20240530

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: 0.7211

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: 1e-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
1.8627	18.35	4000	1.5604
1.4599	36.7	8000	1.1805
1.2256	55.05	12000	0.9678
1.1121	73.39	16000	0.8867
1.0312	91.74	20000	0.8539
1.016	110.09	24000	0.8169
0.9564	128.44	28000	0.8027
0.9438	146.79	32000	0.7773
0.9099	165.14	36000	0.7705
0.8781	183.49	40000	0.7570
0.8743	201.83	44000	0.7558
0.8581	220.18	48000	0.7424
0.8447	238.53	52000	0.7356
0.8207	256.88	56000	0.7324
0.8018	275.23	60000	0.7266
0.793	293.58	64000	0.7279
0.7987	311.93	68000	0.7250
0.7643	330.28	72000	0.7245
0.7673	348.62	76000	0.7297
0.7509	366.97	80000	0.7169
0.758	385.32	84000	0.7202
0.7355	403.67	88000	0.7180
0.738	422.02	92000	0.7202
0.7296	440.37	96000	0.7229
0.7107	458.72	100000	0.7164
0.6961	477.06	104000	0.7161
0.7096	495.41	108000	0.7156
0.6837	513.76	112000	0.7145
0.7034	532.11	116000	0.7147
0.6868	550.46	120000	0.7201
0.6814	568.81	124000	0.7164
0.6896	587.16	128000	0.7167
0.6809	605.5	132000	0.7149
0.6583	623.85	136000	0.7196
0.6696	642.2	140000	0.7185
0.6704	660.55	144000	0.7156
0.6761	678.9	148000	0.7235
0.6577	697.25	152000	0.7207
0.6649	715.6	156000	0.7211
0.6589	733.94	160000	0.7203
0.6461	752.29	164000	0.7190
0.6406	770.64	168000	0.7213
0.638	788.99	172000	0.7191
0.6523	807.34	176000	0.7232
0.6336	825.69	180000	0.7177
0.6382	844.04	184000	0.7199
0.6394	862.39	188000	0.7241
0.6406	880.73	192000	0.7239
0.6366	899.08	196000	0.7226
0.65	917.43	200000	0.7198
0.6382	935.78	204000	0.7198
0.6257	954.13	208000	0.7241
0.6242	972.48	212000	0.7211
0.6405	990.83	216000	0.7211

Framework versions

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