convnextv2-tiny-1k-224-finetuned-neck-style

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: 0.6084
• Accuracy: 0.8493
• Precision: 0.8507
• Recall: 0.8493

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: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall
1.613	0.9897	24	1.5833	0.2928	0.3248	0.2928
1.5494	1.9794	48	1.4944	0.3681	0.4410	0.3681
1.3989	2.9691	72	1.3424	0.5159	0.5262	0.5159
1.2238	4.0	97	1.1162	0.6261	0.6666	0.6261
0.9585	4.9897	121	0.8966	0.6986	0.7014	0.6986
0.8934	5.9794	145	0.7638	0.7507	0.7490	0.7507
0.7589	6.9691	169	0.6776	0.7652	0.7719	0.7652
0.6746	8.0	194	0.6127	0.7623	0.7628	0.7623
0.6048	8.9897	218	0.5221	0.8203	0.8217	0.8203
0.531	9.9794	242	0.4931	0.8116	0.8204	0.8116
0.57	10.9691	266	0.4480	0.8319	0.8345	0.8319
0.4624	12.0	291	0.4214	0.8464	0.8460	0.8464
0.417	12.9897	315	0.4439	0.8493	0.8486	0.8493
0.3814	13.9794	339	0.4138	0.8464	0.8478	0.8464
0.3737	14.9691	363	0.4139	0.8464	0.8466	0.8464
0.3971	16.0	388	0.4119	0.8638	0.8665	0.8638
0.343	16.9897	412	0.4421	0.8609	0.8659	0.8609
0.3311	17.9794	436	0.4581	0.8493	0.8504	0.8493
0.2652	18.9691	460	0.4563	0.8406	0.8441	0.8406
0.3026	20.0	485	0.4536	0.8522	0.8549	0.8522
0.2562	20.9897	509	0.4409	0.8464	0.8493	0.8464
0.2282	21.9794	533	0.4389	0.8435	0.8451	0.8435
0.2374	22.9691	557	0.4452	0.8580	0.8589	0.8580
0.216	24.0	582	0.4375	0.8580	0.8581	0.8580
0.2127	24.9897	606	0.4422	0.8580	0.8588	0.8580
0.2004	25.9794	630	0.4635	0.8522	0.8519	0.8522
0.2029	26.9691	654	0.5215	0.8493	0.8546	0.8493
0.1794	28.0	679	0.4756	0.8638	0.8669	0.8638
0.1835	28.9897	703	0.4728	0.8609	0.8650	0.8609
0.1781	29.9794	727	0.4637	0.8551	0.8568	0.8551
0.1671	30.9691	751	0.4856	0.8580	0.8599	0.8580
0.1762	32.0	776	0.5008	0.8667	0.8684	0.8667
0.1867	32.9897	800	0.5058	0.8580	0.8585	0.8580
0.1409	33.9794	824	0.5490	0.8406	0.8409	0.8406
0.1315	34.9691	848	0.5284	0.8348	0.8356	0.8348
0.1315	36.0	873	0.5415	0.8464	0.8488	0.8464
0.1974	36.9897	897	0.5194	0.8493	0.8536	0.8493
0.1337	37.9794	921	0.5088	0.8609	0.8603	0.8609
0.173	38.9691	945	0.4912	0.8667	0.8680	0.8667
0.1409	40.0	970	0.5223	0.8493	0.8502	0.8493
0.1379	40.9897	994	0.5204	0.8493	0.8487	0.8493
0.1437	41.9794	1018	0.5860	0.8522	0.8551	0.8522
0.1022	42.9691	1042	0.5461	0.8464	0.8492	0.8464
0.1181	44.0	1067	0.5411	0.8551	0.8566	0.8551
0.1212	44.9897	1091	0.5294	0.8580	0.8580	0.8580
0.1049	45.9794	1115	0.5667	0.8493	0.8492	0.8493
0.1132	46.9691	1139	0.5908	0.8464	0.8491	0.8464
0.1313	48.0	1164	0.5996	0.8522	0.8582	0.8522
0.1312	48.9897	1188	0.5430	0.8580	0.8607	0.8580
0.0996	49.9794	1212	0.5777	0.8522	0.8561	0.8522
0.1389	50.9691	1236	0.5758	0.8435	0.8486	0.8435
0.1079	52.0	1261	0.5540	0.8580	0.8611	0.8580
0.0972	52.9897	1285	0.5600	0.8551	0.8559	0.8551
0.0985	53.9794	1309	0.5392	0.8638	0.8656	0.8638
0.1112	54.9691	1333	0.5411	0.8638	0.8656	0.8638
0.1308	56.0	1358	0.5445	0.8638	0.8654	0.8638
0.1005	56.9897	1382	0.5554	0.8551	0.8551	0.8551
0.0871	57.9794	1406	0.5966	0.8406	0.8441	0.8406
0.1102	58.9691	1430	0.5807	0.8522	0.8543	0.8522
0.1028	60.0	1455	0.5654	0.8435	0.8491	0.8435
0.107	60.9897	1479	0.5779	0.8435	0.8461	0.8435
0.0848	61.9794	1503	0.5843	0.8551	0.8569	0.8551
0.0976	62.9691	1527	0.6162	0.8435	0.8454	0.8435
0.0977	64.0	1552	0.5822	0.8464	0.8469	0.8464
0.1256	64.9897	1576	0.5757	0.8493	0.8514	0.8493
0.0883	65.9794	1600	0.5716	0.8464	0.8467	0.8464
0.0808	66.9691	1624	0.5726	0.8551	0.8562	0.8551
0.1034	68.0	1649	0.5413	0.8551	0.8549	0.8551
0.0845	68.9897	1673	0.5826	0.8435	0.8477	0.8435
0.0916	69.9794	1697	0.5661	0.8522	0.8522	0.8522
0.0912	70.9691	1721	0.5771	0.8493	0.8498	0.8493
0.0863	72.0	1746	0.5769	0.8551	0.8550	0.8551
0.083	72.9897	1770	0.5860	0.8493	0.8486	0.8493
0.0839	73.9794	1794	0.5647	0.8551	0.8551	0.8551
0.0903	74.9691	1818	0.6012	0.8551	0.8535	0.8551
0.074	76.0	1843	0.6048	0.8464	0.8461	0.8464
0.0907	76.9897	1867	0.5807	0.8493	0.8495	0.8493
0.0613	77.9794	1891	0.5775	0.8377	0.8382	0.8377
0.0964	78.9691	1915	0.5759	0.8667	0.8676	0.8667
0.0735	80.0	1940	0.5962	0.8551	0.8566	0.8551
0.0663	80.9897	1964	0.5769	0.8435	0.8441	0.8435
0.0719	81.9794	1988	0.5826	0.8493	0.8507	0.8493
0.0718	82.9691	2012	0.5880	0.8580	0.8590	0.8580
0.0925	84.0	2037	0.5986	0.8493	0.8513	0.8493
0.0621	84.9897	2061	0.5915	0.8493	0.8497	0.8493
0.059	85.9794	2085	0.5779	0.8580	0.8577	0.8580
0.0806	86.9691	2109	0.5928	0.8493	0.8501	0.8493
0.0617	88.0	2134	0.6062	0.8522	0.8520	0.8522
0.0651	88.9897	2158	0.6067	0.8522	0.8519	0.8522
0.0754	89.9794	2182	0.6108	0.8551	0.8553	0.8551
0.0682	90.9691	2206	0.6185	0.8493	0.8489	0.8493
0.0763	92.0	2231	0.6168	0.8580	0.8575	0.8580
0.0703	92.9897	2255	0.6259	0.8522	0.8521	0.8522
0.0861	93.9794	2279	0.6128	0.8551	0.8553	0.8551
0.0807	94.9691	2303	0.6140	0.8551	0.8547	0.8551
0.0621	96.0	2328	0.6133	0.8522	0.8532	0.8522
0.0831	96.9897	2352	0.6101	0.8493	0.8507	0.8493
0.0625	97.9794	2376	0.6097	0.8493	0.8507	0.8493
0.0571	98.9691	2400	0.6084	0.8493	0.8507	0.8493

Framework versions

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