You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

WMAC1

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

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 8
eval_batch_size: 12
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 650
training_steps: 5000
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer
2.1976	0.4757	1000	0.5681	68.4004
1.3001	0.9514	2000	0.3560	52.8999
0.9103	1.4267	3000	0.3106	49.7007
0.6961	1.9024	4000	0.2851	46.8008
0.5684	2.3777	5000	0.2915	47.7090

Safetensors

Model size

0.7B params

Tensor type

F32

Finetunes