Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Allow conversion of Llama / Mistral HF models #6144

Merged
merged 8 commits into from
Mar 29, 2024
Merged

Allow conversion of Llama / Mistral HF models #6144

Changes from 1 commit
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
77f313e
Allow conversion of Mistral HF models
pcuenca Mar 18, 2024
eecaf58
Homogenize Llama, Mistral, Mixtral under the same entry.
pcuenca Mar 18, 2024
876b70d
Fix tokenizer, permute tensors
pcuenca Mar 26, 2024
23ffda0
Merge remote-tracking branch 'origin/master' into mistral-hf-conversion
pcuenca Mar 28, 2024
16ede02
Use sentencepiece tokenizer, or fall back to hfft.
pcuenca Mar 28, 2024
6dba2de
convert-hf : small fix for mypy
cebtenzzre Mar 28, 2024
9bf29e6
convert-hf : fix duplicated block_count
cebtenzzre Mar 28, 2024
f33b6c7
convert-hf : add vocab size to metadata
cebtenzzre Mar 28, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions convert-hf-to-gguf.py
View file
Open in desktop
Copy link
Collaborator

@cebtenzzre cebtenzzre Mar 18, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why not just use:

@Model.register("LlamaForCausalLM", "MistralForCausalLM", "MixtralForCausalLM")

on the existing MixtralModel (call it LlamaModel)? I don't see a point in supporting Mistral in this script without supporting Llama, and these classes are identical so they can be merged.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure, we can do that. I tried to mimic what I saw in the existing entries.

In my opinion, it could be interesting to use different architectures for Mistral (and Mixtral) for informative purposes. If you click on the GGUF information for any Mistral file on the Hugging Face Hub, there's nothing that refers to mistral except the filename. But that would also require additional changes, so happy to use the same entry for the three of them!

Original file line number Diff line number Diff line change
Expand Up @@ -1051,6 +1051,14 @@ def set_vocab(self):
self._set_vocab_sentencepiece()


@Model.register("MistralForCausalLM")
class MistralModel(Model):
model_arch = gguf.MODEL_ARCH.LLAMA

def set_vocab(self):
self._set_vocab_sentencepiece()


@Model.register("MiniCPMForCausalLM")
class MiniCPMModel(Model):
model_arch = gguf.MODEL_ARCH.MINICPM
Expand Down