Fine-tuning pretrained VILBERT on a custom dataset

facebookresearch / mmf

A modular framework for vision & language multimodal research from Facebook AI Research (FAIR)

https://mmf.sh/

Other

5.45k stars 925 forks source link

Fine-tuning pretrained VILBERT on a custom dataset #1248

Open ir0nt0ad opened 2 years ago

ir0nt0ad commented 2 years ago

❓ Questions and Help

Hi, I want to use the Hateful Memes VILBERT checkpoint from the zoo to fine-tune it on my custom dataset. I run the following command, and it appears to work:

mmf_run config=mmf/projects/my_dataset/configs/vilbert/direct.yaml model=vilbert dataset=my_dataset checkpoint.resume_zoo=vilbert.finetuned.hateful_memes.direct checkpoint.resume_pretrained=True run_type=train_val

Just wanted to make sure that it actually does what I want it to do. Thanks!

ApoorvVerma5 commented 1 year ago

Hey, can you tell me how you converted your dataset "my_dataset" in the repo, share the script too if possible.