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
How do I merge trained Lora an Llama7b weight? Is there a script? Would make it much easier to share weights, increase portability, file management, etc….
Would be an amazing feature of the training tab as well!
The text was updated successfully, but these errors were encountered:
I'm looking for that option too.. I want to merge back into HF weights so I can convert to ggml and use with llama.cpp eventualy. Did you manage to do it with the scriots in alpaca-weight repo ?
How do I merge trained Lora an Llama7b weight? Is there a script? Would make it much easier to share weights, increase portability, file management, etc….
Would be an amazing feature of the training tab as well!
The text was updated successfully, but these errors were encountered: