libri-alpha-1-Temp-1-processor-change

This model is a fine-tuned version of on the None dataset. It achieves the following results on the evaluation set:

• Loss: 107.5632
• Wer: 0.1210

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
800.4601	0.75	100	189.2537	0.1508
680.6167	1.49	200	151.4276	0.1445
596.8935	2.24	300	144.2152	0.1401
576.7275	2.99	400	135.5783	0.1354
558.4325	3.73	500	121.5539	0.1324
512.5202	4.48	600	118.9303	0.1302
480.8142	5.22	700	117.4402	0.1295
480.9125	5.97	800	113.6219	0.1266
425.4858	6.72	900	112.1535	0.1274
443.0105	7.46	1000	112.7184	0.1240
436.3363	8.21	1100	110.7517	0.1268
416.3612	8.96	1200	112.7510	0.1272
418.833	9.7	1300	108.8896	0.1208
417.0277	10.45	1400	108.8765	0.1208
432.6803	11.19	1500	108.7693	0.1213
378.0122	11.94	1600	107.9976	0.1217
407.0373	12.69	1700	107.8557	0.1211
398.8923	13.43	1800	107.5632	0.1210

Framework versions

• Transformers 4.24.0
• Pytorch 1.12.1
• Datasets 2.7.0
• Tokenizers 0.11.0