Amoros_Beaugosse_batch_64_epochs_150 Open-source Model - Performs Excellently on Specific Tasks after Fine-tuning

Amoros Beaugosse Batch 64 Epochs 150 Test Large 2025 05 31 74882 Bs64 Freeze

Developed by Amoros

This model is a fine-tuned version based on facebook/dinov2-large, which has been optimized and trained on specific tasks.

Image Classification

Transformers

Open Source License:Apache-2.0 #DINOv2 Fine-tuning #Multi-class Classification Task #150 Epochs of Training

Downloads 1,215

Release Time : 5/31/2025

Model Overview

This model is a fine-tuned version of facebook/dinov2-large, mainly used for vision tasks, and the specific usage is not clearly stated.

Model Features

Fine-tuned Based on DINOv2-large

Fine-tuning is carried out using the powerful DINOv2-large architecture, which may inherit its excellent visual feature extraction ability.

Long-term Training

After 150 epochs of training, the model may have good convergence.

Mixed Precision Training

Native AMP is used for mixed precision training, which improves the training efficiency.

Model Capabilities

Visual Feature Extraction

Image Classification (Inference)

Use Cases

Computer Vision

Image Classification

May be applicable to image classification tasks

The F1 micro score is 0.6595 and the accuracy is 0.5713 on the evaluation set

🚀 Amoros_Beaugosse_batch_64_epochs_150_test-large-2025_05_31_74882-bs64_freeze

This model is a fine - tuned version of facebook/dinov2-large on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0794
F1 Micro: 0.6595
F1 Macro: 0.5426
Accuracy: 0.5713
Learning Rate: 0.0000

🚀 Quick Start

This section provides a brief introduction to the Amoros_Beaugosse model, which is a fine - tuned version of the facebook/dinov2-large model.

✨ Features

Fine - tuned from the facebook/dinov2-large model.
Achieved certain evaluation metrics on the evaluation set, including Loss, F1 Micro, F1 Macro, Accuracy, etc.

📚 Documentation

Model description

This model is a fine - tuned version of facebook/dinov2-large on the None dataset.

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.001
train_batch_size: 64
eval_batch_size: 64
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon = 1e - 08 and optimizer_args = No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 150
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Micro	F1 Macro	Accuracy	Rate
No log	1.0	489	0.1054	0.4744	0.1919	0.3282	0.001
0.2259	2.0	978	0.0981	0.5294	0.2757	0.3986	0.001
0.1119	3.0	1467	0.0943	0.5452	0.3540	0.4125	0.001
0.1063	4.0	1956	0.0947	0.5346	0.3396	0.4013	0.001
0.104	5.0	2445	0.0943	0.5597	0.3631	0.4329	0.001
0.1031	6.0	2934	0.0946	0.5400	0.3305	0.4131	0.001
0.1029	7.0	3423	0.0934	0.5622	0.3664	0.4370	0.001
0.1022	8.0	3912	0.0955	0.5511	0.3686	0.4273	0.001
0.1021	9.0	4401	0.0934	0.5745	0.3633	0.4571	0.001
0.102	10.0	4890	0.0930	0.5688	0.3772	0.4417	0.001
0.1027	11.0	5379	0.0930	0.5624	0.3707	0.4352	0.001
0.102	12.0	5868	0.0920	0.5713	0.3767	0.4449	0.001
0.1017	13.0	6357	0.0924	0.5641	0.3580	0.4338	0.001
0.1014	14.0	6846	0.0917	0.5733	0.3675	0.4502	0.001
0.1006	15.0	7335	0.0904	0.5817	0.3965	0.4611	0.001
0.1011	16.0	7824	0.0906	0.5759	0.4032	0.4497	0.001
0.1007	17.0	8313	0.0917	0.5629	0.3868	0.4328	0.001
0.1009	18.0	8802	0.0910	0.5791	0.3982	0.4546	0.001
0.1007	19.0	9291	0.0909	0.5657	0.3833	0.4363	0.001
0.1006	20.0	9780	0.0905	0.5832	0.3929	0.4619	0.001
0.1008	21.0	10269	0.0917	0.5678	0.4099	0.4367	0.001
0.0998	22.0	10758	0.0868	0.6078	0.4424	0.4921	0.0001
0.0947	23.0	11247	0.0861	0.6140	0.4472	0.5008	0.0001
0.0937	24.0	11736	0.0853	0.6166	0.4589	0.5022	0.0001
0.0932	25.0	12225	0.0849	0.6163	0.4571	0.5025	0.0001
0.0922	26.0	12714	0.0845	0.6215	0.4645	0.5119	0.0001
0.0912	27.0	13203	0.0842	0.6259	0.4661	0.5159	0.0001
0.091	28.0	13692	0.0839	0.6245	0.4658	0.5133	0.0001
0.0905	29.0	14181	0.0839	0.6248	0.4696	0.5141	0.0001
0.0903	30.0	14670	0.0835	0.6276	0.4716	0.5202	0.0001
0.09	31.0	15159	0.0832	0.6303	0.4792	0.5222	0.0001
0.0892	32.0	15648	0.0831	0.6310	0.4858	0.5261	0.0001
0.0893	33.0	16137	0.0826	0.6338	0.4873	0.5307	0.0001
0.0884	34.0	16626	0.0826	0.6320	0.4740	0.5241	0.0001
0.0882	35.0	17115	0.0824	0.6342	0.4855	0.5302	0.0001
0.0886	36.0	17604	0.0823	0.6351	0.4845	0.5326	0.0001
0.0881	37.0	18093	0.0822	0.6340	0.4825	0.5273	0.0001
0.0882	38.0	18582	0.0823	0.6383	0.4913	0.5369	0.0001
0.0876	39.0	19071	0.0819	0.6400	0.4970	0.5369	0.0001
0.0877	40.0	19560	0.0819	0.6372	0.4887	0.5315	0.0001
0.0864	41.0	20049	0.0821	0.6317	0.4833	0.5239	0.0001
0.0867	42.0	20538	0.0814	0.6395	0.5033	0.5369	0.0001
0.0871	43.0	21027	0.0812	0.6456	0.5000	0.5460	0.0001
0.087	44.0	21516	0.0812	0.6400	0.4966	0.5371	0.0001
0.0863	45.0	22005	0.0815	0.6392	0.5049	0.5344	0.0001
0.0863	46.0	22494	0.0812	0.6419	0.5045	0.5395	0.0001
0.0859	47.0	22983	0.0809	0.6452	0.5071	0.5442	0.0001
0.0858	48.0	23472	0.0811	0.6451	0.5108	0.5449	0.0001
0.0861	49.0	23961	0.0812	0.6415	0.4906	0.5406	0.0001
0.0856	50.0	24450	0.0808	0.6449	0.5024	0.5432	0.0001
0.0857	51.0	24939	0.0807	0.6466	0.5080	0.5475	0.0001
0.0857	52.0	25428	0.0808	0.6432	0.5082	0.5414	0.0001
0.0852	53.0	25917	0.0806	0.6507	0.5132	0.5525	0.0001
0.0847	54.0	26406	0.0806	0.6436	0.5143	0.5420	0.0001
0.0849	55.0	26895	0.0809	0.6429	0.5096	0.5409	0.0001
0.0847	56.0	27384	0.0807	0.6486	0.5029	0.5485	0.0001
0.0845	57.0	27873	0.0807	0.6439	0.5007	0.5412	0.0001
0.0848	58.0	28362	0.0806	0.6497	0.4993	0.5520	0.0001
0.0843	59.0	28851	0.0804	0.6445	0.4995	0.5391	0.0001
0.0839	60.0	29340	0.0801	0.6549	0.5226	0.5597	0.0001
0.0844	61.0	29829	0.0807	0.6450	0.4941	0.5454	0.0001
0.0832	62.0	30318	0.0801	0.6470	0.5134	0.5438	0.0001
0.084	63.0	30807	0.0804	0.6494	0.5026	0.5517	0.0001
0.0834	64.0	31296	0.0802	0.6448	0.5091	0.5431	0.0001
0.0841	65.0	31785	0.0804	0.6504	0.5109	0.5513	0.0001
0.0837	66.0	32274	0.0802	0.6483	0.5137	0.5487	0.0001
0.0833	67.0	32763	0.0801	0.6517	0.5166	0.5557	0.0001
0.0836	68.0	33252	0.0798	0.6553	0.5184	0.5574	0.0001
0.0835	69.0	33741	0.0802	0.6516	0.5112	0.5549	0.0001
0.0827	70.0	34230	0.0798	0.6536	0.5232	0.5561	0.0001
0.0832	71.0	34719	0.0801	0.6510	0.5223	0.5536	0.0001
0.0831	72.0	35208	0.0799	0.6534	0.5130	0.5583	0.0001
0.0832	73.0	35697	0.0799	0.6489	0.5129	0.5487	0.0001
0.0836	74.0	36186	0.0799	0.6451	0.5035	0.5437	0.0001
0.0827	75.0	36675	0.0798	0.6520	0.5196	0.5533	0.0001
0.0827	76.0	37164	0.0797	0.6507	0.5247	0.5498	0.0001
0.0832	77.0	37653	0.0797	0.6537	0.5186	0.5574	0.0001
0.0824	78.0	38142	0.0796	0.6520	0.5284	0.5534	0.0001
0.0828	79.0	38631	0.0795	0.6536	0.5135	0.5572	0.0001
0.0824	80.0	39120	0.0797	0.6519	0.5117	0.5523	0.0001
0.0822	81.0	39609	0.0795	0.6548	0.5192	0.5586	0.0001
0.0824	82.0	40098	0.0796	0.6550	0.5164	0.5610	0.0001
0.0822	83.0	40587	0.0796	0.6556	0.5408	0.5607	0.0001
0.0818	84.0	41076	0.0792	0.6562	0.5267	0.5631	0.0001
0.0826	85.0	41565	0.0795	0.6517	0.5200	0.5559	0.0001
0.0819	86.0	42054	0.0794	0.6546	0.5127	0.5579	0.0001
0.0822	87.0	42543	0.0794	0.6566	0.5185	0.5613	0.0001
0.0818	88.0	43032	0.0794	0.6549	0.5269	0.5598	0.0001
0.0817	89.0	43521	0.0795	0.6555	0.5239	0.5585	0.0001
0.082	90.0	44010	0.0794	0.6518	0.5180	0.5536	0.0001
0.082	91.0	44499	0.0787	0.6601	0.5286	0.5699	1e - 05
0.0804	92.0	44988	0.0786	0.6590	0.5243	0.5656	1e - 05
0.0803	93.0	45477	0.0785	0.6585	0.5256	0.5645	1e - 05
0.0792	94.0	45966	0.0785	0.6592	0.5281	0.5646	1e - 05
0.0789	95.0	46455	0.0785	0.6603	0.5329	0.5696	1e - 05
0.0788	96.0	46944	0.0785	0.6602	0.5236	0.5685	1e - 05
0.0786	97.0	47433	0.0785	0.6590	0.5270	0.5653	1e - 05
0.0789	98.0	47922	0.0784	0.6629	0.5348	0.5727	1e - 05
0.0783	99.0	48411	0.0784	0.6626	0.5344	0.5726	1e - 05
0.0789	100.0	48900	0.0785	0.6607	0.5257	0.5709	1e - 05
0.0783	101.0	49389	0.0783	0.6620	0.5332	0.5723	1e - 05
0.0783	102.0	49878	0.0783	0.6644	0.5335	0.5750	1e - 05
0.0781	103.0	50367	0.0783	0.6652	0.5375	0.5796	1e - 05
0.0782	104.0	50856	0.0783	0.6644	0.5414	0.5751	1e - 05
0.0776	105.0	51345	0.0783	0.6646	0.5412	0.5776	1e - 05
0.0778	106.0	51834	0.0782	0.6670	0.5439	0.5803	1e - 05
0.0777	107.0	52323	0.0781	0.6652	0.5333	0.5771	1e - 05
0.0778	108.0	52812	0.0782	0.6628	0.5354	0.5716	1e - 05
0.078	109.0	53301	0.0781	0.6640	0.5352	0.5752	1e - 05
0.0785	110.0	53790	0.0780	0.6655	0.5345	0.5752	1e - 05
0.0772	111.0	54279	0.0781	0.6639	0.5403	0.5748	1e - 05
0.0779	112.0	54768	0.0780	0.6648	0.5373	0.5767	1e - 05
0.0774	113.0	55257	0.0781	0.6658	0.5446	0.5792	1e - 05
0.0774	114.0	55746	0.0780	0.6672	0.5445	0.5801	1e - 05
0.078	115.0	56235	0.0782	0.6671	0.5445	0.5816	1e - 05
0.0773	116.0	56724	0.0782	0.6647	0.5352	0.5756	1e - 05
0.0779	117.0	57213	0.0781	0.6641	0.5323	0.5758	1e - 05
0.0769	118.0	57702	0.0781	0.6655	0.5342	0.5774	1e - 05
0.0773	119.0	58191	0.0780	0.6655	0.5362	0.5769	0.0000
0.0771	120.0	58680	0.0780	0.6663	0.5425	0.5777	0.0000
0.0769	121.0	59169	0.0781	0.6668	0.5404	0.5806	0.0000
0.0769	122.0	59658	0.0780	0.6680	0.5436	0.5817	0.0000
0.0771	123.0	60147	0.0780	0.6667	0.5441	0.5798	0.0000
0.0773	124.0	60636	0.0780	0.6664	0.5436	0.5784	0.0000
0.0773	125.0	61125	0.0780	0.6660	0.5453	0.5777	0.0000
0.077	126.0	61614	0.0779	0.6632	0.5347	0.5726	0.0000
0.0774	127.0	62103	0.0780	0.6649	0.5324	0.5757	0.0000
0.0767	128.0	62592	0.0780	0.6662	0.5357	0.5765	0.0000
0.077	129.0	63081	0.0779	0.6664	0.5404	0.5773	0.0000
0.0773	130.0	63570	0.0781	0.6670	0.5409	0.5802	0.0000
0.0772	131.0	64059	0.0779	0.6686	0.5461	0.5828	0.0000
0.0772	132.0	64548	0.0779	0.6671	0.5430	0.5789	0.0000
0.077	133.0	65037	0.0780	0.6678	0.5418	0.5817	0.0000
0.0769	134.0	65526	0.0780	0.6670	0.5429	0.5796	0.0000
0.0766	135.0	66015	0.0779	0.6676	0.5453	0.5783	0.0000
0.0772	136.0	66504	0.0779	0.6646	0.5399	0.5750	0.0000
0.0772	137.0	66993	0.0780	0.6651	0.5299	0.5755	0.0000
0.0773	138.0	67482	0.0780	0.6664	0.5401	0.5793	0.0000
0.0771	139.0	67971	0.0780	0.6657	0.5310	0.5784	0.0000

Framework versions

Transformers 4.48.0
Pytorch 2.6.0+cu118
Datasets 3.0.2
Tokenizers 0.21.1

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご