How to enable the feature_extractor in the cvt pre-training model to read the datasets_load image of hugging face?
#1
by
xcy2336
- opened
I want to fine-tune a pretrained model cvt. Downloaded the cvt from hugging face, took the required photos in jpg format, and organized the file directory for storage according to the instructions in the datasets in the docs on hugging face. Use load_datasets to load generated datasets from a file directory. In the cvt pre-training model, after using Image.open to obtain image data, feature_extractor must be performed in advance before feeding it to the pre-training model. My question is how to open the image content in datasets?
I've known how to get the image data.
xcy2336
changed discussion status to
closed