detr_resnet_50_finetuned_grape

This model is a fine-tuned version of facebook/detr-resnet-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9501
• Map: 0.4128
• Map 50: 0.7569
• Map 75: 0.4085
• Map Small: -1.0
• Map Medium: 0.2598
• Map Large: 0.4291
• Mar 1: 0.05
• Mar 10: 0.378
• Mar 100: 0.5668
• Mar Small: -1.0
• Mar Medium: 0.396
• Mar Large: 0.5842
• Map Grape: 0.4128
• Mar 100 Grape: 0.5668

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: 2
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Grape	Mar 100 Grape
No log	1.0	121	1.4508	0.1575	0.3359	0.1351	-1.0	0.0959	0.1663	0.0373	0.166	0.4958	-1.0	0.228	0.523	0.1575	0.4958
No log	2.0	242	1.3198	0.173	0.3472	0.1551	-1.0	0.1227	0.1817	0.038	0.1948	0.557	-1.0	0.3067	0.5824	0.173	0.557
No log	3.0	363	1.2969	0.1867	0.3509	0.1808	-1.0	0.1446	0.1942	0.0432	0.1999	0.5699	-1.0	0.3467	0.5926	0.1867	0.5699
No log	4.0	484	1.2968	0.255	0.5278	0.2251	-1.0	0.1824	0.2685	0.0423	0.2721	0.5314	-1.0	0.2427	0.5608	0.255	0.5314
1.4189	5.0	605	1.1961	0.3081	0.6141	0.2781	-1.0	0.2025	0.3219	0.0442	0.3025	0.5466	-1.0	0.3093	0.5706	0.3081	0.5466
1.4189	6.0	726	1.1386	0.3408	0.6426	0.3311	-1.0	0.2289	0.3547	0.0473	0.3242	0.5733	-1.0	0.4	0.5909	0.3408	0.5733
1.4189	7.0	847	1.1412	0.3338	0.6733	0.3062	-1.0	0.2106	0.3494	0.0467	0.3302	0.5369	-1.0	0.3173	0.5591	0.3338	0.5369
1.4189	8.0	968	1.1755	0.3302	0.6427	0.3021	-1.0	0.1893	0.349	0.0469	0.3201	0.5123	-1.0	0.2733	0.5367	0.3302	0.5123
1.2457	9.0	1089	1.0411	0.3704	0.7163	0.3533	-1.0	0.2249	0.3861	0.0461	0.3442	0.5661	-1.0	0.3853	0.5844	0.3704	0.5661
1.2457	10.0	1210	1.0536	0.371	0.7168	0.3467	-1.0	0.2475	0.3851	0.0448	0.3565	0.5408	-1.0	0.368	0.5583	0.371	0.5408
1.2457	11.0	1331	1.0660	0.3744	0.7236	0.3503	-1.0	0.2152	0.3915	0.0455	0.3604	0.5396	-1.0	0.3347	0.5604	0.3744	0.5396
1.2457	12.0	1452	1.0016	0.3852	0.7426	0.3582	-1.0	0.2276	0.4024	0.0453	0.364	0.548	-1.0	0.3507	0.5681	0.3852	0.548
1.0907	13.0	1573	0.9670	0.3984	0.7545	0.3857	-1.0	0.2424	0.4153	0.0478	0.3654	0.5615	-1.0	0.372	0.5808	0.3984	0.5615
1.0907	14.0	1694	0.9716	0.4129	0.7626	0.4058	-1.0	0.2418	0.4303	0.0484	0.3787	0.5693	-1.0	0.372	0.5893	0.4129	0.5693
1.0907	15.0	1815	0.9822	0.406	0.7479	0.4057	-1.0	0.2301	0.4244	0.0482	0.3781	0.5579	-1.0	0.356	0.5783	0.406	0.5579
1.0907	16.0	1936	0.9546	0.4092	0.7627	0.3995	-1.0	0.2388	0.4267	0.0469	0.3724	0.5737	-1.0	0.3853	0.5928	0.4092	0.5737
1.0056	17.0	2057	0.9427	0.4207	0.7715	0.4092	-1.0	0.2467	0.4385	0.0483	0.3797	0.5776	-1.0	0.3973	0.5959	0.4207	0.5776
1.0056	18.0	2178	0.9442	0.4177	0.7642	0.4243	-1.0	0.2529	0.4349	0.0488	0.3805	0.5679	-1.0	0.384	0.5866	0.4177	0.5679
1.0056	19.0	2299	0.9367	0.4168	0.7689	0.408	-1.0	0.2604	0.4333	0.0502	0.3757	0.5698	-1.0	0.396	0.5874	0.4168	0.5698
1.0056	20.0	2420	0.9501	0.4128	0.7569	0.4085	-1.0	0.2598	0.4291	0.05	0.378	0.5668	-1.0	0.396	0.5842	0.4128	0.5668

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2