wav2vec2-base-960h-musiccaps

This model is a fine-tuned version of facebook/wav2vec2-base-960h on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.9625
• Exact Accuracy: 0.0
• Partial Accuracy: 0.6056

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: 0.05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Accuracy	Partial Accuracy
15.7112	0.9	9	15.9578	0.06	0.7733
11.1514	1.8	18	10.0587	0.04	0.7667
6.6287	2.7	27	6.6632	0.0	0.5911
4.2256	3.6	36	1.9485	0.0	0.3878
5.6467	4.5	45	4.0850	0.0	0.4422
5.7963	5.4	54	3.0922	0.0	0.5133
4.2273	6.3	63	4.0553	0.0	0.5933
4.0305	7.2	72	3.7869	0.0	0.4600
3.1243	8.1	81	2.4867	0.21	0.8244
2.1803	9.0	90	1.9611	0.0	0.6822
4.1344	9.9	99	2.0692	0.0	0.4978
2.4153	10.8	108	2.4877	0.0	0.4756
2.558	11.7	117	3.8951	0.0	0.46
3.2152	12.6	126	2.9583	0.0	0.4689
2.7496	13.5	135	2.0802	0.0	0.6833
3.3934	14.4	144	4.6528	0.0	0.5344
3.7926	15.3	153	2.9359	0.0	0.7156
3.7156	16.2	162	1.8099	0.0	0.5856
3.0527	17.1	171	2.1170	0.0	0.5956
3.0344	18.0	180	2.8589	0.0	0.6544
3.3441	18.9	189	2.8075	0.0	0.6089
2.553	19.8	198	2.0358	0.03	0.7456
3.7821	20.7	207	3.4526	0.0	0.5911
3.4076	21.6	216	2.6764	0.0	0.5733
3.1877	22.5	225	1.6248	0.0	0.5678
3.4089	23.4	234	4.6811	0.03	0.7167
4.6228	24.3	243	3.3227	0.0	0.5044
3.5935	25.2	252	3.7636	0.0	0.5267
4.5658	26.1	261	5.2628	0.0	0.6489
4.9328	27.0	270	4.7943	0.0	0.6267
5.6449	27.9	279	4.3986	0.0	0.6589
4.6314	28.8	288	3.1873	0.02	0.74
5.4069	29.7	297	3.5047	0.06	0.7678
5.4739	30.6	306	8.9623	0.0	0.4611
4.9211	31.5	315	4.1060	0.0	0.5644
5.1984	32.4	324	4.0939	0.0	0.4478
4.3039	33.3	333	2.6682	0.0	0.5267
3.5481	34.2	342	1.9895	0.01	0.7156
2.1797	35.1	351	2.6432	0.0	0.6033
2.0572	36.0	360	2.0274	0.0	0.5733
2.4059	36.9	369	1.1969	0.11	0.7989
3.4561	37.8	378	4.1893	0.06	0.7944
2.7969	38.7	387	2.3889	0.0	0.6222
2.3617	39.6	396	2.9330	0.0	0.5333
3.0501	40.5	405	2.8672	0.0	0.6322
3.1292	41.4	414	4.0559	0.0	0.5278
2.95	42.3	423	2.4140	0.0	0.6989
2.8121	43.2	432	3.9192	0.0	0.4622
4.0374	44.1	441	4.3913	0.0	0.6322
3.4575	45.0	450	2.8873	0.0	0.5878
2.8763	45.9	459	6.7531	0.0	0.4956
5.0513	46.8	468	5.0384	0.04	0.7544
4.0507	47.7	477	3.3518	0.0	0.5600
2.479	48.6	486	4.3941	0.0	0.5356
3.752	49.5	495	3.9625	0.0	0.6056

Framework versions

• Transformers 4.49.0
• Pytorch 2.7.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1