Giters

google-research / scenic

Scenic: A Jax Library for Computer Vision Research and Beyond

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Full training example with custom dataset

unrue opened this issue · comments

commented

Do you have a owl-vit full training example with custom dataset from scratch? I don't understand what to do from the manual:


`python -m scenic.projects.owl_vit.main \
  --alsologtostderr=true \
  --workdir=/tmp/training \
  --config=scenic/projects/owl_vit/configs/clip_b32_finetune.py`

I have a dataset with 30k images. I need description of each images instead a simple label? I need to make some kind of conversion? I need to train a CLIP model before? Someone could explain all step to make a training on custom dataset? Thanks.

commented

@unrue I have the same confusion. Do you know the process of building a custom dataset and fine-tuning it? My task is the fine-tuning of object detection. Do I need to provide the target text and target coordinates for each image? Thanks!