Modelpublic

OSRSEnthusiast/trainer_output

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes15downloads

Model Card

trainer_output

This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

—Loss: 0.3427
—Accuracy: 0.8776

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: 0.001
—trainbatchsize: 20
—evalbatchsize: 8
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	39	0.4124	0.7938
No log	2.0	78	0.3294	0.8454
0.4497	3.0	117	0.2932	0.8454
0.4497	4.0	156	0.2799	0.8557
0.4497	5.0	195	0.2692	0.8969
0.2764	6.0	234	0.2604	0.8969
0.2764	7.0	273	0.2583	0.9175
0.2192	8.0	312	0.2546	0.9072
0.2192	9.0	351	0.2506	0.9072
0.2192	10.0	390	0.2536	0.9072
0.1936	11.0	429	0.2530	0.8866
0.1936	12.0	468	0.2503	0.9072
0.1731	13.0	507	0.2480	0.9072
0.1731	14.0	546	0.2496	0.9072
0.1731	15.0	585	0.2498	0.9072
0.155	16.0	624	0.2498	0.9072
0.155	17.0	663	0.2495	0.9072
0.1442	18.0	702	0.2488	0.9072
0.1442	19.0	741	0.2493	0.9072
0.1442	20.0	780	0.2490	0.9072

Framework versions

—Transformers 4.51.3
—Pytorch 2.6.0+cu124
—Datasets 3.5.1
—Tokenizers 0.21.1