detr_finetuned_bccd

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

• Loss: 0.5911
• Map: 0.5587
• Map 50: 0.8202
• Map 75: 0.6128
• Map Small: 0.2752
• Map Medium: 0.5141
• Map Large: 0.7114
• Mar 1: 0.4061
• Mar 10: 0.644
• Mar 100: 0.7208
• Mar Small: 0.4679
• Mar Medium: 0.6882
• Mar Large: 0.8062
• Map Platelets: 0.3331
• Mar 100 Platelets: 0.5556
• Map Rbc: 0.5785
• Mar 100 Rbc: 0.7543
• Map Wbc: 0.7646
• Mar 100 Wbc: 0.8525

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: 0.0001
• train_batch_size: 8
• 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: cosine
• num_epochs: 30

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 Platelets	Mar 100 Platelets	Map Rbc	Mar 100 Rbc	Map Wbc	Mar 100 Wbc
No log	1.0	26	1.0745	0.0791	0.1484	0.076	0.0	0.0782	0.1361	0.014	0.0973	0.2035	0.0	0.1939	0.3557	0.0	0.0	0.2372	0.6104	0.0	0.0
No log	2.0	52	0.9278	0.1125	0.1831	0.1244	0.0	0.1171	0.1836	0.0194	0.1209	0.2319	0.0	0.2248	0.3856	0.0	0.0	0.3376	0.6958	0.0	0.0
No log	3.0	78	0.9425	0.1198	0.2245	0.1231	0.0039	0.1425	0.1743	0.032	0.1437	0.2465	0.0321	0.2464	0.3827	0.0199	0.0778	0.3394	0.6618	0.0	0.0
No log	4.0	104	0.8965	0.1344	0.245	0.1438	0.0109	0.1677	0.2048	0.0305	0.1928	0.2909	0.2179	0.2829	0.3606	0.0214	0.2028	0.3818	0.67	0.0	0.0
No log	5.0	130	0.8165	0.1802	0.3282	0.1851	0.0431	0.2059	0.2791	0.0791	0.2832	0.3782	0.3357	0.3734	0.3918	0.0778	0.3778	0.4536	0.7232	0.0093	0.0338
No log	6.0	156	0.8918	0.1743	0.3513	0.15	0.0796	0.1922	0.2253	0.0987	0.3532	0.4564	0.3286	0.3935	0.4931	0.105	0.45	0.4008	0.6716	0.0173	0.2475
No log	7.0	182	0.7744	0.2888	0.4691	0.3181	0.0538	0.2335	0.4002	0.2864	0.554	0.6618	0.4714	0.4136	0.7749	0.103	0.5222	0.451	0.7081	0.3123	0.755
No log	8.0	208	0.7615	0.3452	0.5449	0.3663	0.1073	0.2042	0.4669	0.3263	0.5709	0.6729	0.3286	0.6103	0.8227	0.1204	0.4625	0.4565	0.7075	0.4586	0.8487
No log	9.0	234	0.7013	0.4743	0.7399	0.5305	0.1527	0.4342	0.6532	0.3489	0.6016	0.6871	0.3821	0.6618	0.7912	0.2204	0.4917	0.506	0.7271	0.6966	0.8425
No log	10.0	260	0.6729	0.4978	0.7687	0.5505	0.1813	0.464	0.6733	0.3692	0.6146	0.7015	0.4214	0.68	0.774	0.2453	0.5292	0.5233	0.734	0.725	0.8413
No log	11.0	286	0.6773	0.4691	0.7251	0.5163	0.1795	0.2625	0.6377	0.3542	0.6153	0.6964	0.4714	0.6676	0.7592	0.2314	0.5347	0.5234	0.717	0.6525	0.8375
No log	12.0	312	0.6581	0.5059	0.7516	0.5704	0.1709	0.4819	0.6766	0.3816	0.6253	0.7074	0.4643	0.6775	0.7854	0.251	0.5458	0.5416	0.7326	0.725	0.8438
No log	13.0	338	0.6483	0.514	0.7879	0.5714	0.1836	0.4694	0.688	0.3821	0.6225	0.7026	0.4964	0.6524	0.7954	0.2628	0.5125	0.5389	0.7327	0.7404	0.8625
No log	14.0	364	0.6440	0.5223	0.7986	0.5865	0.2324	0.4881	0.6661	0.387	0.626	0.7069	0.5429	0.6719	0.7583	0.3157	0.5667	0.5396	0.7229	0.7114	0.8313
No log	15.0	390	0.6845	0.5126	0.7979	0.5487	0.2364	0.4547	0.6495	0.3829	0.6194	0.6939	0.5643	0.6301	0.7302	0.3246	0.5722	0.5149	0.6984	0.6984	0.8112
No log	16.0	416	0.6464	0.5213	0.8164	0.5725	0.2683	0.4662	0.6724	0.3855	0.6211	0.7027	0.4643	0.6797	0.7676	0.3011	0.5486	0.5368	0.7295	0.726	0.83
No log	17.0	442	0.6218	0.5267	0.8101	0.5803	0.2126	0.5258	0.6783	0.3862	0.6234	0.7034	0.4964	0.6768	0.7769	0.3032	0.55	0.5597	0.7441	0.7173	0.8163
No log	18.0	468	0.6168	0.5384	0.8041	0.5897	0.2356	0.5064	0.6929	0.4012	0.6352	0.7101	0.5143	0.6716	0.7907	0.3075	0.5472	0.558	0.7456	0.7496	0.8375
No log	19.0	494	0.6173	0.5408	0.8057	0.5921	0.2388	0.5413	0.6875	0.3997	0.635	0.709	0.4536	0.7108	0.8094	0.3092	0.5444	0.561	0.7402	0.7522	0.8425
0.9152	20.0	520	0.6054	0.541	0.8049	0.5917	0.2594	0.505	0.7075	0.3964	0.6354	0.7166	0.4821	0.6822	0.8011	0.2986	0.5569	0.5609	0.7442	0.7636	0.8487
0.9152	21.0	546	0.5996	0.547	0.8152	0.5945	0.2915	0.5	0.7014	0.4018	0.6383	0.7138	0.4857	0.6833	0.7896	0.3255	0.5569	0.5676	0.7456	0.7479	0.8388
0.9152	22.0	572	0.6045	0.5503	0.828	0.5925	0.2747	0.5212	0.7048	0.3997	0.6391	0.7141	0.4429	0.6881	0.803	0.3225	0.5528	0.5714	0.7457	0.757	0.8438
0.9152	23.0	598	0.6001	0.5523	0.8268	0.6055	0.2581	0.5133	0.7072	0.4009	0.6432	0.7149	0.4893	0.6839	0.7892	0.3332	0.5583	0.5729	0.7465	0.7508	0.84
0.9152	24.0	624	0.6008	0.5545	0.8279	0.6021	0.2653	0.5119	0.7072	0.4051	0.6453	0.7221	0.4786	0.6921	0.7969	0.3337	0.5681	0.5727	0.7496	0.757	0.8487
0.9152	25.0	650	0.5943	0.5568	0.8286	0.6102	0.2926	0.511	0.7093	0.4082	0.6479	0.7241	0.4857	0.6915	0.8	0.3384	0.5694	0.5768	0.7503	0.7551	0.8525
0.9152	26.0	676	0.5919	0.5562	0.8211	0.6161	0.273	0.5281	0.7125	0.4062	0.6453	0.7213	0.4786	0.6849	0.8065	0.3267	0.5556	0.5778	0.7509	0.764	0.8575
0.9152	27.0	702	0.5939	0.5555	0.8187	0.6006	0.275	0.5102	0.7088	0.4047	0.643	0.7197	0.4786	0.6843	0.8053	0.3282	0.5556	0.5745	0.7498	0.7636	0.8537
0.9152	28.0	728	0.5925	0.5556	0.8192	0.6068	0.2763	0.5107	0.71	0.4044	0.6414	0.7166	0.4714	0.6816	0.8035	0.329	0.5472	0.5766	0.7501	0.7613	0.8525
0.9152	29.0	754	0.5916	0.5597	0.8205	0.612	0.2759	0.514	0.713	0.4065	0.6442	0.7214	0.4679	0.6884	0.8068	0.3332	0.5556	0.5788	0.7548	0.767	0.8537
0.9152	30.0	780	0.5911	0.5587	0.8202	0.6128	0.2752	0.5141	0.7114	0.4061	0.644	0.7208	0.4679	0.6882	0.8062	0.3331	0.5556	0.5785	0.7543	0.7646	0.8525

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0