12/13/2022 14:12:12 - WARNING - __main__ - Process rank: -1, device: cuda:0, n_gpu: 1distributed training: False, 16-bits training: True
12/13/2022 14:12:12 - INFO - __main__ - Training/evaluation parameters Seq2SeqTrainingArguments(
_n_gpu=1,
adafactor=False,
adam_beta1=0.9,
adam_beta2=0.999,
adam_epsilon=1e-08,
auto_find_batch_size=False,
bf16=False,
bf16_full_eval=False,
data_seed=None,
dataloader_drop_last=False,
dataloader_num_workers=0,
dataloader_pin_memory=True,
ddp_bucket_cap_mb=None,
ddp_find_unused_parameters=None,
ddp_timeout=1800,
debug=[],
deepspeed=None,
disable_tqdm=False,
do_eval=True,
do_predict=False,
do_train=True,
eval_accumulation_steps=None,
eval_delay=0,
eval_steps=1000,
evaluation_strategy=steps,
fp16=True,
fp16_backend=auto,
fp16_full_eval=False,
fp16_opt_level=O1,
fsdp=[],
fsdp_min_num_params=0,
fsdp_transformer_layer_cls_to_wrap=None,
full_determinism=False,
generation_max_length=225,
generation_num_beams=None,
gradient_accumulation_steps=8,
gradient_checkpointing=True,
greater_is_better=False,
group_by_length=False,
half_precision_backend=auto,
hub_model_id=None,
hub_private_repo=False,
hub_strategy=every_save,
hub_token=<HUB_TOKEN>,
ignore_data_skip=False,
include_inputs_for_metrics=False,
jit_mode_eval=False,
label_names=None,
label_smoothing_factor=0.0,
learning_rate=1e-05,
length_column_name=input_length,
load_best_model_at_end=True,
local_rank=-1,
log_level=passive,
log_level_replica=passive,
log_on_each_node=True,
logging_dir=./runs/Dec13_14-12-12_4b942bf2873e,
logging_first_step=False,
logging_nan_inf_filter=True,
logging_steps=25,
logging_strategy=steps,
lr_scheduler_type=linear,
max_grad_norm=1.0,
max_steps=5000,
metric_for_best_model=wer,
mp_parameters=,
no_cuda=False,
num_train_epochs=3.0,
optim=adamw_hf,
optim_args=None,
output_dir=./,
overwrite_output_dir=True,
past_index=-1,
per_device_eval_batch_size=8,
per_device_train_batch_size=8,
predict_with_generate=True,
prediction_loss_only=False,
push_to_hub=True,
push_to_hub_model_id=None,
push_to_hub_organization=None,
push_to_hub_token=<PUSH_TO_HUB_TOKEN>,
ray_scope=last,
remove_unused_columns=True,
report_to=['tensorboard'],
resume_from_checkpoint=None,
run_name=./,
save_on_each_node=False,
save_steps=1000,
save_strategy=steps,
save_total_limit=None,
seed=42,
sharded_ddp=[],
skip_memory_metrics=True,
sortish_sampler=False,
tf32=None,
torchdynamo=None,
tpu_metrics_debug=False,
tpu_num_cores=None,
use_ipex=False,
use_legacy_prediction_loop=False,
use_mps_device=False,
warmup_ratio=0.0,
warmup_steps=500,
weight_decay=0.0,
xpu_backend=None,
)
12/13/2022 14:12:12 - INFO - __main__ - Training/evaluation parameters Seq2SeqTrainingArguments(
_n_gpu=1,
adafactor=False,
adam_beta1=0.9,
adam_beta2=0.999,
adam_epsilon=1e-08,
auto_find_batch_size=False,
bf16=False,
bf16_full_eval=False,
data_seed=None,
dataloader_drop_last=False,
dataloader_num_workers=0,
dataloader_pin_memory=True,
ddp_bucket_cap_mb=None,
ddp_find_unused_parameters=None,
ddp_timeout=1800,
debug=[],
deepspeed=None,
disable_tqdm=False,
do_eval=True,
do_predict=False,
do_train=True,
eval_accumulation_steps=None,
eval_delay=0,
eval_steps=1000,
evaluation_strategy=steps,
fp16=True,
fp16_backend=auto,
fp16_full_eval=False,
fp16_opt_level=O1,
fsdp=[],
fsdp_min_num_params=0,
fsdp_transformer_layer_cls_to_wrap=None,
full_determinism=False,
generation_max_length=225,
generation_num_beams=None,
gradient_accumulation_steps=8,
gradient_checkpointing=True,
greater_is_better=False,
group_by_length=False,
half_precision_backend=auto,
hub_model_id=None,
hub_private_repo=False,
hub_strategy=every_save,
hub_token=<HUB_TOKEN>,
ignore_data_skip=False,
include_inputs_for_metrics=False,
jit_mode_eval=False,
label_names=None,
label_smoothing_factor=0.0,
learning_rate=1e-05,
length_column_name=input_length,
load_best_model_at_end=True,
local_rank=-1,
log_level=passive,
log_level_replica=passive,
log_on_each_node=True,
logging_dir=./runs/Dec13_14-12-12_4b942bf2873e,
logging_first_step=False,
logging_nan_inf_filter=True,
logging_steps=25,
logging_strategy=steps,
lr_scheduler_type=linear,
max_grad_norm=1.0,
max_steps=5000,
metric_for_best_model=wer,
mp_parameters=,
no_cuda=False,
num_train_epochs=3.0,
optim=adamw_hf,
optim_args=None,
output_dir=./,
overwrite_output_dir=True,
past_index=-1,
per_device_eval_batch_size=8,
per_device_train_batch_size=8,
predict_with_generate=True,
prediction_loss_only=False,
push_to_hub=True,
push_to_hub_model_id=None,
push_to_hub_organization=None,
push_to_hub_token=<PUSH_TO_HUB_TOKEN>,
ray_scope=last,
remove_unused_columns=True,
report_to=['tensorboard'],
resume_from_checkpoint=None,
run_name=./,
save_on_each_node=False,
save_steps=1000,
save_strategy=steps,
save_total_limit=None,
seed=42,
sharded_ddp=[],
skip_memory_metrics=True,
sortish_sampler=False,
tf32=None,
torchdynamo=None,
tpu_metrics_debug=False,
tpu_num_cores=None,
use_ipex=False,
use_legacy_prediction_loop=False,
use_mps_device=False,
warmup_ratio=0.0,
warmup_steps=500,
weight_decay=0.0,
xpu_backend=None,
)
12/13/2022 14:12:15 - INFO - datasets.info - Loading Dataset Infos from /root/.cache/huggingface/modules/datasets_modules/datasets/mozilla-foundation--common_voice_11_0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:15 - INFO - datasets.builder - Overwrite dataset info from restored data version.
12/13/2022 14:12:15 - INFO - datasets.info - Loading Dataset info from /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:15 - WARNING - datasets.builder - Found cached dataset common_voice_11_0 (/root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f)
12/13/2022 14:12:15 - INFO - datasets.info - Loading Dataset info from /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:19 - INFO - datasets.info - Loading Dataset Infos from /root/.cache/huggingface/modules/datasets_modules/datasets/mozilla-foundation--common_voice_11_0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:19 - INFO - datasets.builder - Overwrite dataset info from restored data version.
12/13/2022 14:12:19 - INFO - datasets.info - Loading Dataset info from /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:19 - WARNING - datasets.builder - Found cached dataset common_voice_11_0 (/root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f)
12/13/2022 14:12:19 - INFO - datasets.info - Loading Dataset info from /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:22 - INFO - datasets.info - Loading Dataset Infos from /root/.cache/huggingface/modules/datasets_modules/datasets/mozilla-foundation--common_voice_11_0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:22 - INFO - datasets.builder - Overwrite dataset info from restored data version.
12/13/2022 14:12:22 - INFO - datasets.info - Loading Dataset info from /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:12:22 - WARNING - datasets.builder - Found cached dataset common_voice_11_0 (/root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f)
12/13/2022 14:12:22 - INFO - datasets.info - Loading Dataset info from /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f
12/13/2022 14:14:43 - INFO - datasets.arrow_dataset - Caching processed dataset at /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f/cache-f4ebdf4955d452c1.arrow
12/13/2022 15:03:44 - INFO - datasets.arrow_dataset - Caching processed dataset at /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f/cache-0bdc3bb096868b69.arrow
12/13/2022 15:31:19 - INFO - datasets.arrow_dataset - Caching processed dataset at /root/.cache/huggingface/datasets/mozilla-foundation___common_voice_11_0/cs/11.0.0/f8e47235d9b4e68fa24ed71d63266a02018ccf7194b2a8c9c598a5f3ab304d9f/cache-beeaca47803ef85b.arrow
12/13/2022 16:01:48 - WARNING - huggingface_hub.repository - /usr/src/app/models/whisper-large2-czech-cv11/./ is already a clone of https://huggingface.co/mikr/whisper-large2-czech-cv11. Make sure you pull the latest changes with `repo.git_pull()`.
{'loss': 0.7151, 'learning_rate': 4.6000000000000004e-07, 'epoch': 0.11}
{'loss': 0.4297, 'learning_rate': 9.400000000000001e-07, 'epoch': 0.21}
{'loss': 0.2399, 'learning_rate': 1.42e-06, 'epoch': 0.32}
{'loss': 0.1871, 'learning_rate': 1.9200000000000003e-06, 'epoch': 0.42}
{'loss': 0.161, 'learning_rate': 2.42e-06, 'epoch': 0.53}
{'loss': 0.1511, 'learning_rate': 2.92e-06, 'epoch': 0.64}
{'loss': 0.1481, 'learning_rate': 3.4200000000000007e-06, 'epoch': 0.74}
{'loss': 0.1392, 'learning_rate': 3.920000000000001e-06, 'epoch': 0.85}
{'loss': 0.1328, 'learning_rate': 4.42e-06, 'epoch': 0.95}
{'loss': 0.1152, 'learning_rate': 4.92e-06, 'epoch': 1.06}
{'loss': 0.0998, 'learning_rate': 5.420000000000001e-06, 'epoch': 1.17}
{'loss': 0.0936, 'learning_rate': 5.92e-06, 'epoch': 1.28}
{'loss': 0.0953, 'learning_rate': 6.4000000000000006e-06, 'epoch': 1.38}
{'loss': 0.0962, 'learning_rate': 6.9e-06, 'epoch': 1.49}
{'loss': 0.0971, 'learning_rate': 7.4e-06, 'epoch': 1.59}
{'loss': 0.1003, 'learning_rate': 7.9e-06, 'epoch': 1.7}
{'loss': 0.0931, 'learning_rate': 8.400000000000001e-06, 'epoch': 1.81}
{'loss': 0.0925, 'learning_rate': 8.900000000000001e-06, 'epoch': 1.91}
{'loss': 0.092, 'learning_rate': 9.4e-06, 'epoch': 2.02}
{'loss': 0.0509, 'learning_rate': 9.9e-06, 'epoch': 2.13}
{'loss': 0.0563, 'learning_rate': 9.955555555555556e-06, 'epoch': 2.23}
{'loss': 0.0563, 'learning_rate': 9.9e-06, 'epoch': 2.34}
{'loss': 0.0581, 'learning_rate': 9.844444444444446e-06, 'epoch': 2.45}
{'loss': 0.0557, 'learning_rate': 9.78888888888889e-06, 'epoch': 2.55}
{'loss': 0.0561, 'learning_rate': 9.733333333333334e-06, 'epoch': 2.66}
{'loss': 0.0548, 'learning_rate': 9.677777777777778e-06, 'epoch': 2.76}
{'loss': 0.056, 'learning_rate': 9.622222222222222e-06, 'epoch': 2.87}
{'loss': 0.0582, 'learning_rate': 9.566666666666668e-06, 'epoch': 2.98}
{'loss': 0.0341, 'learning_rate': 9.511111111111112e-06, 'epoch': 3.08}
{'loss': 0.0256, 'learning_rate': 9.455555555555557e-06, 'epoch': 3.19}
{'loss': 0.0263, 'learning_rate': 9.4e-06, 'epoch': 3.3}
{'loss': 0.0275, 'learning_rate': 9.344444444444446e-06, 'epoch': 3.4}
{'loss': 0.0289, 'learning_rate': 9.28888888888889e-06, 'epoch': 3.51}
{'loss': 0.0274, 'learning_rate': 9.233333333333334e-06, 'epoch': 3.62}
{'loss': 0.0309, 'learning_rate': 9.17777777777778e-06, 'epoch': 3.72}
{'loss': 0.0307, 'learning_rate': 9.122222222222223e-06, 'epoch': 3.83}
{'loss': 0.0304, 'learning_rate': 9.066666666666667e-06, 'epoch': 3.93}
{'loss': 0.026, 'learning_rate': 9.011111111111111e-06, 'epoch': 4.04}
{'loss': 0.0139, 'learning_rate': 8.955555555555555e-06, 'epoch': 4.15}
{'loss': 0.0149, 'learning_rate': 8.900000000000001e-06, 'epoch': 4.25}
{'eval_loss': 0.16215506196022034, 'eval_wer': 10.040330748973316, 'eval_runtime': 17422.9164, 'eval_samples_per_second': 0.443, 'eval_steps_per_second': 0.055, 'epoch': 4.25}
{'loss': 0.0141, 'learning_rate': 8.844444444444445e-06, 'epoch': 4.36}
{'loss': 0.0159, 'learning_rate': 8.788888888888891e-06, 'epoch': 4.47}
{'loss': 0.0165, 'learning_rate': 8.733333333333333e-06, 'epoch': 4.57}
{'loss': 0.0134, 'learning_rate': 8.677777777777779e-06, 'epoch': 4.68}
{'loss': 0.0174, 'learning_rate': 8.622222222222223e-06, 'epoch': 4.78}
{'loss': 0.0147, 'learning_rate': 8.566666666666667e-06, 'epoch': 4.89}
{'loss': 0.0179, 'learning_rate': 8.511111111111113e-06, 'epoch': 5.0}
{'loss': 0.01, 'learning_rate': 8.455555555555555e-06, 'epoch': 5.11}
{'loss': 0.0094, 'learning_rate': 8.400000000000001e-06, 'epoch': 5.21}
{'loss': 0.0087, 'learning_rate': 8.344444444444445e-06, 'epoch': 5.32}
{'loss': 0.0087, 'learning_rate': 8.288888888888889e-06, 'epoch': 5.42}
{'loss': 0.0113, 'learning_rate': 8.233333333333335e-06, 'epoch': 5.53}
{'loss': 0.0126, 'learning_rate': 8.177777777777779e-06, 'epoch': 5.64}
{'loss': 0.0107, 'learning_rate': 8.122222222222223e-06, 'epoch': 5.74}
{'loss': 0.0105, 'learning_rate': 8.066666666666667e-06, 'epoch': 5.85}
{'loss': 0.0104, 'learning_rate': 8.011111111111113e-06, 'epoch': 5.95}
{'loss': 0.0082, 'learning_rate': 7.955555555555557e-06, 'epoch': 6.06}
{'loss': 0.0056, 'learning_rate': 7.9e-06, 'epoch': 6.17}
{'loss': 0.0056, 'learning_rate': 7.844444444444446e-06, 'epoch': 6.28}
{'loss': 0.007, 'learning_rate': 7.788888888888889e-06, 'epoch': 6.38}
{'loss': 0.0062, 'learning_rate': 7.733333333333334e-06, 'epoch': 6.49}
{'loss': 0.0058, 'learning_rate': 7.677777777777778e-06, 'epoch': 6.59}
{'loss': 0.0061, 'learning_rate': 7.622222222222223e-06, 'epoch': 6.7}
{'loss': 0.0055, 'learning_rate': 7.566666666666667e-06, 'epoch': 6.81}
{'loss': 0.006, 'learning_rate': 7.511111111111111e-06, 'epoch': 6.91}
{'loss': 0.0063, 'learning_rate': 7.455555555555556e-06, 'epoch': 7.02}
{'loss': 0.0043, 'learning_rate': 7.4e-06, 'epoch': 7.13}
{'loss': 0.0035, 'learning_rate': 7.344444444444445e-06, 'epoch': 7.23}
{'loss': 0.0035, 'learning_rate': 7.28888888888889e-06, 'epoch': 7.34}
{'loss': 0.004, 'learning_rate': 7.233333333333334e-06, 'epoch': 7.45}
{'loss': 0.0034, 'learning_rate': 7.177777777777778e-06, 'epoch': 7.55}
{'loss': 0.0042, 'learning_rate': 7.122222222222222e-06, 'epoch': 7.66}
{'loss': 0.0042, 'learning_rate': 7.066666666666667e-06, 'epoch': 7.76}
{'loss': 0.0042, 'learning_rate': 7.011111111111112e-06, 'epoch': 7.87}
{'loss': 0.0043, 'learning_rate': 6.955555555555557e-06, 'epoch': 7.98}
{'loss': 0.0034, 'learning_rate': 6.9e-06, 'epoch': 8.08}
{'loss': 0.0029, 'learning_rate': 6.844444444444445e-06, 'epoch': 8.19}
{'loss': 0.0034, 'learning_rate': 6.788888888888889e-06, 'epoch': 8.3}
{'loss': 0.0033, 'learning_rate': 6.733333333333334e-06, 'epoch': 8.4}
{'loss': 0.0027, 'learning_rate': 6.677777777777779e-06, 'epoch': 8.51}
{'eval_loss': 0.18483087420463562, 'eval_wer': 9.51363694959577, 'eval_runtime': 21193.0144, 'eval_samples_per_second': 0.364, 'eval_steps_per_second': 0.046, 'epoch': 8.51}
{'loss': 0.0042, 'learning_rate': 6.6222222222222236e-06, 'epoch': 8.62}
{'loss': 0.0042, 'learning_rate': 6.566666666666667e-06, 'epoch': 8.72}
{'loss': 0.0046, 'learning_rate': 6.511111111111112e-06, 'epoch': 8.83}
{'loss': 0.004, 'learning_rate': 6.455555555555556e-06, 'epoch': 8.93}
{'loss': 0.004, 'learning_rate': 6.4000000000000006e-06, 'epoch': 9.04}
{'loss': 0.0029, 'learning_rate': 6.3444444444444454e-06, 'epoch': 9.15}
{'loss': 0.0027, 'learning_rate': 6.28888888888889e-06, 'epoch': 9.25}
{'loss': 0.003, 'learning_rate': 6.2333333333333335e-06, 'epoch': 9.36}
{'loss': 0.003, 'learning_rate': 6.177777777777778e-06, 'epoch': 9.47}
{'loss': 0.0031, 'learning_rate': 6.1222222222222224e-06, 'epoch': 9.57}
{'loss': 0.0028, 'learning_rate': 6.066666666666667e-06, 'epoch': 9.68}
{'loss': 0.0029, 'learning_rate': 6.011111111111112e-06, 'epoch': 9.78}
{'loss': 0.003, 'learning_rate': 5.955555555555555e-06, 'epoch': 9.89}
{'loss': 0.0035, 'learning_rate': 5.9e-06, 'epoch': 10.0}
{'loss': 0.0023, 'learning_rate': 5.844444444444445e-06, 'epoch': 10.11}
{'loss': 0.0019, 'learning_rate': 5.788888888888889e-06, 'epoch': 10.21}
{'loss': 0.002, 'learning_rate': 5.733333333333334e-06, 'epoch': 10.32}
{'loss': 0.0024, 'learning_rate': 5.677777777777779e-06, 'epoch': 10.42}
{'loss': 0.0021, 'learning_rate': 5.622222222222222e-06, 'epoch': 10.53}
{'loss': 0.0021, 'learning_rate': 5.566666666666667e-06, 'epoch': 10.64}
{'loss': 0.0019, 'learning_rate': 5.511111111111112e-06, 'epoch': 10.74}
{'loss': 0.002, 'learning_rate': 5.455555555555556e-06, 'epoch': 10.85}
{'loss': 0.0018, 'learning_rate': 5.400000000000001e-06, 'epoch': 10.95}
{'loss': 0.0017, 'learning_rate': 5.344444444444446e-06, 'epoch': 11.06}
{'loss': 0.0024, 'learning_rate': 5.288888888888889e-06, 'epoch': 11.17}
{'loss': 0.0019, 'learning_rate': 5.233333333333334e-06, 'epoch': 11.28}
{'loss': 0.0018, 'learning_rate': 5.177777777777779e-06, 'epoch': 11.38}
{'loss': 0.0017, 'learning_rate': 5.122222222222223e-06, 'epoch': 11.49}
{'loss': 0.0014, 'learning_rate': 5.0666666666666676e-06, 'epoch': 11.59}
{'loss': 0.0014, 'learning_rate': 5.011111111111111e-06, 'epoch': 11.7}
{'loss': 0.0015, 'learning_rate': 4.9555555555555565e-06, 'epoch': 11.81}
{'loss': 0.0017, 'learning_rate': 4.9000000000000005e-06, 'epoch': 11.91}
{'loss': 0.0016, 'learning_rate': 4.8444444444444446e-06, 'epoch': 12.02}
{'loss': 0.001, 'learning_rate': 4.7888888888888894e-06, 'epoch': 12.13}
{'loss': 0.001, 'learning_rate': 4.7333333333333335e-06, 'epoch': 12.23}
{'loss': 0.0018, 'learning_rate': 4.677777777777778e-06, 'epoch': 12.34}
{'loss': 0.0014, 'learning_rate': 4.622222222222222e-06, 'epoch': 12.45}
{'loss': 0.0012, 'learning_rate': 4.566666666666667e-06, 'epoch': 12.55}
{'loss': 0.0012, 'learning_rate': 4.511111111111111e-06, 'epoch': 12.66}
{'loss': 0.0008, 'learning_rate': 4.455555555555555e-06, 'epoch': 12.76}
{'eval_loss': 0.19296467304229736, 'eval_wer': 9.31658717150697, 'eval_runtime': 16619.4936, 'eval_samples_per_second': 0.464, 'eval_steps_per_second': 0.058, 'epoch': 12.76}
{'loss': 0.0009, 'learning_rate': 4.4e-06, 'epoch': 12.87}
{'loss': 0.001, 'learning_rate': 4.344444444444445e-06, 'epoch': 12.98}
{'loss': 0.0007, 'learning_rate': 4.288888888888889e-06, 'epoch': 13.08}
{'loss': 0.0007, 'learning_rate': 4.233333333333334e-06, 'epoch': 13.19}
{'loss': 0.0009, 'learning_rate': 4.177777777777778e-06, 'epoch': 13.3}
{'loss': 0.0006, 'learning_rate': 4.122222222222222e-06, 'epoch': 13.4}
{'loss': 0.0007, 'learning_rate': 4.066666666666667e-06, 'epoch': 13.51}
{'loss': 0.0006, 'learning_rate': 4.011111111111111e-06, 'epoch': 13.62}
{'loss': 0.0008, 'learning_rate': 3.955555555555556e-06, 'epoch': 13.72}
{'loss': 0.0007, 'learning_rate': 3.900000000000001e-06, 'epoch': 13.83}
{'loss': 0.0007, 'learning_rate': 3.844444444444445e-06, 'epoch': 13.93}
{'loss': 0.0006, 'learning_rate': 3.7888888888888893e-06, 'epoch': 14.04}
{'loss': 0.0005, 'learning_rate': 3.7333333333333337e-06, 'epoch': 14.15}
{'loss': 0.0005, 'learning_rate': 3.6777777777777778e-06, 'epoch': 14.25}
{'loss': 0.0005, 'learning_rate': 3.6222222222222226e-06, 'epoch': 14.36}
{'loss': 0.0005, 'learning_rate': 3.566666666666667e-06, 'epoch': 14.47}
{'loss': 0.0005, 'learning_rate': 3.511111111111111e-06, 'epoch': 14.57}
{'loss': 0.0005, 'learning_rate': 3.455555555555556e-06, 'epoch': 14.68}
{'loss': 0.0011, 'learning_rate': 3.4000000000000005e-06, 'epoch': 14.78}
{'loss': 0.0005, 'learning_rate': 3.3444444444444445e-06, 'epoch': 14.89}
{'loss': 0.0008, 'learning_rate': 3.2888888888888894e-06, 'epoch': 15.0}
{'loss': 0.0008, 'learning_rate': 3.2333333333333334e-06, 'epoch': 15.11}
{'loss': 0.0005, 'learning_rate': 3.177777777777778e-06, 'epoch': 15.21}
{'loss': 0.0007, 'learning_rate': 3.1222222222222228e-06, 'epoch': 15.32}
{'loss': 0.0005, 'learning_rate': 3.066666666666667e-06, 'epoch': 15.42}
{'loss': 0.0005, 'learning_rate': 3.0111111111111113e-06, 'epoch': 15.53}
{'loss': 0.0009, 'learning_rate': 2.955555555555556e-06, 'epoch': 15.64}
{'loss': 0.0005, 'learning_rate': 2.9e-06, 'epoch': 15.74}
{'loss': 0.0005, 'learning_rate': 2.8444444444444446e-06, 'epoch': 15.85}
{'loss': 0.0006, 'learning_rate': 2.788888888888889e-06, 'epoch': 15.95}
{'loss': 0.0005, 'learning_rate': 2.7333333333333336e-06, 'epoch': 16.06}
{'loss': 0.0004, 'learning_rate': 2.677777777777778e-06, 'epoch': 16.17}
{'loss': 0.0007, 'learning_rate': 2.6222222222222225e-06, 'epoch': 16.28}
{'loss': 0.0005, 'learning_rate': 2.566666666666667e-06, 'epoch': 16.38}
{'loss': 0.0005, 'learning_rate': 2.5111111111111114e-06, 'epoch': 16.49}
{'loss': 0.0004, 'learning_rate': 2.455555555555556e-06, 'epoch': 16.59}
{'loss': 0.0005, 'learning_rate': 2.4000000000000003e-06, 'epoch': 16.7}
{'loss': 0.0004, 'learning_rate': 2.3444444444444448e-06, 'epoch': 16.81}
{'loss': 0.0004, 'learning_rate': 2.2888888888888892e-06, 'epoch': 16.91}
{'loss': 0.0004, 'learning_rate': 2.2333333333333333e-06, 'epoch': 17.02}
{'eval_loss': 0.20616495609283447, 'eval_wer': 9.032982817995986, 'eval_runtime': 12625.6787, 'eval_samples_per_second': 0.611, 'eval_steps_per_second': 0.076, 'epoch': 17.02}
{'loss': 0.0004, 'learning_rate': 2.1777777777777777e-06, 'epoch': 17.13}
{'loss': 0.0004, 'learning_rate': 2.1222222222222226e-06, 'epoch': 17.23}
{'loss': 0.0004, 'learning_rate': 2.0666666666666666e-06, 'epoch': 17.34}
{'loss': 0.0004, 'learning_rate': 2.011111111111111e-06, 'epoch': 17.45}
{'loss': 0.0004, 'learning_rate': 1.955555555555556e-06, 'epoch': 17.55}
{'loss': 0.0004, 'learning_rate': 1.9000000000000002e-06, 'epoch': 17.66}
{'loss': 0.0004, 'learning_rate': 1.8444444444444445e-06, 'epoch': 17.76}
{'loss': 0.0004, 'learning_rate': 1.788888888888889e-06, 'epoch': 17.87}
{'loss': 0.0004, 'learning_rate': 1.7333333333333336e-06, 'epoch': 17.98}
{'loss': 0.0004, 'learning_rate': 1.6777777777777779e-06, 'epoch': 18.08}
{'loss': 0.0004, 'learning_rate': 1.6222222222222223e-06, 'epoch': 18.19}
{'loss': 0.0004, 'learning_rate': 1.566666666666667e-06, 'epoch': 18.3}
{'loss': 0.0003, 'learning_rate': 1.5111111111111112e-06, 'epoch': 18.4}
{'loss': 0.0003, 'learning_rate': 1.4555555555555557e-06, 'epoch': 18.51}
{'loss': 0.0004, 'learning_rate': 1.4000000000000001e-06, 'epoch': 18.62}
{'loss': 0.0003, 'learning_rate': 1.3444444444444446e-06, 'epoch': 18.72}
{'loss': 0.0003, 'learning_rate': 1.288888888888889e-06, 'epoch': 18.83}
{'loss': 0.0004, 'learning_rate': 1.2333333333333335e-06, 'epoch': 18.93}
{'loss': 0.0003, 'learning_rate': 1.1777777777777778e-06, 'epoch': 19.04}
{'loss': 0.0003, 'learning_rate': 1.1222222222222222e-06, 'epoch': 19.15}
{'loss': 0.0003, 'learning_rate': 1.066666666666667e-06, 'epoch': 19.25}
{'loss': 0.0003, 'learning_rate': 1.0111111111111111e-06, 'epoch': 19.36}
{'loss': 0.0003, 'learning_rate': 9.555555555555556e-07, 'epoch': 19.47}
{'loss': 0.0003, 'learning_rate': 9.000000000000001e-07, 'epoch': 19.57}
{'loss': 0.0003, 'learning_rate': 8.444444444444445e-07, 'epoch': 19.68}
{'loss': 0.0003, 'learning_rate': 7.888888888888889e-07, 'epoch': 19.78}
{'loss': 0.0003, 'learning_rate': 7.333333333333334e-07, 'epoch': 19.89}
{'loss': 0.0003, 'learning_rate': 6.777777777777779e-07, 'epoch': 20.0}
{'loss': 0.0003, 'learning_rate': 6.222222222222223e-07, 'epoch': 20.11}
{'loss': 0.0003, 'learning_rate': 5.666666666666667e-07, 'epoch': 20.21}
{'loss': 0.0003, 'learning_rate': 5.111111111111112e-07, 'epoch': 20.32}
{'loss': 0.0003, 'learning_rate': 4.5555555555555563e-07, 'epoch': 20.42}
{'loss': 0.0003, 'learning_rate': 4.0000000000000003e-07, 'epoch': 20.53}
{'loss': 0.0003, 'learning_rate': 3.444444444444445e-07, 'epoch': 20.64}
{'loss': 0.0003, 'learning_rate': 2.888888888888889e-07, 'epoch': 20.74}
{'loss': 0.0003, 'learning_rate': 2.3333333333333336e-07, 'epoch': 20.85}
{'loss': 0.0003, 'learning_rate': 1.777777777777778e-07, 'epoch': 20.95}
{'loss': 0.0003, 'learning_rate': 1.2222222222222225e-07, 'epoch': 21.06}
{'loss': 0.0003, 'learning_rate': 6.666666666666668e-08, 'epoch': 21.17}
{'loss': 0.0003, 'learning_rate': 1.1111111111111112e-08, 'epoch': 21.28}