Fork of ms-swift to get it working with Qwen3-Omni and Earth Species Project NatureLM dataset.
https://spacecruft.org/deepcrayon/ms-swift
deepseek-r1
embedding
grpo
internvl
liger
llama
llama4
llm
lora
megatron
moe
multimodal
open-r1
peft
qwen3
qwen3-next
qwen3-omni
qwen3-vl
reranker
sft
- Python 99.8%
- Shell 0.2%
| .dev_scripts | ||
| .github | ||
| asset | ||
| docs | ||
| examples | ||
| requirements | ||
| scripts | ||
| swift | ||
| tests | ||
| .gitignore | ||
| .pre-commit-config.yaml | ||
| .pre-commit-config_local.yaml | ||
| CODE_OF_CONDUCT.md | ||
| CONTRIBUTING.md | ||
| CONTRIBUTING_CN.md | ||
| LICENSE | ||
| Makefile | ||
| MANIFEST.in | ||
| README-upstream.md | ||
| README.md | ||
| README_CN.md | ||
| requirements.txt | ||
| setup.cfg | ||
| setup.py | ||
ms-swift
This repo contains a fork of ModelScope's ms-swift. The fork has changes to train the Qwen3-Omni model with the Earth Species Project NatureLM dataset.
AniMUL
This code is used to create the AniMUL model.
Upstream
- https://github.com/modelscope/ms-swift
- https://huggingface.co/datasets/EarthSpeciesProject/NatureLM-audio-training
License
Apache 2.0.
Copyright © 2025 Jeff Moe.