emotion_model

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.3497
• Accuracy: 0.6

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0823	1.0	10	2.0560	0.1625
2.0479	2.0	20	2.0218	0.2812
1.9636	3.0	30	1.8882	0.4062
1.7902	4.0	40	1.6881	0.4313
1.5792	5.0	50	1.6159	0.3688
1.4429	6.0	60	1.3871	0.5687
1.2854	7.0	70	1.2973	0.5437
1.1487	8.0	80	1.2303	0.6
1.0374	9.0	90	1.2661	0.5375
0.9584	10.0	100	1.1662	0.5563
0.8108	11.0	110	1.2135	0.5312
0.7402	12.0	120	1.2117	0.5813
0.6349	13.0	130	1.1176	0.6062
0.5674	14.0	140	1.1794	0.575
0.5103	15.0	150	1.0948	0.6375
0.4826	16.0	160	1.1833	0.5875
0.4128	17.0	170	1.2601	0.5375
0.3664	18.0	180	1.3378	0.55
0.3112	19.0	190	1.2789	0.5437
0.335	20.0	200	1.2913	0.5625
0.3261	21.0	210	1.1114	0.6
0.3443	22.0	220	1.2177	0.5938
0.2642	23.0	230	1.2299	0.5938
0.2895	24.0	240	1.2339	0.5813
0.266	25.0	250	1.2384	0.5875
0.2725	26.0	260	1.2100	0.6062
0.2725	27.0	270	1.3073	0.575
0.2637	28.0	280	1.3019	0.5875
0.2561	29.0	290	1.3597	0.5437
0.2375	30.0	300	1.3404	0.5563
0.2188	31.0	310	1.2922	0.5813
0.2141	32.0	320	1.3778	0.5312
0.198	33.0	330	1.3473	0.5875
0.1805	34.0	340	1.3984	0.5437
0.1888	35.0	350	1.3508	0.5813
0.1867	36.0	360	1.3531	0.575
0.1596	37.0	370	1.5846	0.4875
0.1564	38.0	380	1.3380	0.5687
0.1719	39.0	390	1.5206	0.5312
0.1678	40.0	400	1.2929	0.5875
0.136	41.0	410	1.5031	0.55
0.1602	42.0	420	1.3855	0.5625
0.174	43.0	430	1.4385	0.5875
0.179	44.0	440	1.3153	0.575
0.1284	45.0	450	1.4295	0.5875
0.1419	46.0	460	1.4126	0.575
0.1425	47.0	470	1.3760	0.5687
0.1602	48.0	480	1.4374	0.5875
0.1473	49.0	490	1.3126	0.5813
0.153	50.0	500	1.3497	0.6

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3