detr_finetunned_air

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

• Loss: 0.8865
• Map: 0.32
• Map 50: 0.778
• Map 75: 0.1885
• Map Small: 0.3219
• Map Medium: 0.0139
• Map Large: -1.0
• Mar 1: 0.0252
• Mar 10: 0.1994
• Mar 100: 0.487
• Mar Small: 0.4896
• Mar Medium: 0.0122
• Mar Large: -1.0
• Map Falciparum Trophozoite: 0.32
• Mar 100 Falciparum Trophozoite: 0.487
• Map Wbc: -1.0
• Mar 100 Wbc: -1.0

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: 5e-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: 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 Falciparum Trophozoite	Mar 100 Falciparum Trophozoite	Map Wbc	Mar 100 Wbc
No log	1.0	209	1.2045	0.1719	0.4768	0.0896	0.1728	0.0167	-1.0	0.0188	0.139	0.4221	0.4242	0.0184	-1.0	0.1719	0.4221	-1.0	-1.0
No log	2.0	418	1.1177	0.2155	0.5922	0.1044	0.2165	0.005	-1.0	0.021	0.1601	0.4249	0.4272	0.0041	-1.0	0.2155	0.4249	-1.0	-1.0
1.339	3.0	627	1.0571	0.249	0.6576	0.1328	0.2503	0.0229	-1.0	0.0236	0.1739	0.44	0.4422	0.0224	-1.0	0.249	0.44	-1.0	-1.0
1.339	4.0	836	1.0675	0.2358	0.6473	0.1175	0.2369	0.0139	-1.0	0.0216	0.166	0.4242	0.4265	0.0122	-1.0	0.2358	0.4242	-1.0	-1.0
1.0423	5.0	1045	1.0266	0.2509	0.6759	0.1224	0.2525	0.0	-1.0	0.0228	0.1711	0.4318	0.4341	0.0	-1.0	0.2509	0.4318	-1.0	-1.0
1.0423	6.0	1254	1.0054	0.2569	0.6847	0.1269	0.2579	0.0091	-1.0	0.0226	0.1754	0.4467	0.4491	0.0122	-1.0	0.2569	0.4467	-1.0	-1.0
1.0423	7.0	1463	0.9812	0.2767	0.7163	0.1454	0.2782	0.0114	-1.0	0.0238	0.1831	0.4539	0.4562	0.0224	-1.0	0.2767	0.4539	-1.0	-1.0
0.99	8.0	1672	1.0019	0.271	0.7169	0.1358	0.2724	0.0096	-1.0	0.0236	0.1801	0.4515	0.4538	0.0143	-1.0	0.271	0.4515	-1.0	-1.0
0.99	9.0	1881	0.9623	0.2873	0.731	0.1597	0.2886	0.0064	-1.0	0.0251	0.1865	0.4608	0.4633	0.0061	-1.0	0.2873	0.4608	-1.0	-1.0
0.9521	10.0	2090	0.9763	0.273	0.711	0.1419	0.2742	0.011	-1.0	0.0229	0.178	0.4482	0.4506	0.0122	-1.0	0.273	0.4482	-1.0	-1.0
0.9521	11.0	2299	0.9551	0.2906	0.7354	0.1634	0.2925	0.0064	-1.0	0.0248	0.188	0.4654	0.4679	0.0082	-1.0	0.2906	0.4654	-1.0	-1.0
0.92	12.0	2508	0.9430	0.2956	0.7454	0.1685	0.297	0.0052	-1.0	0.0248	0.1886	0.4696	0.4721	0.0061	-1.0	0.2956	0.4696	-1.0	-1.0
0.92	13.0	2717	0.9434	0.2953	0.7445	0.1721	0.2968	0.0233	-1.0	0.0245	0.1895	0.4673	0.4697	0.0265	-1.0	0.2953	0.4673	-1.0	-1.0
0.92	14.0	2926	0.9228	0.3001	0.7498	0.1716	0.3015	0.0132	-1.0	0.0244	0.1899	0.4767	0.4792	0.0143	-1.0	0.3001	0.4767	-1.0	-1.0
0.8986	15.0	3135	0.9194	0.3036	0.7566	0.1757	0.3055	0.0119	-1.0	0.0243	0.1929	0.4778	0.4803	0.0143	-1.0	0.3036	0.4778	-1.0	-1.0
0.8986	16.0	3344	0.9166	0.3063	0.7558	0.1791	0.3081	0.0129	-1.0	0.0253	0.1928	0.4762	0.4787	0.0122	-1.0	0.3063	0.4762	-1.0	-1.0
0.875	17.0	3553	0.9218	0.3021	0.7573	0.1653	0.3041	0.0089	-1.0	0.0251	0.1908	0.4721	0.4746	0.0082	-1.0	0.3021	0.4721	-1.0	-1.0
0.875	18.0	3762	0.9094	0.3032	0.7548	0.1705	0.3052	0.0079	-1.0	0.0242	0.1921	0.4769	0.4795	0.0061	-1.0	0.3032	0.4769	-1.0	-1.0
0.875	19.0	3971	0.8965	0.3156	0.7713	0.1873	0.3171	0.0187	-1.0	0.0247	0.1963	0.484	0.4865	0.0184	-1.0	0.3156	0.484	-1.0	-1.0
0.8575	20.0	4180	0.8995	0.3101	0.7674	0.1854	0.3116	0.0069	-1.0	0.0247	0.1964	0.4803	0.4829	0.0061	-1.0	0.3101	0.4803	-1.0	-1.0
0.8575	21.0	4389	0.8992	0.3118	0.7676	0.1852	0.3131	0.0119	-1.0	0.0252	0.1954	0.4794	0.4819	0.0102	-1.0	0.3118	0.4794	-1.0	-1.0
0.834	22.0	4598	0.8912	0.3169	0.7744	0.1894	0.3186	0.0089	-1.0	0.0254	0.1977	0.4876	0.4902	0.0082	-1.0	0.3169	0.4876	-1.0	-1.0
0.834	23.0	4807	0.8922	0.3175	0.7761	0.1895	0.3195	0.0083	-1.0	0.0255	0.1984	0.4881	0.4907	0.0102	-1.0	0.3175	0.4881	-1.0	-1.0
0.8217	24.0	5016	0.8946	0.3153	0.7735	0.1809	0.3167	0.0119	-1.0	0.0249	0.1972	0.4819	0.4844	0.0102	-1.0	0.3153	0.4819	-1.0	-1.0
0.8217	25.0	5225	0.8891	0.3198	0.7801	0.1877	0.3213	0.0089	-1.0	0.025	0.1981	0.4878	0.4903	0.0082	-1.0	0.3198	0.4878	-1.0	-1.0
0.8217	26.0	5434	0.8867	0.3206	0.7794	0.1894	0.3223	0.0139	-1.0	0.0254	0.1987	0.4875	0.4901	0.0122	-1.0	0.3206	0.4875	-1.0	-1.0
0.8153	27.0	5643	0.8859	0.3207	0.7787	0.1897	0.3224	0.0139	-1.0	0.0255	0.1991	0.4879	0.4905	0.0122	-1.0	0.3207	0.4879	-1.0	-1.0
0.8153	28.0	5852	0.8862	0.3203	0.7785	0.1882	0.3222	0.0139	-1.0	0.0255	0.1994	0.4867	0.4892	0.0122	-1.0	0.3203	0.4867	-1.0	-1.0
0.8022	29.0	6061	0.8864	0.32	0.7776	0.1892	0.3219	0.0139	-1.0	0.0253	0.1994	0.4871	0.4897	0.0122	-1.0	0.32	0.4871	-1.0	-1.0
0.8022	30.0	6270	0.8865	0.32	0.778	0.1885	0.3219	0.0139	-1.0	0.0252	0.1994	0.487	0.4896	0.0122	-1.0	0.32	0.487	-1.0	-1.0

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1