.936666666666667e-05,
|
| 391 |
+
"loss": 1.0305,
|
| 392 |
"step": 4900
|
| 393 |
},
|
| 394 |
{
|
| 395 |
+
"epoch": 16.08,
|
| 396 |
+
"learning_rate": 4.870694444444445e-05,
|
| 397 |
+
"loss": 1.0164,
|
| 398 |
"step": 5000
|
| 399 |
},
|
| 400 |
{
|
| 401 |
+
"epoch": 16.08,
|
| 402 |
+
"eval_cer": 0.044392419340476684,
|
| 403 |
+
"eval_loss": 0.14910832047462463,
|
| 404 |
+
"eval_runtime": 205.8325,
|
| 405 |
+
"eval_samples_per_second": 28.188,
|
| 406 |
+
"eval_steps_per_second": 0.442,
|
| 407 |
+
"eval_wer": 0.23372976544433433,
|
| 408 |
"step": 5000
|
| 409 |
},
|
| 410 |
{
|
| 411 |
+
"epoch": 16.4,
|
| 412 |
+
"learning_rate": 4.804722222222223e-05,
|
| 413 |
+
"loss": 1.0029,
|
| 414 |
"step": 5100
|
| 415 |
},
|
| 416 |
{
|
| 417 |
+
"epoch": 16.72,
|
| 418 |
+
"learning_rate": 4.73875e-05,
|
| 419 |
+
"loss": 0.9924,
|
| 420 |
"step": 5200
|
| 421 |
},
|
| 422 |
{
|
| 423 |
+
"epoch": 17.04,
|
| 424 |
+
"learning_rate": 4.672777777777778e-05,
|
| 425 |
+
"loss": 1.0058,
|
| 426 |
"step": 5300
|
| 427 |
},
|
| 428 |
{
|
| 429 |
+
"epoch": 17.36,
|
| 430 |
+
"learning_rate": 4.606805555555556e-05,
|
| 431 |
+
"loss": 0.996,
|
| 432 |
"step": 5400
|
| 433 |
},
|
| 434 |
{
|
| 435 |
+
"epoch": 17.68,
|
| 436 |
+
"learning_rate": 4.540833333333334e-05,
|
| 437 |
+
"loss": 0.9935,
|
| 438 |
"step": 5500
|
| 439 |
},
|
| 440 |
{
|
| 441 |
+
"epoch": 17.68,
|
| 442 |
+
"eval_cer": 0.045754058947493396,
|
| 443 |
+
"eval_loss": 0.15390604734420776,
|
| 444 |
+
"eval_runtime": 206.7044,
|
| 445 |
+
"eval_samples_per_second": 28.069,
|
| 446 |
+
"eval_steps_per_second": 0.44,
|
| 447 |
+
"eval_wer": 0.23729293501345036,
|
| 448 |
"step": 5500
|
| 449 |
},
|
| 450 |
{
|
| 451 |
+
"epoch": 18.01,
|
| 452 |
+
"learning_rate": 4.4748611111111116e-05,
|
| 453 |
+
"loss": 0.9993,
|
| 454 |
"step": 5600
|
| 455 |
},
|
| 456 |
{
|
| 457 |
+
"epoch": 18.33,
|
| 458 |
+
"learning_rate": 4.408888888888889e-05,
|
| 459 |
+
"loss": 0.983,
|
| 460 |
"step": 5700
|
| 461 |
},
|
| 462 |
{
|
| 463 |
+
"epoch": 18.65,
|
| 464 |
+
"learning_rate": 4.342916666666667e-05,
|
| 465 |
+
"loss": 0.9794,
|
| 466 |
"step": 5800
|
| 467 |
},
|
| 468 |
{
|
| 469 |
+
"epoch": 18.97,
|
| 470 |
+
"learning_rate": 4.2769444444444447e-05,
|
| 471 |
+
"loss": 0.9719,
|
| 472 |
"step": 5900
|
| 473 |
},
|
| 474 |
{
|
| 475 |
+
"epoch": 19.29,
|
| 476 |
+
"learning_rate": 4.2109722222222226e-05,
|
| 477 |
+
"loss": 0.9626,
|
| 478 |
"step": 6000
|
| 479 |
},
|
| 480 |
{
|
| 481 |
+
"epoch": 19.29,
|
| 482 |
+
"eval_cer": 0.04342777002760381,
|
| 483 |
+
"eval_loss": 0.1458132266998291,
|
| 484 |
+
"eval_runtime": 201.2355,
|
| 485 |
+
"eval_samples_per_second": 28.832,
|
| 486 |
+
"eval_steps_per_second": 0.452,
|
| 487 |
+
"eval_wer": 0.2305441502666478,
|
| 488 |
"step": 6000
|
| 489 |
},
|
| 490 |
{
|
| 491 |
+
"epoch": 19.61,
|
| 492 |
+
"learning_rate": 4.145e-05,
|
| 493 |
+
"loss": 0.9542,
|
| 494 |
"step": 6100
|
| 495 |
},
|
| 496 |
{
|
| 497 |
+
"epoch": 19.94,
|
| 498 |
+
"learning_rate": 4.079027777777778e-05,
|
| 499 |
+
"loss": 0.978,
|
| 500 |
"step": 6200
|
| 501 |
},
|
| 502 |
{
|
| 503 |
+
"epoch": 20.26,
|
| 504 |
+
"learning_rate": 4.013055555555556e-05,
|
| 505 |
+
"loss": 0.9536,
|
| 506 |
"step": 6300
|
| 507 |
},
|
| 508 |
{
|
| 509 |
+
"epoch": 20.58,
|
| 510 |
+
"learning_rate": 3.9470833333333335e-05,
|
| 511 |
+
"loss": 0.9627,
|
| 512 |
"step": 6400
|
| 513 |
},
|
| 514 |
{
|
| 515 |
+
"epoch": 20.9,
|
| 516 |
+
"learning_rate": 3.8811111111111114e-05,
|
| 517 |
+
"loss": 0.9505,
|
| 518 |
"step": 6500
|
| 519 |
},
|
| 520 |
{
|
| 521 |
+
"epoch": 20.9,
|
| 522 |
+
"eval_cer": 0.04073046214122466,
|
| 523 |
+
"eval_loss": 0.13684287667274475,
|
| 524 |
+
"eval_runtime": 202.0319,
|
| 525 |
+
"eval_samples_per_second": 28.718,
|
| 526 |
+
"eval_steps_per_second": 0.45,
|
| 527 |
+
"eval_wer": 0.21565434895464627,
|
| 528 |
"step": 6500
|
| 529 |
},
|
| 530 |
{
|
| 531 |
+
"epoch": 21.22,
|
| 532 |
+
"learning_rate": 3.815138888888889e-05,
|
| 533 |
+
"loss": 0.9395,
|
| 534 |
"step": 6600
|
| 535 |
},
|
| 536 |
{
|
| 537 |
+
"epoch": 21.54,
|
| 538 |
+
"learning_rate": 3.749166666666667e-05,
|
| 539 |
+
"loss": 0.9393,
|
| 540 |
"step": 6700
|
| 541 |
},
|
| 542 |
{
|
| 543 |
+
"epoch": 21.86,
|
| 544 |
+
"learning_rate": 3.6831944444444444e-05,
|
| 545 |
+
"loss": 0.9541,
|
| 546 |
"step": 6800
|
| 547 |
},
|
| 548 |
{
|
| 549 |
+
"epoch": 22.19,
|
| 550 |
+
"learning_rate": 3.6172222222222224e-05,
|
| 551 |
+
"loss": 0.9538,
|
| 552 |
"step": 6900
|
| 553 |
},
|
| 554 |
{
|
| 555 |
+
"epoch": 22.51,
|
| 556 |
+
"learning_rate": 3.55125e-05,
|
| 557 |
+
"loss": 0.9389,
|
| 558 |
"step": 7000
|
| 559 |
},
|
| 560 |
{
|
| 561 |
+
"epoch": 22.51,
|
| 562 |
+
"eval_cer": 0.042626369059986347,
|
| 563 |
+
"eval_loss": 0.14371351897716522,
|
| 564 |
+
"eval_runtime": 197.7954,
|
| 565 |
+
"eval_samples_per_second": 29.333,
|
| 566 |
+
"eval_steps_per_second": 0.46,
|
| 567 |
+
"eval_wer": 0.22306385388645053,
|
| 568 |
"step": 7000
|
| 569 |
},
|
| 570 |
{
|
| 571 |
+
"epoch": 22.83,
|
| 572 |
+
"learning_rate": 3.485277777777778e-05,
|
| 573 |
+
"loss": 0.9429,
|
| 574 |
"step": 7100
|
| 575 |
},
|
| 576 |
{
|
| 577 |
+
"epoch": 23.15,
|
| 578 |
+
"learning_rate": 3.419965277777778e-05,
|
| 579 |
+
"loss": 0.9407,
|
| 580 |
"step": 7200
|
| 581 |
},
|
| 582 |
{
|
| 583 |
+
"epoch": 23.47,
|
| 584 |
+
"learning_rate": 3.353993055555556e-05,
|
| 585 |
+
"loss": 0.9224,
|
| 586 |
"step": 7300
|
| 587 |
},
|
| 588 |
{
|
| 589 |
+
"epoch": 23.79,
|
| 590 |
+
"learning_rate": 3.288020833333334e-05,
|
| 591 |
+
"loss": 0.9197,
|
| 592 |
"step": 7400
|
| 593 |
},
|
| 594 |
{
|
| 595 |
+
"epoch": 24.12,
|
| 596 |
+
"learning_rate": 3.2220486111111115e-05,
|
| 597 |
+
"loss": 0.9129,
|
| 598 |
"step": 7500
|
| 599 |
},
|
| 600 |
{
|
| 601 |
+
"epoch": 24.12,
|
| 602 |
+
"eval_cer": 0.039372532723872845,
|
| 603 |
+
"eval_loss": 0.13133755326271057,
|
| 604 |
+
"eval_runtime": 209.4773,
|
| 605 |
+
"eval_samples_per_second": 27.698,
|
| 606 |
+
"eval_steps_per_second": 0.434,
|
| 607 |
+
"eval_wer": 0.20760772098730473,
|
| 608 |
"step": 7500
|
| 609 |
},
|
| 610 |
{
|
| 611 |
+
"epoch": 24.44,
|
| 612 |
+
"learning_rate": 3.156076388888889e-05,
|
| 613 |
+
"loss": 0.9169,
|
| 614 |
"step": 7600
|
| 615 |
},
|
| 616 |
{
|
| 617 |
+
"epoch": 24.76,
|
| 618 |
+
"learning_rate": 3.090763888888889e-05,
|
| 619 |
+
"loss": 0.9133,
|
| 620 |
"step": 7700
|
| 621 |
},
|
| 622 |
{
|
| 623 |
+
"epoch": 25.08,
|
| 624 |
+
"learning_rate": 3.024791666666667e-05,
|
| 625 |
+
"loss": 0.9068,
|
| 626 |
"step": 7800
|
| 627 |
},
|
| 628 |
{
|
| 629 |
+
"epoch": 25.4,
|
| 630 |
+
"learning_rate": 2.958819444444445e-05,
|
| 631 |
+
"loss": 0.9137,
|
| 632 |
"step": 7900
|
| 633 |
},
|
| 634 |
{
|
| 635 |
+
"epoch": 25.72,
|
| 636 |
+
"learning_rate": 2.8928472222222224e-05,
|
| 637 |
+
"loss": 0.9118,
|
| 638 |
"step": 8000
|
| 639 |
},
|
| 640 |
{
|
| 641 |
+
"epoch": 25.72,
|
| 642 |
+
"eval_cer": 0.03844869549731382,
|
| 643 |
+
"eval_loss": 0.12918178737163544,
|
| 644 |
+
"eval_runtime": 197.6149,
|
| 645 |
+
"eval_samples_per_second": 29.36,
|
| 646 |
+
"eval_steps_per_second": 0.46,
|
| 647 |
+
"eval_wer": 0.2040445514181887,
|
| 648 |
"step": 8000
|
| 649 |
},
|
| 650 |
{
|
| 651 |
+
"epoch": 26.05,
|
| 652 |
+
"learning_rate": 2.826875e-05,
|
| 653 |
+
"loss": 0.9057,
|
| 654 |
"step": 8100
|
| 655 |
},
|
| 656 |
{
|
| 657 |
+
"epoch": 26.37,
|
| 658 |
+
"learning_rate": 2.7609027777777785e-05,
|
| 659 |
+
"loss": 0.8956,
|
| 660 |
"step": 8200
|
| 661 |
},
|
| 662 |
{
|
| 663 |
+
"epoch": 26.69,
|
| 664 |
+
"learning_rate": 2.694930555555556e-05,
|
| 665 |
+
"loss": 0.9088,
|
| 666 |
"step": 8300
|
| 667 |
},
|
| 668 |
{
|
| 669 |
+
"epoch": 27.01,
|
| 670 |
+
"learning_rate": 2.6289583333333333e-05,
|
| 671 |
+
"loss": 0.8997,
|
| 672 |
"step": 8400
|
| 673 |
},
|
| 674 |
{
|
| 675 |
+
"epoch": 27.33,
|
| 676 |
+
"learning_rate": 2.5629861111111116e-05,
|
| 677 |
+
"loss": 0.8848,
|
| 678 |
"step": 8500
|
| 679 |
},
|
| 680 |
{
|
| 681 |
+
"epoch": 27.33,
|
| 682 |
+
"eval_cer": 0.03840788341099997,
|
| 683 |
+
"eval_loss": 0.1298777312040329,
|
| 684 |
+
"eval_runtime": 197.318,
|
| 685 |
+
"eval_samples_per_second": 29.404,
|
| 686 |
+
"eval_steps_per_second": 0.461,
|
| 687 |
+
"eval_wer": 0.20281749964604276,
|
| 688 |
"step": 8500
|
| 689 |
},
|
| 690 |
{
|
| 691 |
+
"epoch": 27.65,
|
| 692 |
+
"learning_rate": 2.4970138888888895e-05,
|
| 693 |
+
"loss": 0.8926,
|
| 694 |
"step": 8600
|
| 695 |
},
|
| 696 |
{
|
| 697 |
+
"epoch": 27.97,
|
| 698 |
+
"learning_rate": 2.431041666666667e-05,
|
| 699 |
+
"loss": 0.8802,
|
| 700 |
"step": 8700
|
| 701 |
},
|
| 702 |
{
|
| 703 |
+
"epoch": 28.3,
|
| 704 |
+
"learning_rate": 2.365069444444445e-05,
|
| 705 |
+
"loss": 0.8784,
|
| 706 |
"step": 8800
|
| 707 |
},
|
| 708 |
{
|
| 709 |
+
"epoch": 28.62,
|
| 710 |
+
"learning_rate": 2.2990972222222225e-05,
|
| 711 |
+
"loss": 0.8749,
|
| 712 |
"step": 8900
|
| 713 |
},
|
| 714 |
{
|
| 715 |
+
"epoch": 28.94,
|
| 716 |
+
"learning_rate": 2.2331250000000004e-05,
|
| 717 |
+
"loss": 0.8667,
|
| 718 |
"step": 9000
|
| 719 |
},
|
| 720 |
{
|
| 721 |
+
"epoch": 28.94,
|
| 722 |
+
"eval_cer": 0.03673829806179692,
|
| 723 |
+
"eval_loss": 0.12283530086278915,
|
| 724 |
+
"eval_runtime": 199.3855,
|
| 725 |
+
"eval_samples_per_second": 29.099,
|
| 726 |
+
"eval_steps_per_second": 0.456,
|
| 727 |
+
"eval_wer": 0.1945113030345934,
|
| 728 |
"step": 9000
|
| 729 |
},
|
| 730 |
{
|
| 731 |
+
"epoch": 29.26,
|
| 732 |
+
"learning_rate": 2.1671527777777783e-05,
|
| 733 |
+
"loss": 0.8628,
|
| 734 |
"step": 9100
|
| 735 |
},
|
| 736 |
{
|
| 737 |
+
"epoch": 29.58,
|
| 738 |
+
"learning_rate": 2.101180555555556e-05,
|
| 739 |
+
"loss": 0.8775,
|
| 740 |
"step": 9200
|
| 741 |
},
|
| 742 |
{
|
| 743 |
+
"epoch": 29.9,
|
| 744 |
+
"learning_rate": 2.0352083333333338e-05,
|
| 745 |
+
"loss": 0.8661,
|
| 746 |
"step": 9300
|
| 747 |
},
|
| 748 |
{
|
| 749 |
+
"epoch": 30.23,
|
| 750 |
+
"learning_rate": 1.9692361111111114e-05,
|
| 751 |
+
"loss": 0.8624,
|
| 752 |
"step": 9400
|
| 753 |
},
|
| 754 |
{
|
| 755 |
+
"epoch": 30.55,
|
| 756 |
+
"learning_rate": 1.9032638888888893e-05,
|
| 757 |
+
"loss": 0.8641,
|
| 758 |
"step": 9500
|
| 759 |
},
|
| 760 |
{
|
| 761 |
+
"epoch": 30.55,
|
| 762 |
+
"eval_cer": 0.036352438336647766,
|
| 763 |
+
"eval_loss": 0.12234856933355331,
|
| 764 |
+
"eval_runtime": 202.2537,
|
| 765 |
+
"eval_samples_per_second": 28.687,
|
| 766 |
+
"eval_steps_per_second": 0.45,
|
| 767 |
+
"eval_wer": 0.19385058284959178,
|
| 768 |
"step": 9500
|
| 769 |
},
|
| 770 |
{
|
| 771 |
+
"epoch": 30.87,
|
| 772 |
+
"learning_rate": 1.837291666666667e-05,
|
| 773 |
+
"loss": 0.8637,
|
| 774 |
"step": 9600
|
| 775 |
},
|
| 776 |
{
|
| 777 |
+
"epoch": 31.19,
|
| 778 |
+
"learning_rate": 1.7713194444444447e-05,
|
| 779 |
+
"loss": 0.8608,
|
| 780 |
"step": 9700
|
| 781 |
},
|
| 782 |
{
|
| 783 |
+
"epoch": 31.51,
|
| 784 |
+
"learning_rate": 1.7053472222222226e-05,
|
| 785 |
+
"loss": 0.8556,
|
| 786 |
"step": 9800
|
| 787 |
},
|
| 788 |
{
|
| 789 |
+
"epoch": 31.83,
|
| 790 |
+
"learning_rate": 1.6393750000000002e-05,
|
| 791 |
+
"loss": 0.854,
|
| 792 |
"step": 9900
|
| 793 |
},
|
| 794 |
{
|
| 795 |
+
"epoch": 32.15,
|
| 796 |
+
"learning_rate": 1.573402777777778e-05,
|
| 797 |
+
"loss": 0.8516,
|
| 798 |
"step": 10000
|
| 799 |
},
|
| 800 |
{
|
| 801 |
+
"epoch": 32.15,
|
| 802 |
+
"eval_cer": 0.03494627645365231,
|
| 803 |
+
"eval_loss": 0.11841931194067001,
|
| 804 |
+
"eval_runtime": 199.2371,
|
| 805 |
+
"eval_samples_per_second": 29.121,
|
| 806 |
+
"eval_steps_per_second": 0.457,
|
| 807 |
+
"eval_wer": 0.18762093539100477,
|
| 808 |
"step": 10000
|
| 809 |
},
|
| 810 |
{
|
| 811 |
+
"epoch": 32.48,
|
| 812 |
+
"learning_rate": 1.5074305555555557e-05,
|
| 813 |
+
"loss": 0.8433,
|
| 814 |
+
"step": 10100
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 32.8,
|
| 818 |
+
"learning_rate": 1.4414583333333338e-05,
|
| 819 |
+
"loss": 0.8507,
|
| 820 |
+
"step": 10200
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 33.12,
|
| 824 |
+
"learning_rate": 1.3754861111111117e-05,
|
| 825 |
+
"loss": 0.8419,
|
| 826 |
+
"step": 10300
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 33.44,
|
| 830 |
+
"learning_rate": 1.3095138888888892e-05,
|
| 831 |
+
"loss": 0.8344,
|
| 832 |
+
"step": 10400
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 33.76,
|
| 836 |
+
"learning_rate": 1.2435416666666671e-05,
|
| 837 |
+
"loss": 0.8379,
|
| 838 |
+
"step": 10500
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 33.76,
|
| 842 |
+
"eval_cer": 0.03375159538155591,
|
| 843 |
+
"eval_loss": 0.11372008919715881,
|
| 844 |
+
"eval_runtime": 199.4785,
|
| 845 |
+
"eval_samples_per_second": 29.086,
|
| 846 |
+
"eval_steps_per_second": 0.456,
|
| 847 |
+
"eval_wer": 0.18207560526688377,
|
| 848 |
+
"step": 10500
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 34.08,
|
| 852 |
+
"learning_rate": 1.1782291666666672e-05,
|
| 853 |
+
"loss": 0.8302,
|
| 854 |
+
"step": 10600
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 34.41,
|
| 858 |
+
"learning_rate": 1.1122569444444448e-05,
|
| 859 |
+
"loss": 0.8294,
|
| 860 |
+
"step": 10700
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 34.73,
|
| 864 |
+
"learning_rate": 1.0462847222222227e-05,
|
| 865 |
+
"loss": 0.8225,
|
| 866 |
+
"step": 10800
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 35.05,
|
| 870 |
+
"learning_rate": 9.803125000000001e-06,
|
| 871 |
+
"loss": 0.8237,
|
| 872 |
+
"step": 10900
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 35.37,
|
| 876 |
+
"learning_rate": 9.143402777777782e-06,
|
| 877 |
+
"loss": 0.8235,
|
| 878 |
+
"step": 11000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 35.37,
|
| 882 |
+
"eval_cer": 0.03308005105220979,
|
| 883 |
+
"eval_loss": 0.11269930005073547,
|
| 884 |
+
"eval_runtime": 198.8276,
|
| 885 |
+
"eval_samples_per_second": 29.181,
|
| 886 |
+
"eval_steps_per_second": 0.458,
|
| 887 |
+
"eval_wer": 0.1778753126622304,
|
| 888 |
+
"step": 11000
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 35.69,
|
| 892 |
+
"learning_rate": 8.483680555555563e-06,
|
| 893 |
+
"loss": 0.8205,
|
| 894 |
+
"step": 11100
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 36.01,
|
| 898 |
+
"learning_rate": 7.823958333333337e-06,
|
| 899 |
+
"loss": 0.826,
|
| 900 |
+
"step": 11200
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 36.33,
|
| 904 |
+
"learning_rate": 7.1642361111111165e-06,
|
| 905 |
+
"loss": 0.8207,
|
| 906 |
+
"step": 11300
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 36.66,
|
| 910 |
+
"learning_rate": 6.504513888888891e-06,
|
| 911 |
+
"loss": 0.8129,
|
| 912 |
+
"step": 11400
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 36.98,
|
| 916 |
+
"learning_rate": 5.844791666666671e-06,
|
| 917 |
+
"loss": 0.8112,
|
| 918 |
+
"step": 11500
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 36.98,
|
| 922 |
+
"eval_cer": 0.03268677094773085,
|
| 923 |
+
"eval_loss": 0.11033473163843155,
|
| 924 |
+
"eval_runtime": 201.8103,
|
| 925 |
+
"eval_samples_per_second": 28.75,
|
| 926 |
+
"eval_steps_per_second": 0.451,
|
| 927 |
+
"eval_wer": 0.17662466374062014,
|
| 928 |
+
"step": 11500
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 37.3,
|
| 932 |
+
"learning_rate": 5.185069444444451e-06,
|
| 933 |
+
"loss": 0.805,
|
| 934 |
+
"step": 11600
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 37.62,
|
| 938 |
+
"learning_rate": 4.525347222222226e-06,
|
| 939 |
+
"loss": 0.8108,
|
| 940 |
+
"step": 11700
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 37.94,
|
| 944 |
+
"learning_rate": 3.865625000000006e-06,
|
| 945 |
+
"loss": 0.8025,
|
| 946 |
+
"step": 11800
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 38.26,
|
| 950 |
+
"learning_rate": 3.2059027777777807e-06,
|
| 951 |
+
"loss": 0.8018,
|
| 952 |
+
"step": 11900
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 38.59,
|
| 956 |
+
"learning_rate": 2.5461805555555606e-06,
|
| 957 |
+
"loss": 0.8069,
|
| 958 |
+
"step": 12000
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 38.59,
|
| 962 |
+
"eval_cer": 0.032260099136267845,
|
| 963 |
+
"eval_loss": 0.10924588888883591,
|
| 964 |
+
"eval_runtime": 199.7196,
|
| 965 |
+
"eval_samples_per_second": 29.051,
|
| 966 |
+
"eval_steps_per_second": 0.456,
|
| 967 |
+
"eval_wer": 0.17520883477275945,
|
| 968 |
+
"step": 12000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 38.59,
|
| 972 |
+
"step": 12000,
|
| 973 |
+
"total_flos": 1.0363087195555613e+21,
|
| 974 |
+
"train_loss": 1.0786900800069172,
|
| 975 |
+
"train_runtime": 133237.4383,
|
| 976 |
+
"train_samples_per_second": 11.528,
|
| 977 |
+
"train_steps_per_second": 0.09
|
| 978 |
}
|
| 979 |
],
|
| 980 |
+
"max_steps": 12000,
|
| 981 |
+
"num_train_epochs": 39,
|
| 982 |
+
"total_flos": 1.0363087195555613e+21,
|
| 983 |
"trial_name": null,
|
| 984 |
"trial_params": null
|
| 985 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3055
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1bc1a8adc9b904ef39fbc4de06d5ac7b060a28d098cdb7eb65adcb5cd1d34935
|
| 3 |
size 3055
|