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: 0.6123
• Map: 0.5265
• Map 50: 0.7826
• Map 75: 0.5744
• Map Small: 0.2152
• Map Medium: 0.4756
• Map Large: 0.6938
• Mar 1: 0.3992
• Mar 10: 0.6443
• Mar 100: 0.7154
• Mar Small: 0.4536
• Mar Medium: 0.6485
• Mar Large: 0.8102
• Map Platelets: 0.2758
• Mar 100 Platelets: 0.5431
• Map Rbc: 0.5698
• Mar 100 Rbc: 0.7495
• Map Wbc: 0.7337
• Mar 100 Wbc: 0.8537

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: 16
• 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	13	1.1546	0.0333	0.0707	0.0249	0.0	0.0271	0.1446	0.0064	0.0421	0.2061	0.0	0.2014	0.3274	0.0067	0.0069	0.0833	0.605	0.0099	0.0063
No log	2.0	26	0.9313	0.0482	0.0836	0.0527	0.0	0.0438	0.2079	0.0087	0.0602	0.2274	0.0	0.2266	0.3448	0.0	0.0	0.1446	0.6821	0.0	0.0
No log	3.0	39	0.8843	0.0595	0.0931	0.0709	0.0	0.0531	0.2013	0.0122	0.0742	0.24	0.0	0.238	0.3705	0.0	0.0	0.1784	0.72	0.0	0.0
No log	4.0	52	0.8400	0.076	0.1146	0.093	0.0139	0.0696	0.1556	0.0179	0.0936	0.2514	0.025	0.2439	0.395	0.0069	0.0097	0.2212	0.7445	0.0	0.0
No log	5.0	65	0.8139	0.0959	0.1534	0.111	0.0302	0.0836	0.2469	0.0246	0.1213	0.2634	0.1	0.252	0.3775	0.0152	0.0556	0.2725	0.7346	0.0	0.0
No log	6.0	78	0.7977	0.146	0.2567	0.1544	0.0786	0.1605	0.2663	0.0631	0.2102	0.3363	0.2786	0.3275	0.3838	0.0966	0.2625	0.3335	0.7363	0.0079	0.01
No log	7.0	91	0.7647	0.2556	0.4191	0.274	0.1006	0.1952	0.4013	0.2166	0.4411	0.5528	0.475	0.4013	0.6032	0.1285	0.4653	0.3944	0.7468	0.2439	0.4462
No log	8.0	104	0.7444	0.291	0.4664	0.3125	0.1267	0.2016	0.4213	0.2909	0.5785	0.6785	0.3143	0.4412	0.7984	0.147	0.4875	0.4502	0.7368	0.2757	0.8112
No log	9.0	117	0.7266	0.2924	0.485	0.3148	0.1268	0.2331	0.4203	0.2878	0.6042	0.7038	0.3536	0.6701	0.819	0.1681	0.4903	0.4673	0.7373	0.2416	0.8838
No log	10.0	130	0.7287	0.2959	0.4868	0.3108	0.1195	0.2577	0.4443	0.2997	0.5952	0.6973	0.5107	0.6539	0.7842	0.1684	0.5194	0.4671	0.735	0.2521	0.8375
No log	11.0	143	0.7002	0.3551	0.5616	0.3713	0.1406	0.2663	0.4856	0.335	0.6162	0.7094	0.4393	0.6747	0.8098	0.1974	0.5292	0.4936	0.7429	0.3745	0.8562
No log	12.0	156	0.7192	0.3779	0.602	0.4117	0.1314	0.2696	0.4934	0.3521	0.5962	0.6861	0.3964	0.6607	0.8063	0.1954	0.5042	0.4779	0.7254	0.4604	0.8288
No log	13.0	169	0.6769	0.3992	0.6147	0.4365	0.1841	0.2564	0.5507	0.3647	0.6143	0.7046	0.4464	0.6746	0.7946	0.2153	0.5347	0.5117	0.7353	0.4705	0.8438
No log	14.0	182	0.6575	0.4304	0.6396	0.4777	0.1562	0.3608	0.5942	0.3753	0.63	0.7125	0.45	0.7042	0.8135	0.2083	0.525	0.5272	0.7437	0.5557	0.8687
No log	15.0	195	0.6548	0.4677	0.706	0.5117	0.1505	0.3777	0.6348	0.3825	0.629	0.7109	0.4393	0.7014	0.8238	0.2287	0.5194	0.5292	0.7432	0.6453	0.87
No log	16.0	208	0.6408	0.4898	0.726	0.5319	0.1853	0.4655	0.6638	0.3887	0.6322	0.7091	0.425	0.7033	0.8108	0.2375	0.5125	0.542	0.7434	0.69	0.8712
No log	17.0	221	0.6440	0.5086	0.7475	0.5633	0.2068	0.5118	0.6656	0.3957	0.6419	0.716	0.4464	0.7125	0.8189	0.2637	0.5458	0.5399	0.7384	0.7221	0.8637
No log	18.0	234	0.6331	0.4936	0.7323	0.5389	0.1818	0.4665	0.6635	0.3808	0.6353	0.707	0.4321	0.673	0.8094	0.2339	0.525	0.5437	0.7409	0.7031	0.855
No log	19.0	247	0.6370	0.4993	0.7416	0.5503	0.1747	0.4907	0.6774	0.3913	0.634	0.7071	0.4964	0.6626	0.7922	0.2353	0.5264	0.546	0.7387	0.7166	0.8562
No log	20.0	260	0.6358	0.5138	0.7644	0.5612	0.2402	0.4921	0.6847	0.3973	0.6412	0.7106	0.4821	0.6697	0.7934	0.2691	0.5361	0.547	0.7356	0.7254	0.86
No log	21.0	273	0.6298	0.5153	0.7664	0.5838	0.252	0.475	0.6879	0.3969	0.6429	0.7162	0.4821	0.6788	0.7913	0.2728	0.5458	0.5511	0.7439	0.7221	0.8587
No log	22.0	286	0.6224	0.5235	0.7714	0.5621	0.2005	0.4708	0.6961	0.3993	0.6465	0.7172	0.5107	0.6344	0.8087	0.2725	0.5431	0.5574	0.741	0.7405	0.8675
No log	23.0	299	0.6223	0.5284	0.7818	0.5855	0.2549	0.4685	0.6927	0.3978	0.6473	0.72	0.4821	0.648	0.806	0.2869	0.5556	0.5603	0.7431	0.7379	0.8612
No log	24.0	312	0.6161	0.5177	0.7654	0.5585	0.2368	0.4677	0.6801	0.3982	0.6446	0.7191	0.4607	0.6496	0.8169	0.2787	0.55	0.5626	0.7486	0.7118	0.8587
No log	25.0	325	0.6119	0.5295	0.7813	0.5905	0.2328	0.4801	0.6872	0.3988	0.647	0.722	0.475	0.6528	0.8113	0.2968	0.5583	0.5671	0.7502	0.7245	0.8575
No log	26.0	338	0.6147	0.5268	0.7812	0.5798	0.2004	0.4825	0.6897	0.3992	0.6483	0.721	0.4893	0.6459	0.8119	0.2863	0.5514	0.5677	0.7492	0.7264	0.8625
No log	27.0	351	0.6134	0.5246	0.7813	0.57	0.2177	0.4736	0.6895	0.3994	0.6459	0.7178	0.4643	0.6457	0.8126	0.2763	0.5417	0.5681	0.7491	0.7294	0.8625
No log	28.0	364	0.6120	0.5261	0.7826	0.5809	0.2197	0.4755	0.6931	0.3991	0.6444	0.7167	0.4607	0.6483	0.811	0.2773	0.5458	0.5693	0.7491	0.7317	0.855
No log	29.0	377	0.6123	0.526	0.7821	0.5743	0.2131	0.4755	0.6935	0.3987	0.6435	0.7157	0.4536	0.6484	0.8108	0.2754	0.5431	0.5696	0.7491	0.7332	0.855
No log	30.0	390	0.6123	0.5265	0.7826	0.5744	0.2152	0.4756	0.6938	0.3992	0.6443	0.7154	0.4536	0.6485	0.8102	0.2758	0.5431	0.5698	0.7495	0.7337	0.8537

Framework versions

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