ssc-cgg-mms-model-mix-adapt-max3

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 0.0005
train_batch_size: 1
eval_batch_size: 6
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 2
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
num_epochs: 5
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
0.3948	0.2261	200	0.6911	0.1446	0.6222
0.425	0.4522	400	0.6826	0.1449	0.6325
0.394	0.6783	600	0.6796	0.1414	0.6274
0.3473	0.9045	800	0.6964	0.1417	0.6157
0.3528	1.1300	1000	0.6941	0.1418	0.6148
0.3965	1.3561	1200	0.6866	0.1405	0.6097
0.3656	1.5822	1400	0.6896	0.1373	0.6030
0.3588	1.8084	1600	0.6852	0.1412	0.6152
0.3796	2.0339	1800	0.6782	0.1427	0.6140
0.3501	2.2600	2000	0.6840	0.1393	0.6104
0.3358	2.4862	2200	0.6839	0.1381	0.6085
0.3487	2.7123	2400	0.6792	0.1374	0.6011
0.3347	2.9384	2600	0.6824	0.1369	0.6029
0.3354	3.1639	2800	0.6687	0.1370	0.6041
0.3271	3.3901	3000	0.6826	0.1371	0.6009
0.3331	3.6162	3200	0.6689	0.1380	0.5960
0.3165	3.8423	3400	0.6756	0.1357	0.5954
0.3263	4.0678	3600	0.6736	0.1360	0.5956
0.2909	4.2940	3800	0.6691	0.1359	0.5917
0.2644	4.5201	4000	0.6744	0.1356	0.5955
0.3013	4.7462	4200	0.6754	0.1363	0.5961
0.3	4.9723	4400	0.6729	0.1357	0.5953

Safetensors

Model size

1.0B params

Tensor type

F32