dinov2-base-finetuned-har

This model is a fine-tuned version of facebook/dinov2-base on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.4424
Accuracy: 0.8968

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: 32
eval_batch_size: 32
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 128
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9155	0.9910	83	0.6204	0.8053
0.749	1.9940	167	0.4433	0.8667
0.8197	2.9970	251	0.4826	0.8571
0.6854	4.0	335	0.4243	0.8725
0.7058	4.9910	418	0.4349	0.8593
0.6717	5.9940	502	0.4984	0.8434
0.6544	6.9970	586	0.4730	0.8545
0.5846	8.0	670	0.4631	0.8630
0.5207	8.9910	753	0.4072	0.8751
0.4977	9.9940	837	0.4627	0.8608
0.4974	10.9970	921	0.4600	0.8661
0.4502	12.0	1005	0.4548	0.8725
0.4051	12.9910	1088	0.4404	0.8709
0.3862	13.9940	1172	0.4498	0.8772
0.351	14.9970	1256	0.4859	0.8677
0.3807	16.0	1340	0.5189	0.8556
0.3538	16.9910	1423	0.4959	0.8646
0.3181	17.9940	1507	0.4831	0.8698
0.3225	18.9970	1591	0.4890	0.8804
0.3257	20.0	1675	0.4817	0.8735
0.2667	20.9910	1758	0.5199	0.8683
0.2863	21.9940	1842	0.4835	0.8683
0.2527	22.9970	1926	0.4764	0.8772
0.2657	24.0	2010	0.4651	0.8767
0.1995	24.9910	2093	0.5079	0.8693
0.2481	25.9940	2177	0.5112	0.8698
0.2072	26.9970	2261	0.5082	0.8831
0.2164	28.0	2345	0.5002	0.8730
0.2198	28.9910	2428	0.4785	0.8778
0.2137	29.9940	2512	0.5012	0.8889
0.1936	30.9970	2596	0.4961	0.8757
0.2255	32.0	2680	0.4987	0.8788
0.1818	32.9910	2763	0.4840	0.8852
0.1644	33.9940	2847	0.4694	0.8862
0.1799	34.9970	2931	0.4599	0.8915
0.1624	36.0	3015	0.5122	0.8852
0.157	36.9910	3098	0.4546	0.8899
0.2165	37.9940	3182	0.5097	0.8836
0.1565	38.9970	3266	0.4566	0.8952
0.1476	40.0	3350	0.4579	0.8915
0.1296	40.9910	3433	0.4595	0.8931
0.1159	41.9940	3517	0.4841	0.8884
0.1071	42.9970	3601	0.4730	0.8820
0.1017	44.0	3685	0.4470	0.8931
0.11	44.9910	3768	0.4557	0.8910
0.126	45.9940	3852	0.4585	0.8926
0.1079	46.9970	3936	0.4551	0.8905
0.1194	48.0	4020	0.4401	0.8947
0.11	48.9910	4103	0.4424	0.8968
0.1104	49.5522	4150	0.4414	0.8958

Framework versions

Transformers 4.44.2
Pytorch 2.4.1+cu121
Datasets 2.21.0
Tokenizers 0.19.1

Loading…

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

gradient_accumulation_steps: 4

total_train_batch_size: 128

optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08

lr_scheduler_type: linear

lr_scheduler_warmup_ratio: 0.1

num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9155	0.9910	83	0.6204	0.8053
0.749	1.9940	167	0.4433	0.8667
0.8197	2.9970	251	0.4826	0.8571
0.6854	4.0	335	0.4243	0.8725
0.7058	4.9910	418	0.4349	0.8593
0.6717	5.9940	502	0.4984	0.8434
0.6544	6.9970	586	0.4730	0.8545
0.5846	8.0	670	0.4631	0.8630
0.5207	8.9910	753	0.4072	0.8751
0.4977	9.9940	837	0.4627	0.8608
0.4974	10.9970	921	0.4600	0.8661
0.4502	12.0	1005	0.4548	0.8725
0.4051	12.9910	1088	0.4404	0.8709
0.3862	13.9940	1172	0.4498	0.8772
0.351	14.9970	1256	0.4859	0.8677
0.3807	16.0	1340	0.5189	0.8556
0.3538	16.9910	1423	0.4959	0.8646
0.3181	17.9940	1507	0.4831	0.8698
0.3225	18.9970	1591	0.4890	0.8804
0.3257	20.0	1675	0.4817	0.8735
0.2667	20.9910	1758	0.5199	0.8683
0.2863	21.9940	1842	0.4835	0.8683
0.2527	22.9970	1926	0.4764	0.8772
0.2657	24.0	2010	0.4651	0.8767
0.1995	24.9910	2093	0.5079	0.8693
0.2481	25.9940	2177	0.5112	0.8698
0.2072	26.9970	2261	0.5082	0.8831
0.2164	28.0	2345	0.5002	0.8730
0.2198	28.9910	2428	0.4785	0.8778
0.2137	29.9940	2512	0.5012	0.8889
0.1936	30.9970	2596	0.4961	0.8757
0.2255	32.0	2680	0.4987	0.8788
0.1818	32.9910	2763	0.4840	0.8852
0.1644	33.9940	2847	0.4694	0.8862
0.1799	34.9970	2931	0.4599	0.8915
0.1624	36.0	3015	0.5122	0.8852
0.157	36.9910	3098	0.4546	0.8899
0.2165	37.9940	3182	0.5097	0.8836
0.1565	38.9970	3266	0.4566	0.8952
0.1476	40.0	3350	0.4579	0.8915
0.1296	40.9910	3433	0.4595	0.8931
0.1159	41.9940	3517	0.4841	0.8884
0.1071	42.9970	3601	0.4730	0.8820
0.1017	44.0	3685	0.4470	0.8931
0.11	44.9910	3768	0.4557	0.8910
0.126	45.9940	3852	0.4585	0.8926
0.1079	46.9970	3936	0.4551	0.8905
0.1194	48.0	4020	0.4401	0.8947
0.11	48.9910	4103	0.4424	0.8968
0.1104	49.5522	4150	0.4414	0.8958

Framework versions

Transformers 4.44.2

Pytorch 2.4.1+cu121

Datasets 2.21.0

Tokenizers 0.19.1

amauriciogonzalez/dinov2-base-finetuned-har

dinov2-base-finetuned-har

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

dinov2-base-finetuned-har

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions