Yoru1010/sh_model

model_sh_intit_model

This model is a fine-tuned version of facebook/wav2vec2-base-960h on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.4880
• Wer: 0.3617
• Cer: 0.9396

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.9874	20.0	100	1.6657	0.5234	0.9457
0.5576	40.0	200	1.0859	0.4426	0.9422
0.399	60.0	300	1.2627	0.3957	0.9406
0.2509	80.0	400	1.3391	0.3830	0.9405
0.2643	100.0	500	1.4182	0.3787	0.9401
0.1931	120.0	600	1.3800	0.3915	0.9403
0.1553	140.0	700	1.4751	0.3957	0.9402
0.1679	160.0	800	1.4633	0.3660	0.9397
0.1642	180.0	900	1.5003	0.3617	0.9397
0.1286	200.0	1000	1.4880	0.3617	0.9396

Framework versions

• Transformers 4.35.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.7
• Tokenizers 0.14.1