convnextv2-tiny-1k-224-finetuned-sleeve-length

This model is a fine-tuned version of facebook/convnextv2-tiny-1k-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.5496
• Accuracy: 0.8621

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.96	6	1.7957	0.2299
1.8656	1.92	12	1.7704	0.2759
1.8656	2.88	18	1.7382	0.3218
1.7835	4.0	25	1.6674	0.3793
1.664	4.96	31	1.5982	0.4253
1.664	5.92	37	1.4861	0.4368
1.5072	6.88	43	1.3645	0.4713
1.3304	8.0	50	1.2859	0.4598
1.3304	8.96	56	1.2796	0.4713
1.1651	9.92	62	1.2456	0.5172
1.1651	10.88	68	1.1667	0.5402
1.0876	12.0	75	1.1510	0.5632
1.0046	12.96	81	1.0510	0.6092
1.0046	13.92	87	1.0338	0.5862
0.9465	14.88	93	0.9883	0.5862
0.8699	16.0	100	0.9882	0.5632
0.8699	16.96	106	0.9276	0.5747
0.7969	17.92	112	0.9145	0.5862
0.7969	18.88	118	0.8144	0.6667
0.7254	20.0	125	0.7587	0.6667
0.6447	20.96	131	0.6990	0.7471
0.6447	21.92	137	0.7042	0.7241
0.6021	22.88	143	0.6526	0.7701
0.516	24.0	150	0.6485	0.8046
0.516	24.96	156	0.5803	0.8161
0.4497	25.92	162	0.6085	0.8046
0.4497	26.88	168	0.6095	0.8046
0.3935	28.0	175	0.5372	0.8276
0.3321	28.96	181	0.5829	0.8161
0.3321	29.92	187	0.6205	0.8161
0.3007	30.88	193	0.5150	0.8276
0.2618	32.0	200	0.6069	0.8391
0.2618	32.96	206	0.5273	0.8391
0.2411	33.92	212	0.4727	0.8621
0.2411	34.88	218	0.4611	0.8736
0.2108	36.0	225	0.5696	0.8506
0.2143	36.96	231	0.4944	0.8621
0.2143	37.92	237	0.5628	0.8161
0.1663	38.88	243	0.6131	0.8046
0.1714	40.0	250	0.4962	0.8506
0.1714	40.96	256	0.5023	0.8391
0.174	41.92	262	0.4842	0.8276
0.174	42.88	268	0.4679	0.8276
0.138	44.0	275	0.6271	0.8161
0.1437	44.96	281	0.5326	0.8506
0.1437	45.92	287	0.5655	0.8161
0.136	46.88	293	0.4672	0.8391
0.1401	48.0	300	0.4990	0.8621
0.1401	48.96	306	0.5445	0.8276
0.1281	49.92	312	0.4761	0.8736
0.1281	50.88	318	0.5665	0.8506
0.1156	52.0	325	0.5090	0.8506
0.0981	52.96	331	0.5152	0.8506
0.0981	53.92	337	0.5466	0.8161
0.1055	54.88	343	0.5390	0.8276
0.112	56.0	350	0.5574	0.8506
0.112	56.96	356	0.5449	0.8506
0.0855	57.92	362	0.5390	0.8506
0.0855	58.88	368	0.5206	0.8506
0.0899	60.0	375	0.5476	0.8621
0.1026	60.96	381	0.5344	0.8506
0.1026	61.92	387	0.5531	0.8391
0.0799	62.88	393	0.5723	0.8276
0.0844	64.0	400	0.5340	0.8161
0.0844	64.96	406	0.5236	0.8736
0.0724	65.92	412	0.6137	0.8391
0.0724	66.88	418	0.5825	0.8276
0.0867	68.0	425	0.5105	0.8621
0.071	68.96	431	0.5272	0.8506
0.071	69.92	437	0.5524	0.8506
0.0723	70.88	443	0.5508	0.8391
0.0748	72.0	450	0.5689	0.8161
0.0748	72.96	456	0.5556	0.8506
0.0589	73.92	462	0.5452	0.8506
0.0589	74.88	468	0.5475	0.8621
0.0719	76.0	475	0.5484	0.8621
0.0801	76.8	480	0.5496	0.8621

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1