Dogs-Breed-Image-Classification-V0

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

• Loss: 1.8210
• Accuracy: 0.7444

Model description

This model was trained using dataset from Kaggle - Standford dogs dataset

Quotes from the website: The Stanford Dogs dataset contains images of 120 breeds of dogs from around the world. This dataset has been built using images and annotation from ImageNet for the task of fine-grained image categorization. It was originally collected for fine-grain image categorization, a challenging problem as certain dog breeds have near identical features or differ in colour and age.

citation: Aditya Khosla, Nityananda Jayadevaprakash, Bangpeng Yao and Li Fei-Fei. Novel dataset for Fine-Grained Image Categorization. First Workshop on Fine-Grained Visual Categorization (FGVC), IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2011. [pdf] [poster] [BibTex]

Secondary: J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li and L. Fei-Fei, ImageNet: A Large-Scale Hierarchical Image Database. IEEE Computer Vision and Pattern Recognition (CVPR), 2009. [pdf] [BibTex]

Intended uses & limitations

This model is fined tune solely for classifiying 120 species of dogs.

Training and evaluation data

75% training data, 25% testing data.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
13.4902	1.0	515	4.7822	0.0104
4.7159	2.0	1030	4.6822	0.0323
4.6143	3.0	1545	4.5940	0.0554
4.4855	4.0	2060	4.5027	0.0935
4.36	5.0	2575	4.3961	0.1239
4.2198	6.0	3090	4.3112	0.1528
4.0882	7.0	3605	4.1669	0.1747
3.9314	8.0	4120	4.0775	0.2021
3.7863	9.0	4635	3.9487	0.2310
3.6511	10.0	5150	3.9028	0.2466
3.5168	11.0	5665	3.8635	0.2626
3.3999	12.0	6180	3.7550	0.2767
3.3037	13.0	6695	3.6973	0.2884
3.1613	14.0	7210	3.6315	0.3037
3.0754	15.0	7725	3.4839	0.3188
2.9441	16.0	8240	3.4406	0.3302
2.8579	17.0	8755	3.3528	0.3406
2.7531	18.0	9270	3.3132	0.3472
2.6477	19.0	9785	3.2736	0.3567
2.5422	20.0	10300	3.1950	0.3756
2.4629	21.0	10815	3.1174	0.4004
2.3735	22.0	11330	2.9916	0.4225
2.2436	23.0	11845	2.9205	0.4509
2.1578	24.0	12360	2.9197	0.4689
2.0671	25.0	12875	2.8196	0.4866
1.9902	26.0	13390	2.7117	0.4961
1.8737	27.0	13905	2.7129	0.5078
1.7945	28.0	14420	2.6654	0.5143
1.7092	29.0	14935	2.6273	0.5301
1.6228	30.0	15450	2.5407	0.5454
1.5744	31.0	15965	2.5412	0.5559
1.4761	32.0	16480	2.4658	0.5658
1.4084	33.0	16995	2.4247	0.5673
1.2624	34.0	17510	2.3766	0.5758
1.2066	35.0	18025	2.2879	0.5843
1.124	36.0	18540	2.2039	0.5872
1.074	37.0	19055	2.2469	0.5965
0.9937	38.0	19570	2.1575	0.6011
0.9418	39.0	20085	2.0854	0.6122
0.8812	40.0	20600	1.9991	0.6254
0.819	41.0	21115	2.0161	0.6312
0.771	42.0	21630	1.9253	0.6375
0.7128	43.0	22145	1.9412	0.6390
0.6434	44.0	22660	1.8463	0.6509
0.6138	45.0	23175	1.8163	0.6650
0.5325	46.0	23690	1.7881	0.6710
0.498	47.0	24205	1.7526	0.6744
0.4565	48.0	24720	1.7155	0.6859
0.4109	49.0	25235	1.6874	0.6946
0.3681	50.0	25750	1.7386	0.6997
0.3306	51.0	26265	1.6578	0.7104
0.2913	52.0	26780	1.6641	0.7104
0.2598	53.0	27295	1.6823	0.7162
0.2311	54.0	27810	1.6835	0.7157
0.2115	55.0	28325	1.6581	0.7206
0.1843	56.0	28840	1.6286	0.7274
0.1668	57.0	29355	1.6358	0.7225
0.1483	58.0	29870	1.6422	0.7250
0.132	59.0	30385	1.6618	0.7284
0.1164	60.0	30900	1.6894	0.7262
0.1043	61.0	31415	1.6923	0.7276
0.0937	62.0	31930	1.6627	0.7323
0.0826	63.0	32445	1.6280	0.7342
0.0743	64.0	32960	1.6204	0.7366
0.0638	65.0	33475	1.6890	0.7383
0.0603	66.0	33990	1.6967	0.7335
0.0491	67.0	34505	1.6975	0.7306
0.0459	68.0	35020	1.7242	0.7337
0.0416	69.0	35535	1.7019	0.7374
0.0382	70.0	36050	1.7098	0.7381
0.0378	71.0	36565	1.7188	0.7383
0.0326	72.0	37080	1.8212	0.7376
0.0323	73.0	37595	1.7965	0.7393
0.0299	74.0	38110	1.7934	0.7301
0.0259	75.0	38625	1.7799	0.7335
0.0276	76.0	39140	1.8456	0.7301
0.0257	77.0	39655	1.8551	0.7391
0.0234	78.0	40170	1.7780	0.7391
0.0222	79.0	40685	1.8216	0.7362
0.0195	80.0	41200	1.8333	0.7352
0.0214	81.0	41715	1.8526	0.7430
0.0207	82.0	42230	1.8581	0.7364
0.0171	83.0	42745	1.8329	0.7393
0.0175	84.0	43260	1.8841	0.7396
0.0165	85.0	43775	1.8381	0.7345
0.0152	86.0	44290	1.8192	0.7379
0.0168	87.0	44805	1.8538	0.7388
0.0158	88.0	45320	1.8390	0.7371
0.0181	89.0	45835	1.8555	0.7374
0.0142	90.0	46350	1.7987	0.7352
0.0147	91.0	46865	1.8446	0.7427
0.0142	92.0	47380	1.8210	0.7444
0.0124	93.0	47895	1.8233	0.7405
0.0128	94.0	48410	1.8517	0.7393
0.0135	95.0	48925	1.8408	0.7413
0.0122	96.0	49440	1.8153	0.7396
0.0141	97.0	49955	1.8645	0.7432
0.0121	98.0	50470	1.8526	0.7430
0.0124	99.0	50985	1.8693	0.7388
0.0113	100.0	51500	1.8051	0.7427

Framework versions

• Transformers 4.37.2
• Pytorch 2.3.0
• Datasets 2.15.0
• Tokenizers 0.15.1