git-large-r-coco-IDB2-V2-IDB2-VAtlasv3

This model is a fine-tuned version of ooliverz/git-large-r-coco-IDB2-V2 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2691

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 500
• num_epochs: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2629	5.0	5	0.2399
0.2271	10.0	10	0.2413
0.2241	15.0	15	0.2413
0.1901	20.0	20	0.2415
0.2116	25.0	25	0.2410
0.1891	30.0	30	0.2420
0.2193	35.0	35	0.2416
0.204	40.0	40	0.2406
0.1697	45.0	45	0.2407
0.1814	50.0	50	0.2419
0.1576	55.0	55	0.2421
0.1371	60.0	60	0.2437
0.1329	65.0	65	0.2426
0.1385	70.0	70	0.2434
0.1355	75.0	75	0.2451
0.1156	80.0	80	0.2472
0.0942	85.0	85	0.2440
0.1222	90.0	90	0.2474
0.1068	95.0	95	0.2477
0.0886	100.0	100	0.2473
0.0735	105.0	105	0.2469
0.0792	110.0	110	0.2501
0.0784	115.0	115	0.2535
0.0686	120.0	120	0.2522
0.058	125.0	125	0.2517
0.0539	130.0	130	0.2549
0.0533	135.0	135	0.2569
0.0508	140.0	140	0.2571
0.0515	145.0	145	0.2570
0.0438	150.0	150	0.2544
0.0397	155.0	155	0.2620
0.0358	160.0	160	0.2584
0.0341	165.0	165	0.2687
0.032	170.0	170	0.2632
0.0339	175.0	175	0.2651
0.0263	180.0	180	0.2674
0.0247	185.0	185	0.2672
0.0257	190.0	190	0.2698
0.0192	195.0	195	0.2687
0.0162	200.0	200	0.2691

Framework versions

• Transformers 4.46.1
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.20.2