README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: assis
    results: []

assis

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

• Loss: 0.3440
• Wer: 1

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: 3000
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
16.8292	1.56	100	16.7197	1
15.1534	3.12	200	14.3410	1
10.7755	4.69	300	9.9820	1
6.4859	6.25	400	6.1913	1
4.0464	7.81	500	3.8280	1
3.3418	9.38	600	3.2733	1
3.217	10.94	700	3.1409	1
3.0927	12.5	800	3.0469	1
3.0235	14.06	900	3.0015	1
2.9902	15.62	1000	2.9748	1
2.945	17.19	1100	2.9550	1
2.9293	18.75	1200	2.9262	1
2.9139	20.31	1300	2.9230	1
2.9084	21.88	1400	2.9067	1
2.8941	23.44	1500	2.9077	1
2.8883	25.0	1600	2.8858	1
2.872	26.56	1700	2.8709	1
2.8641	28.12	1800	2.8587	1
2.8548	29.69	1900	2.8537	1
2.8396	31.25	2000	2.8371	1
2.7043	32.81	2100	2.6063	1
2.3905	34.38	2200	2.2233	1
1.9862	35.94	2300	1.7478	1
1.5463	37.5	2400	1.3176	1
1.218	39.06	2500	0.9948	1
0.9606	40.62	2600	0.7820	1
0.7923	42.19	2700	0.6577	1
0.6811	43.75	2800	0.5650	1
0.5927	45.31	2900	0.5204	1
0.5449	46.88	3000	0.4857	1
0.4876	48.44	3100	0.4526	1
0.4646	50.0	3200	0.4281	1
0.4374	51.56	3300	0.4376	1
0.3952	53.12	3400	0.4075	1
0.3952	54.69	3500	0.3937	1
0.3558	56.25	3600	0.3875	1
0.3527	57.81	3700	0.3775	1
0.3349	59.38	3800	0.3701	1
0.3264	60.94	3900	0.3576	1
0.3108	62.5	4000	0.3644	1
0.3104	64.06	4100	0.3548	1
0.3012	65.62	4200	0.3510	1
0.3027	67.19	4300	0.3486	1
0.2967	68.75	4400	0.3431	1
0.2892	70.31	4500	0.3391	1
0.296	71.88	4600	0.3427	1
0.2821	73.44	4700	0.3469	1
0.2701	75.0	4800	0.3428	1
0.2825	76.56	4900	0.3426	1
0.2549	78.12	5000	0.3440	1

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.13.3