Imported weights of google/vit-base-patch16-224 fine-tuned on Kaggle mostafaabla/garbage-classification

5d36c7e verified almost 2 years ago

359 Bytes

Technical details:

Base model: google/vit-base-patch16-224
Dataset: [Recycling Classification (12 classes)] (https://www.kaggle.com/datasets/mostafaabla/garbage-classification)
Learning rate: 0.0001
Effective training batch size: 16 (2 items per TPU core over 8 cores.)

Research supported with Cloud TPUs from Google's TPU Research Cloud (TRC)