cifar10_outputs
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the cifar10 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0806
- Accuracy: 0.9914
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: 0.0001
- train_batch_size: 17
- eval_batch_size: 17
- seed: 1337
- distributed_type: IPU
- gradient_accumulation_steps: 128
- total_train_batch_size: 8704
- total_eval_batch_size: 272
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.25
- num_epochs: 100.0
- training precision: Mixed Precision
Training results
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cpu
- Datasets 2.3.3.dev0
- Tokenizers 0.12.1