conditional-detr-resnet-50_adamw_hf_finetuned_food-roboflow

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

• Loss: 10.2661

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4448.7409	0.77	50	2784.9878
2312.7831	1.54	100	1275.4084
1139.5262	2.31	150	825.0410
857.6314	3.08	200	626.2924
630.5604	3.85	250	492.9028
518.8461	4.62	300	391.5704
418.3785	5.38	350	312.8008
331.9224	6.15	400	250.7763
261.2112	6.92	450	201.8681
207.421	7.69	500	163.3361
172.0157	8.46	550	132.3763
140.2059	9.23	600	108.1960
110.3696	10.0	650	89.2497
95.9428	10.77	700	73.2426
78.812	11.54	750	61.4362
64.1493	12.31	800	51.2599
56.184	13.08	850	43.4741
46.7644	13.85	900	37.4028
38.1726	14.62	950	32.6764
34.7277	15.38	1000	28.1207
29.9978	16.15	1050	25.0045
27.5957	16.92	1100	22.3012
23.6549	17.69	1150	19.9766
21.4961	18.46	1200	18.2427
19.3312	19.23	1250	16.8829
17.8215	20.0	1300	15.5794
16.2877	20.77	1350	14.5258
16.0017	21.54	1400	13.5888
14.6977	22.31	1450	13.0312
13.9111	23.08	1500	12.4389
13.5826	23.85	1550	11.8718
12.5621	24.62	1600	11.6370
12.1993	25.38	1650	11.1523
12.2491	26.15	1700	10.8586
11.5879	26.92	1750	10.6795
11.4736	27.69	1800	10.5273
11.5405	28.46	1850	10.4190
11.5894	29.23	1900	10.3225
11.0308	30.0	1950	10.2661

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1