SER_model_xapiens_binary

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: 1.9238
• Accuracy: 0.6121

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	9	0.6576	0.6186
0.6373	2.0	18	0.6570	0.6239
0.6356	3.0	27	0.6528	0.6252
0.6255	4.0	36	0.6525	0.6212
0.6275	5.0	45	0.6459	0.6239
0.62	6.0	54	0.6564	0.6252
0.6079	7.0	63	0.6575	0.6173
0.6066	8.0	72	0.6592	0.6304
0.6032	9.0	81	0.6540	0.6265
0.5894	10.0	90	0.6553	0.6252
0.5894	11.0	99	0.6764	0.6239
0.5707	12.0	108	0.6768	0.6370
0.5665	13.0	117	0.6877	0.6330
0.5478	14.0	126	0.7169	0.6134
0.5303	15.0	135	0.7437	0.5767
0.5113	16.0	144	0.7235	0.6252
0.4867	17.0	153	0.7914	0.6055
0.4525	18.0	162	0.7906	0.6121
0.4339	19.0	171	0.7946	0.6029
0.4231	20.0	180	0.7982	0.6121
0.4231	21.0	189	0.7991	0.6278
0.3913	22.0	198	0.8757	0.6252
0.3967	23.0	207	0.8746	0.6042
0.3615	24.0	216	0.8731	0.5976
0.3354	25.0	225	0.9350	0.6173
0.3288	26.0	234	0.9604	0.6042
0.3087	27.0	243	0.9534	0.6199
0.2977	28.0	252	0.9491	0.6107
0.2657	29.0	261	1.0096	0.6134
0.2614	30.0	270	1.0069	0.6042
0.2614	31.0	279	1.0388	0.5990
0.2309	32.0	288	1.1028	0.5950
0.2155	33.0	297	1.1211	0.6239
0.2162	34.0	306	1.1376	0.5767
0.2055	35.0	315	1.2133	0.5806
0.1949	36.0	324	1.1371	0.5963
0.1897	37.0	333	1.2791	0.5793
0.1715	38.0	342	1.2176	0.5924
0.1622	39.0	351	1.2985	0.5832
0.1571	40.0	360	1.3201	0.5832
0.1571	41.0	369	1.3364	0.5976
0.1518	42.0	378	1.2600	0.6173
0.1715	43.0	387	1.2334	0.6134
0.1702	44.0	396	1.2732	0.6055
0.1292	45.0	405	1.3833	0.5963
0.1104	46.0	414	1.4520	0.5990
0.1572	47.0	423	1.3802	0.6291
0.132	48.0	432	1.4400	0.6003
0.1528	49.0	441	1.3557	0.6212
0.1219	50.0	450	1.3858	0.5911
0.1219	51.0	459	1.4453	0.5937
0.1014	52.0	468	1.3600	0.6212
0.0951	53.0	477	1.4042	0.6186
0.0962	54.0	486	1.4200	0.6199
0.0937	55.0	495	1.4017	0.6291
0.1005	56.0	504	1.4544	0.6081
0.0832	57.0	513	1.4527	0.6252
0.0896	58.0	522	1.4970	0.6068
0.0754	59.0	531	1.5413	0.6016
0.076	60.0	540	1.5303	0.6186
0.076	61.0	549	1.5076	0.6134
0.0945	62.0	558	1.4725	0.6225
0.0636	63.0	567	1.5103	0.6160
0.0866	64.0	576	1.5384	0.6304
0.0558	65.0	585	1.5868	0.6081
0.0607	66.0	594	1.5560	0.6265
0.0595	67.0	603	1.6161	0.6212
0.0566	68.0	612	1.6855	0.5911
0.042	69.0	621	1.6898	0.6186
0.0506	70.0	630	1.6505	0.6330
0.0506	71.0	639	1.6898	0.6107
0.048	72.0	648	1.6370	0.6225
0.0564	73.0	657	1.6576	0.6147
0.0512	74.0	666	1.7149	0.6003
0.0482	75.0	675	1.7085	0.6199
0.0351	76.0	684	1.7413	0.6134
0.0432	77.0	693	1.7627	0.6016
0.0368	78.0	702	1.7961	0.6291
0.0416	79.0	711	1.7539	0.6199
0.0398	80.0	720	1.7764	0.6160
0.0398	81.0	729	1.8516	0.6252
0.0311	82.0	738	1.8629	0.6199
0.0327	83.0	747	1.8564	0.6147
0.0404	84.0	756	1.8447	0.6291
0.0407	85.0	765	1.8198	0.6173
0.0432	86.0	774	1.8477	0.6029
0.0367	87.0	783	1.8972	0.6317
0.0334	88.0	792	1.8497	0.6121
0.0228	89.0	801	1.8638	0.6199
0.0351	90.0	810	1.8128	0.6107
0.0351	91.0	819	1.8479	0.6186
0.0342	92.0	828	1.8849	0.6225
0.0257	93.0	837	1.9021	0.6160
0.0288	94.0	846	1.9284	0.6147
0.0241	95.0	855	1.9315	0.6107
0.0237	96.0	864	1.9235	0.5976
0.0231	97.0	873	1.9428	0.6225
0.0217	98.0	882	1.9266	0.6199
0.0215	99.0	891	1.9238	0.6160
0.0207	100.0	900	1.9238	0.6121

Framework versions

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