SER_model_xapiens

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: 3.0600
• Accuracy: 0.7

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: 3e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	1	2.6350	0.7167
No log	2.0	2	2.7406	0.7167
No log	3.0	3	2.8632	0.7
No log	4.0	4	2.8745	0.7
No log	5.0	5	2.9112	0.7
No log	6.0	6	2.9644	0.6833
No log	7.0	7	3.3638	0.65
No log	8.0	8	3.8127	0.5833
No log	9.0	9	4.2359	0.5833
0.0038	10.0	10	3.7159	0.6167
0.0038	11.0	11	3.2066	0.65
0.0038	12.0	12	2.9436	0.7333
0.0038	13.0	13	3.1083	0.7
0.0038	14.0	14	3.4327	0.6833
0.0038	15.0	15	3.4017	0.6833
0.0038	16.0	16	3.1088	0.7
0.0038	17.0	17	2.9279	0.7
0.0038	18.0	18	2.9036	0.7333
0.0038	19.0	19	2.9245	0.7167
0.0097	20.0	20	2.8291	0.7333
0.0097	21.0	21	2.8655	0.7
0.0097	22.0	22	2.7505	0.7167
0.0097	23.0	23	2.6406	0.7333
0.0097	24.0	24	2.6017	0.7333
0.0097	25.0	25	2.5550	0.75
0.0097	26.0	26	2.5972	0.75
0.0097	27.0	27	2.6193	0.7333
0.0097	28.0	28	2.6949	0.7167
0.0097	29.0	29	2.8405	0.6833
0.0052	30.0	30	2.8552	0.7
0.0052	31.0	31	2.7041	0.7
0.0052	32.0	32	2.5161	0.7333
0.0052	33.0	33	2.4609	0.7333
0.0052	34.0	34	2.4119	0.7333
0.0052	35.0	35	2.3811	0.75
0.0052	36.0	36	2.4323	0.75
0.0052	37.0	37	2.6068	0.75
0.0052	38.0	38	2.6425	0.75
0.0052	39.0	39	2.6534	0.75
0.004	40.0	40	2.6643	0.7333
0.004	41.0	41	2.6912	0.7167
0.004	42.0	42	2.7284	0.7
0.004	43.0	43	2.7621	0.7
0.004	44.0	44	2.7870	0.7
0.004	45.0	45	2.8085	0.7
0.004	46.0	46	2.8251	0.7
0.004	47.0	47	2.8478	0.7167
0.004	48.0	48	2.8729	0.7
0.004	49.0	49	2.8999	0.7
0.0001	50.0	50	2.9256	0.7
0.0001	51.0	51	2.9287	0.7
0.0001	52.0	52	2.8962	0.7
0.0001	53.0	53	2.8783	0.6833
0.0001	54.0	54	2.8959	0.7167
0.0001	55.0	55	2.8606	0.7167
0.0001	56.0	56	2.7337	0.7167
0.0001	57.0	57	2.7450	0.7
0.0001	58.0	58	2.9678	0.6667
0.0001	59.0	59	3.1560	0.6667
0.0145	60.0	60	3.4292	0.6333
0.0145	61.0	61	3.5474	0.6167
0.0145	62.0	62	3.6343	0.6167
0.0145	63.0	63	3.5805	0.6167
0.0145	64.0	64	3.5623	0.6333
0.0145	65.0	65	3.5532	0.6333
0.0145	66.0	66	3.6041	0.6
0.0145	67.0	67	3.7335	0.6
0.0145	68.0	68	3.8447	0.6
0.0145	69.0	69	3.8997	0.6167
0.0001	70.0	70	3.8381	0.6167
0.0001	71.0	71	3.7951	0.6
0.0001	72.0	72	3.8582	0.6
0.0001	73.0	73	3.8983	0.6
0.0001	74.0	74	3.8225	0.6167
0.0001	75.0	75	3.6746	0.65
0.0001	76.0	76	3.5636	0.65
0.0001	77.0	77	3.5198	0.6667
0.0001	78.0	78	3.4888	0.6667
0.0001	79.0	79	3.4752	0.65
0.004	80.0	80	3.4661	0.65
0.004	81.0	81	3.4177	0.65
0.004	82.0	82	3.3299	0.65
0.004	83.0	83	3.2658	0.6667
0.004	84.0	84	3.2386	0.6833
0.004	85.0	85	3.2177	0.7
0.004	86.0	86	3.1985	0.7
0.004	87.0	87	3.1800	0.7
0.004	88.0	88	3.1624	0.7
0.004	89.0	89	3.1451	0.7
0.0001	90.0	90	3.1292	0.7
0.0001	91.0	91	3.1145	0.7
0.0001	92.0	92	3.1022	0.7
0.0001	93.0	93	3.0917	0.7
0.0001	94.0	94	3.0817	0.7
0.0001	95.0	95	3.0753	0.7
0.0001	96.0	96	3.0691	0.7
0.0001	97.0	97	3.0658	0.7
0.0001	98.0	98	3.0623	0.7
0.0001	99.0	99	3.0608	0.7
0.0	100.0	100	3.0600	0.7

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0
• Datasets 2.19.2.dev0
• Tokenizers 0.19.1