yolo_finetuned_fruits

This model is a fine-tuned version of hustvl/yolos-tiny on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.8378
• Map: 0.5435
• Map 50: 0.7858
• Map 75: 0.6452
• Map Small: -1.0
• Map Medium: 0.4552
• Map Large: 0.5822
• Mar 1: 0.4348
• Mar 10: 0.7199
• Mar 100: 0.7611
• Mar Small: -1.0
• Mar Medium: 0.6043
• Mar Large: 0.784
• Map Banana: 0.4186
• Mar 100 Banana: 0.7125
• Map Orange: 0.5885
• Mar 100 Orange: 0.7619
• Map Apple: 0.6235
• Mar 100 Apple: 0.8088

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: 4
• 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 Banana	Mar 100 Banana	Map Orange	Mar 100 Orange	Map Apple	Mar 100 Apple
No log	1.0	60	2.0506	0.011	0.0317	0.003	-1.0	0.0013	0.0125	0.0342	0.1814	0.2782	-1.0	0.2	0.2838	0.024	0.3375	0.0	0.0	0.009	0.4971
No log	2.0	120	1.8857	0.0241	0.0798	0.0121	-1.0	0.0197	0.0262	0.0796	0.1927	0.3263	-1.0	0.21	0.3366	0.0383	0.3775	0.0244	0.281	0.0095	0.3206
No log	3.0	180	1.7609	0.046	0.1307	0.0172	-1.0	0.0486	0.0478	0.1069	0.3004	0.3731	-1.0	0.2271	0.3854	0.0642	0.4525	0.0311	0.1286	0.0426	0.5382
No log	4.0	240	1.6919	0.0293	0.0807	0.0182	-1.0	0.0947	0.0268	0.1489	0.2855	0.427	-1.0	0.34	0.4361	0.0402	0.4325	0.0236	0.231	0.0241	0.6176
No log	5.0	300	1.4915	0.061	0.1351	0.0471	-1.0	0.2056	0.0521	0.209	0.3625	0.4944	-1.0	0.35	0.509	0.0434	0.54	0.0796	0.2667	0.0601	0.6765
No log	6.0	360	1.3753	0.0974	0.1925	0.0823	-1.0	0.2268	0.0911	0.2553	0.428	0.5633	-1.0	0.4014	0.5843	0.1087	0.5575	0.0969	0.4619	0.0866	0.6706
No log	7.0	420	1.2789	0.1289	0.2472	0.1263	-1.0	0.2379	0.1328	0.258	0.4513	0.6189	-1.0	0.4714	0.6393	0.0852	0.59	0.1844	0.5286	0.1171	0.7382
No log	8.0	480	1.1585	0.1581	0.2788	0.1709	-1.0	0.3042	0.2115	0.3352	0.5374	0.6507	-1.0	0.5171	0.6691	0.1196	0.6275	0.1799	0.5952	0.1748	0.7294
1.4858	9.0	540	1.0415	0.2814	0.4693	0.3269	-1.0	0.2975	0.341	0.3436	0.6008	0.6806	-1.0	0.4714	0.7093	0.1971	0.65	0.3023	0.6595	0.3449	0.7324
1.4858	10.0	600	1.0736	0.3515	0.5773	0.3928	-1.0	0.4128	0.3713	0.3483	0.6074	0.6841	-1.0	0.57	0.7041	0.1648	0.5975	0.4373	0.7048	0.4525	0.75
1.4858	11.0	660	0.9742	0.3939	0.585	0.4417	-1.0	0.4197	0.4288	0.3898	0.6448	0.7088	-1.0	0.6357	0.7235	0.2362	0.64	0.4709	0.7333	0.4746	0.7529
1.4858	12.0	720	0.9998	0.4352	0.6626	0.5084	-1.0	0.38	0.472	0.3865	0.636	0.6966	-1.0	0.5071	0.7254	0.2423	0.6375	0.4987	0.7024	0.5646	0.75
1.4858	13.0	780	0.9837	0.4633	0.7055	0.5407	-1.0	0.4693	0.4817	0.4042	0.6539	0.7077	-1.0	0.61	0.725	0.3007	0.6375	0.4919	0.7238	0.5973	0.7618
1.4858	14.0	840	0.8749	0.4922	0.72	0.5604	-1.0	0.4485	0.5236	0.4124	0.6801	0.7352	-1.0	0.6286	0.7522	0.3219	0.6825	0.5568	0.7524	0.598	0.7706
1.4858	15.0	900	0.9257	0.5118	0.7613	0.5846	-1.0	0.4444	0.5404	0.4054	0.6786	0.7203	-1.0	0.6143	0.7387	0.3419	0.6475	0.5699	0.7429	0.6235	0.7706
1.4858	16.0	960	0.9588	0.4799	0.743	0.5839	-1.0	0.3745	0.5221	0.4059	0.6752	0.734	-1.0	0.6086	0.7536	0.3546	0.675	0.5157	0.7476	0.5693	0.7794
0.7856	17.0	1020	0.9172	0.5192	0.7543	0.6014	-1.0	0.469	0.5495	0.4091	0.6838	0.7366	-1.0	0.6171	0.7546	0.3759	0.6975	0.5607	0.7357	0.6209	0.7765
0.7856	18.0	1080	0.8988	0.5207	0.7676	0.5971	-1.0	0.453	0.5504	0.4114	0.6938	0.7446	-1.0	0.6071	0.7642	0.373	0.715	0.5763	0.7452	0.6128	0.7735
0.7856	19.0	1140	0.8570	0.535	0.7692	0.6236	-1.0	0.4664	0.5639	0.4258	0.7068	0.7665	-1.0	0.6314	0.7867	0.3878	0.72	0.5966	0.7619	0.6205	0.8176
0.7856	20.0	1200	0.8996	0.5314	0.7943	0.6089	-1.0	0.4622	0.5615	0.4188	0.692	0.7453	-1.0	0.6186	0.7644	0.3923	0.7	0.593	0.7595	0.6089	0.7765
0.7856	21.0	1260	0.8521	0.5369	0.7821	0.6274	-1.0	0.4475	0.5734	0.4237	0.7064	0.747	-1.0	0.5971	0.7688	0.4078	0.7075	0.5978	0.7571	0.6051	0.7765
0.7856	22.0	1320	0.8472	0.5558	0.7931	0.6288	-1.0	0.4904	0.5925	0.4378	0.7159	0.7543	-1.0	0.6214	0.7747	0.4059	0.7	0.6095	0.7571	0.652	0.8059
0.7856	23.0	1380	0.8626	0.5332	0.787	0.6246	-1.0	0.4637	0.5691	0.4162	0.7127	0.7553	-1.0	0.5843	0.78	0.3987	0.705	0.5794	0.7667	0.6216	0.7941
0.7856	24.0	1440	0.8798	0.5436	0.7823	0.6218	-1.0	0.4699	0.5807	0.4293	0.7145	0.7498	-1.0	0.5843	0.7739	0.4124	0.6975	0.5917	0.7548	0.6267	0.7971
0.5635	25.0	1500	0.8468	0.5534	0.7888	0.6432	-1.0	0.4615	0.5935	0.4352	0.7166	0.7512	-1.0	0.5671	0.7777	0.4166	0.6975	0.6095	0.7619	0.6341	0.7941
0.5635	26.0	1560	0.8479	0.5436	0.7894	0.6388	-1.0	0.4535	0.583	0.4343	0.7186	0.7576	-1.0	0.5843	0.7827	0.4115	0.705	0.5874	0.7619	0.6318	0.8059
0.5635	27.0	1620	0.8385	0.5429	0.7871	0.6418	-1.0	0.4477	0.5818	0.4365	0.718	0.7608	-1.0	0.6043	0.7834	0.4197	0.7175	0.5857	0.7619	0.6234	0.8029
0.5635	28.0	1680	0.8397	0.5422	0.7857	0.6421	-1.0	0.4552	0.5805	0.4349	0.7174	0.7594	-1.0	0.6043	0.7822	0.4149	0.71	0.5865	0.7595	0.6251	0.8088
0.5635	29.0	1740	0.8379	0.5431	0.7857	0.6451	-1.0	0.4552	0.5816	0.4348	0.7191	0.7611	-1.0	0.6043	0.784	0.4187	0.7125	0.5871	0.7619	0.6235	0.8088
0.5635	30.0	1800	0.8378	0.5435	0.7858	0.6452	-1.0	0.4552	0.5822	0.4348	0.7199	0.7611	-1.0	0.6043	0.784	0.4186	0.7125	0.5885	0.7619	0.6235	0.8088

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1