You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thank for sharing this work. Is also possible to share the checkpoint_model (from HF or anywhere else)? I'm aware of the datasets and algortithm are provided herein.
However, I'm constraint on a smaller GPU (RTX 3060) and even with batch size reduction full training may takes a few weeks I suppose.
ckpt_base_path: ./ # put absolute path here. Checkpoints will be loaded from ${ckpt_base_path}/outputs
The text was updated successfully, but these errors were encountered:
Hello,
Thank for sharing this work. Is also possible to share the checkpoint_model (from HF or anywhere else)? I'm aware of the datasets and algortithm are provided herein.
However, I'm constraint on a smaller GPU (RTX 3060) and even with batch size reduction full training may takes a few weeks I suppose.
ckpt_base_path: ./ # put absolute path here. Checkpoints will be loaded from ${ckpt_base_path}/outputs
The text was updated successfully, but these errors were encountered: