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: 0.5863
• Map: 0.5535
• Map 50: 0.823
• Map 75: 0.6013
• Map Small: -1.0
• Map Medium: 0.3472
• Map Large: 0.638
• Mar 1: 0.4031
• Mar 10: 0.6432
• Mar 100: 0.7115
• Mar Small: -1.0
• Mar Medium: 0.542
• Mar Large: 0.73
• Map Platelets: 0.3468
• Mar 100 Platelets: 0.5444
• Map Rbc: 0.5782
• Mar 100 Rbc: 0.75
• Map Wbc: 0.7356
• Mar 100 Wbc: 0.84

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	0.9642	0.1078	0.177	0.1243	-1.0	0.0	0.113	0.0193	0.1197	0.2165	-1.0	0.0	0.2165	0.0	0.0	0.3235	0.6496	0.0	0.0
No log	2.0	52	0.9589	0.1277	0.2441	0.1207	-1.0	0.0405	0.1148	0.0413	0.1585	0.2468	-1.0	0.1101	0.2116	0.0391	0.1056	0.3441	0.6349	0.0	0.0
No log	3.0	78	0.8392	0.2048	0.363	0.2117	-1.0	0.111	0.2378	0.1376	0.459	0.6004	-1.0	0.4609	0.5277	0.1078	0.4514	0.405	0.6685	0.1018	0.6812
No log	4.0	104	0.7755	0.3901	0.6199	0.436	-1.0	0.149	0.4846	0.3071	0.5539	0.6514	-1.0	0.4754	0.6805	0.1471	0.4792	0.4421	0.6862	0.581	0.7887
No log	5.0	130	0.7384	0.4471	0.7188	0.488	-1.0	0.1917	0.5099	0.3475	0.5664	0.6504	-1.0	0.4507	0.7194	0.1887	0.4597	0.4747	0.694	0.678	0.7975
No log	6.0	156	0.7484	0.4525	0.7284	0.4858	-1.0	0.2084	0.5322	0.3549	0.5697	0.6437	-1.0	0.4449	0.7039	0.2053	0.4528	0.4713	0.6784	0.681	0.8
No log	7.0	182	0.7283	0.4352	0.7337	0.4859	-1.0	0.1647	0.4529	0.3455	0.5486	0.6284	-1.0	0.4	0.6178	0.1618	0.3986	0.4761	0.688	0.6677	0.7987
No log	8.0	208	0.7011	0.4873	0.7735	0.527	-1.0	0.2589	0.5811	0.3658	0.6008	0.6773	-1.0	0.5058	0.7181	0.2553	0.5111	0.5051	0.7121	0.7015	0.8087
No log	9.0	234	0.6765	0.482	0.757	0.5214	-1.0	0.257	0.5759	0.3619	0.5998	0.6877	-1.0	0.5507	0.7247	0.2518	0.5556	0.5037	0.7212	0.6903	0.7862
No log	10.0	260	0.6977	0.4837	0.7833	0.5141	-1.0	0.2669	0.5117	0.373	0.5761	0.654	-1.0	0.4377	0.6952	0.2618	0.4431	0.4919	0.7077	0.6973	0.8112
No log	11.0	286	0.6463	0.5015	0.7766	0.552	-1.0	0.2568	0.5259	0.3789	0.5999	0.6817	-1.0	0.5043	0.6271	0.2527	0.4972	0.54	0.7416	0.7118	0.8062
No log	12.0	312	0.6382	0.5109	0.7939	0.5519	-1.0	0.275	0.565	0.3855	0.612	0.6902	-1.0	0.5101	0.7083	0.2719	0.5125	0.5403	0.7307	0.7206	0.8275
No log	13.0	338	0.6360	0.504	0.7943	0.5412	-1.0	0.2616	0.562	0.3762	0.6118	0.6923	-1.0	0.5188	0.7076	0.2602	0.5208	0.5406	0.7311	0.7111	0.825
No log	14.0	364	0.6422	0.5205	0.8	0.5632	-1.0	0.305	0.6082	0.3915	0.6112	0.6849	-1.0	0.5058	0.715	0.3038	0.5097	0.5381	0.7212	0.7197	0.8238
No log	15.0	390	0.7001	0.4877	0.7964	0.5097	-1.0	0.3002	0.546	0.3682	0.5841	0.6589	-1.0	0.513	0.6866	0.2976	0.5167	0.4969	0.6761	0.6685	0.7837
No log	16.0	416	0.6330	0.5173	0.7955	0.5706	-1.0	0.3087	0.572	0.3811	0.6138	0.6855	-1.0	0.5304	0.6651	0.3025	0.5278	0.5454	0.7236	0.7038	0.805
No log	17.0	442	0.6013	0.5356	0.8156	0.5856	-1.0	0.3084	0.6229	0.3936	0.6348	0.7128	-1.0	0.5565	0.7586	0.3051	0.5625	0.5594	0.7396	0.7423	0.8363
No log	18.0	468	0.6173	0.5414	0.8115	0.5947	-1.0	0.3382	0.6081	0.4042	0.6367	0.7073	-1.0	0.5435	0.736	0.3342	0.5472	0.5528	0.7334	0.7371	0.8413
No log	19.0	494	0.5997	0.5368	0.8061	0.5793	-1.0	0.3236	0.641	0.3981	0.6291	0.7038	-1.0	0.5319	0.7575	0.3217	0.5389	0.564	0.7474	0.7246	0.825
0.7657	20.0	520	0.5929	0.5384	0.8159	0.5793	-1.0	0.3278	0.6412	0.3973	0.6361	0.7103	-1.0	0.5449	0.7386	0.3258	0.5486	0.5732	0.7524	0.716	0.83
0.7657	21.0	546	0.5932	0.5403	0.8167	0.5962	-1.0	0.3445	0.6262	0.401	0.6382	0.7129	-1.0	0.5623	0.7249	0.3424	0.5639	0.571	0.7485	0.7074	0.8263
0.7657	22.0	572	0.5931	0.5422	0.82	0.5838	-1.0	0.3427	0.6289	0.3984	0.6363	0.7068	-1.0	0.5362	0.7272	0.3403	0.5389	0.5722	0.7478	0.7141	0.8338
0.7657	23.0	598	0.5896	0.5454	0.8141	0.6003	-1.0	0.3437	0.6427	0.4001	0.6333	0.7036	-1.0	0.5304	0.7365	0.3421	0.5347	0.5739	0.7486	0.7202	0.8275
0.7657	24.0	624	0.5918	0.5473	0.8161	0.5912	-1.0	0.3487	0.61	0.4027	0.6445	0.7132	-1.0	0.5551	0.7275	0.3462	0.5569	0.5733	0.7464	0.7225	0.8363
0.7657	25.0	650	0.5874	0.5496	0.8117	0.6142	-1.0	0.3503	0.6333	0.4026	0.6448	0.7145	-1.0	0.558	0.7279	0.3494	0.5597	0.5751	0.7487	0.7244	0.835
0.7657	26.0	676	0.5887	0.5453	0.8162	0.6013	-1.0	0.3375	0.6329	0.4004	0.6365	0.7085	-1.0	0.5391	0.7279	0.3376	0.5417	0.5754	0.7476	0.723	0.8363
0.7657	27.0	702	0.5881	0.5518	0.8205	0.6036	-1.0	0.3489	0.6358	0.4014	0.6438	0.7113	-1.0	0.5478	0.728	0.3483	0.55	0.5777	0.7489	0.7293	0.835
0.7657	28.0	728	0.5865	0.5531	0.8225	0.6003	-1.0	0.3494	0.6369	0.4022	0.6443	0.7122	-1.0	0.5478	0.7289	0.3487	0.55	0.5784	0.7504	0.7321	0.8363
0.7657	29.0	754	0.5864	0.5532	0.8226	0.6009	-1.0	0.3464	0.6379	0.4031	0.6431	0.7114	-1.0	0.542	0.73	0.3462	0.5444	0.5779	0.7499	0.7356	0.84
0.7657	30.0	780	0.5863	0.5535	0.823	0.6013	-1.0	0.3472	0.638	0.4031	0.6432	0.7115	-1.0	0.542	0.73	0.3468	0.5444	0.5782	0.75	0.7356	0.84

Framework versions

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