convnextv2-nano-1k-224-finetuned-galaxy10-decals

This model is a fine-tuned version of facebook/convnextv2-nano-1k-224 on the matthieulel/galaxy10_decals dataset. It achieves the following results on the evaluation set:

• Loss: 0.7023
• Accuracy: 0.8591
• Precision: 0.8565
• Recall: 0.8591
• F1: 0.8566

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: 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
1.0509	0.99	62	0.9876	0.6561	0.6692	0.6561	0.6355
0.8315	2.0	125	0.6745	0.7638	0.7726	0.7638	0.7581
0.7659	2.99	187	0.6651	0.7897	0.8092	0.7897	0.7858
0.6645	4.0	250	0.6068	0.8027	0.8117	0.8027	0.8031
0.6212	4.99	312	0.5895	0.8061	0.8158	0.8061	0.8058
0.624	6.0	375	0.5128	0.8269	0.8223	0.8269	0.8220
0.5648	6.99	437	0.5219	0.8196	0.8239	0.8196	0.8171
0.5235	8.0	500	0.5652	0.8275	0.8303	0.8275	0.8270
0.5316	8.99	562	0.4804	0.8427	0.8431	0.8427	0.8396
0.4756	10.0	625	0.5345	0.8213	0.8188	0.8213	0.8169
0.4758	10.99	687	0.5560	0.8326	0.8324	0.8326	0.8302
0.4499	12.0	750	0.5283	0.8416	0.8458	0.8416	0.8423
0.4274	12.99	812	0.5347	0.8348	0.8364	0.8348	0.8337
0.4148	14.0	875	0.5326	0.8422	0.8396	0.8422	0.8383
0.3822	14.99	937	0.5116	0.8410	0.8442	0.8410	0.8412
0.3613	16.0	1000	0.6081	0.8230	0.8245	0.8230	0.8210
0.2903	16.99	1062	0.6212	0.8298	0.8289	0.8298	0.8275
0.3405	18.0	1125	0.6093	0.8377	0.8394	0.8377	0.8368
0.2999	18.99	1187	0.6482	0.8393	0.8352	0.8393	0.8356
0.2792	20.0	1250	0.6473	0.8484	0.8482	0.8484	0.8419
0.2681	20.99	1312	0.6710	0.8467	0.8428	0.8467	0.8425
0.2966	22.0	1375	0.6355	0.8534	0.8513	0.8534	0.8514
0.2609	22.99	1437	0.6850	0.8399	0.8406	0.8399	0.8397
0.2281	24.0	1500	0.7124	0.8444	0.8444	0.8444	0.8440
0.2354	24.99	1562	0.7317	0.8427	0.8394	0.8427	0.8395
0.2188	26.0	1625	0.6753	0.8512	0.8490	0.8512	0.8489
0.2118	26.99	1687	0.6865	0.8506	0.8495	0.8506	0.8494
0.2232	28.0	1750	0.7098	0.8557	0.8531	0.8557	0.8535
0.2104	28.99	1812	0.7023	0.8591	0.8565	0.8591	0.8566
0.1936	29.76	1860	0.7043	0.8557	0.8540	0.8557	0.8541

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.15.1