wav2vec2-1b-adapters-mer-drL-v1.2

This model is a fine-tuned version of facebook/mms-1b-all on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: nan
• Wer: 1.0

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: 0.001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.2542	0.02	100	inf	0.6865
1.0706	0.05	200	inf	0.6425
1.054	0.07	300	inf	0.6273
1.131	0.1	400	inf	0.6269
1.091	0.12	500	inf	0.6120
1.209	0.14	600	inf	0.6234
1.032	0.17	700	inf	0.6036
1.0873	0.19	800	inf	0.5984
1.1786	0.21	900	inf	0.6164
1.0065	0.24	1000	inf	0.5984
1.1682	0.26	1100	inf	0.9431
2.716	0.29	1200	inf	1.0
4.6494	0.31	1300	inf	1.0
6.5668	0.33	1400	inf	1.0
8.2033	0.36	1500	inf	0.9997
8.4349	0.38	1600	inf	1.0
8.8253	0.41	1700	inf	1.0000
10.0956	0.43	1800	inf	1.0000
10.1367	0.45	1900	inf	1.0000
10.2369	0.48	2000	inf	1.0000
10.185	0.5	2100	inf	1.0000
10.1101	0.53	2200	inf	1.0000
10.1882	0.55	2300	inf	1.0000
10.1791	0.57	2400	inf	1.0000
10.095	0.6	2500	inf	1.0000
10.0778	0.62	2600	inf	1.0000
10.1407	0.64	2700	inf	1.0000
10.2779	0.67	2800	inf	1.0000
10.0577	0.69	2900	inf	1.0000
10.1342	0.72	3000	inf	1.0000
10.0638	0.74	3100	inf	1.0000
10.19	0.76	3200	inf	1.0000
10.2655	0.79	3300	inf	1.0000
10.1485	0.81	3400	inf	1.0000
10.2903	0.84	3500	inf	1.0000
10.1934	0.86	3600	inf	1.0000
10.038	0.88	3700	inf	1.0000
10.1181	0.91	3800	inf	1.0000
10.1547	0.93	3900	inf	1.0000
10.2849	0.95	4000	inf	1.0000
10.1119	0.98	4100	inf	1.0000
10.2269	1.0	4200	inf	1.0000
10.1069	1.03	4300	inf	1.0000
10.2036	1.05	4400	inf	1.0000
10.1252	1.07	4500	inf	1.0000
10.0869	1.1	4600	inf	1.0000
9.9904	1.12	4700	inf	1.0000
10.1395	1.15	4800	inf	1.0000
10.0352	1.17	4900	inf	1.0000
10.3128	1.19	5000	inf	1.0000
10.1161	1.22	5100	inf	1.0000
10.1318	1.24	5200	inf	1.0000
10.1863	1.27	5300	inf	1.0000
10.1645	1.29	5400	inf	1.0000
10.3267	1.31	5500	inf	1.0000
9.9707	1.34	5600	inf	1.0000
10.2071	1.36	5700	inf	1.0000
10.0865	1.38	5800	inf	1.0000
10.3051	1.41	5900	inf	1.0000
10.203	1.43	6000	inf	1.0000
10.0152	1.46	6100	inf	1.0000
10.1636	1.48	6200	inf	1.0000
10.1885	1.5	6300	inf	1.0000
10.1876	1.53	6400	inf	1.0000
10.1075	1.55	6500	inf	1.0000
10.1307	1.58	6600	inf	1.0000
10.3877	1.6	6700	inf	1.0000
10.1684	1.62	6800	inf	1.0000
10.0601	1.65	6900	inf	1.0000
10.3244	1.67	7000	inf	1.0000
10.2978	1.69	7100	inf	1.0000
10.2394	1.72	7200	inf	1.0000
10.0721	1.74	7300	inf	1.0000
10.1697	1.77	7400	inf	1.0000
10.3378	1.79	7500	inf	1.0000
10.1207	1.81	7600	inf	1.0000
10.1188	1.84	7700	inf	1.0000
10.0966	1.86	7800	inf	1.0000
10.2581	1.89	7900	inf	1.0000
10.219	1.91	8000	inf	1.0000
10.272	1.93	8100	inf	1.0000
10.1932	1.96	8200	inf	1.0000
10.0127	1.98	8300	nan	1.0
0.0	2.01	8400	nan	1.0
0.0	2.03	8500	nan	1.0
0.0	2.05	8600	nan	1.0
0.0	2.08	8700	nan	1.0
0.0	2.1	8800	nan	1.0
0.0	2.12	8900	nan	1.0
0.0	2.15	9000	nan	1.0
0.0	2.17	9100	nan	1.0
0.0	2.2	9200	nan	1.0
0.0	2.22	9300	nan	1.0
0.0	2.24	9400	nan	1.0
0.0	2.27	9500	nan	1.0
0.0	2.29	9600	nan	1.0
0.0	2.32	9700	nan	1.0
0.0	2.34	9800	nan	1.0
0.0	2.36	9900	nan	1.0
0.0	2.39	10000	nan	1.0
0.0	2.41	10100	nan	1.0
0.0	2.43	10200	nan	1.0
0.0	2.46	10300	nan	1.0
0.0	2.48	10400	nan	1.0
0.0	2.51	10500	nan	1.0
0.0	2.53	10600	nan	1.0
0.0	2.55	10700	nan	1.0
0.0	2.58	10800	nan	1.0
0.0	2.6	10900	nan	1.0
0.0	2.63	11000	nan	1.0
0.0	2.65	11100	nan	1.0
0.0	2.67	11200	nan	1.0
0.0	2.7	11300	nan	1.0
0.0	2.72	11400	nan	1.0
0.0	2.75	11500	nan	1.0
0.0	2.77	11600	nan	1.0
0.0	2.79	11700	nan	1.0
0.0	2.82	11800	nan	1.0
0.0	2.84	11900	nan	1.0
0.0	2.86	12000	nan	1.0
0.0	2.89	12100	nan	1.0
0.0	2.91	12200	nan	1.0
0.0	2.94	12300	nan	1.0
0.0	2.96	12400	nan	1.0
0.0	2.98	12500	nan	1.0
0.0	3.01	12600	nan	1.0
0.0	3.03	12700	nan	1.0
0.0	3.06	12800	nan	1.0
0.0	3.08	12900	nan	1.0
0.0	3.1	13000	nan	1.0
0.0	3.13	13100	nan	1.0
0.0	3.15	13200	nan	1.0
0.0	3.17	13300	nan	1.0
0.0	3.2	13400	nan	1.0
0.0	3.22	13500	nan	1.0
0.0	3.25	13600	nan	1.0
0.0	3.27	13700	nan	1.0
0.0	3.29	13800	nan	1.0
0.0	3.32	13900	nan	1.0
0.0	3.34	14000	nan	1.0
0.0	3.37	14100	nan	1.0
0.0	3.39	14200	nan	1.0
0.0	3.41	14300	nan	1.0
0.0	3.44	14400	nan	1.0
0.0	3.46	14500	nan	1.0
0.0	3.49	14600	nan	1.0
0.0	3.51	14700	nan	1.0
0.0	3.53	14800	nan	1.0
0.0	3.56	14900	nan	1.0
0.0	3.58	15000	nan	1.0
0.0	3.6	15100	nan	1.0
0.0	3.63	15200	nan	1.0
0.0	3.65	15300	nan	1.0
0.0	3.68	15400	nan	1.0
0.0	3.7	15500	nan	1.0
0.0	3.72	15600	nan	1.0
0.0	3.75	15700	nan	1.0
0.0	3.77	15800	nan	1.0
0.0	3.8	15900	nan	1.0
0.0	3.82	16000	nan	1.0
0.0	3.84	16100	nan	1.0
0.0	3.87	16200	nan	1.0
0.0	3.89	16300	nan	1.0
0.0	3.92	16400	nan	1.0
0.0	3.94	16500	nan	1.0
0.0	3.96	16600	nan	1.0
0.0	3.99	16700	nan	1.0

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.14.3
• Tokenizers 0.13.3