20240704

This model is a fine-tuned version of facebook/detr-resnet-50 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2392

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: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.4262	32.0	800	0.3110
0.3444	64.0	1600	0.2795
0.3122	96.0	2400	0.2704
0.2893	128.0	3200	0.2639
0.2823	160.0	4000	0.2518
0.2679	192.0	4800	0.2502
0.2491	224.0	5600	0.2444
0.2354	256.0	6400	0.2495
0.2284	288.0	7200	0.2406
0.2318	320.0	8000	0.2453
0.217	352.0	8800	0.2432
0.2124	384.0	9600	0.2469
0.2097	416.0	10400	0.2516
0.1998	448.0	11200	0.2428
0.1944	480.0	12000	0.2445
0.1909	512.0	12800	0.2468
0.1952	544.0	13600	0.2401
0.1903	576.0	14400	0.2408
0.1867	608.0	15200	0.2478
0.1845	640.0	16000	0.2443
0.179	672.0	16800	0.2432
0.1744	704.0	17600	0.2410
0.1747	736.0	18400	0.2432
0.1674	768.0	19200	0.2386
0.166	800.0	20000	0.2384
0.1686	832.0	20800	0.2399
0.1647	864.0	21600	0.2391
0.1599	896.0	22400	0.2398
0.1633	928.0	23200	0.2401
0.1654	960.0	24000	0.2397
0.1632	992.0	24800	0.2392

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0
• Datasets 2.12.0
• Tokenizers 0.15.1