ggml-org/ggml

[Feature Request] rinna's Japanese GPT model support

Open

#33 opened on Mar 4, 2023

View on GitHub
 (2 comments) (1 reaction) (0 assignees)C++ (14,741 stars) (1,646 forks)auto 404
enhancementgood first issue

Description

Thanks @ggerganov for your sharing.

I want to use GPT on my local pc. rinna Co.,Ltd is Japanese AI company. rinna provides some GPT models on huggingface. https://huggingface.co/rinna/japanese-gpt-1b https://huggingface.co/rinna/japanese-gpt2-xsmall https://huggingface.co/rinna/japanese-gpt2-small https://huggingface.co/rinna/japanese-gpt2-medium I'd like to use these models with ggml gpt. But, I can't convert these models for ggml. I think these models are pytorch bin model and tensorflow H5 weight.

Could you convert and support these models ?

Contributor guide

Tech stack
pythonpytorch
Domain
machine learning
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
3-5 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
needs maintainer response
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
PythonPyTorchGitModel conversion knowledge
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
25
Research direction
Study the architecture of the listed rinna Japanese GPT models (1b, xsmall, small, medium) and understand the current ggml conversion tools (e.g., convert pt to ggml.py). Determine the differences from other supported GPT models and implement the necessary changes to support these models.