Source: https://seerepo.com/r/OptimalScale/LMFlow/built
Markdown source: https://seerepo.com/r/OptimalScale/LMFlow/built.md
Description: LMFlow is a Python repo built with jupyter-core and tornado. It uses AI through cpm-kernels, einops, NVIDIA, open-clip-torch, sentencepiece, and 14 more.

# What is LMFlow built with?

LMFlow is a Python repo built with jupyter-core and tornado. It uses AI through cpm-kernels, einops, NVIDIA, open-clip-torch, sentencepiece, and 14 more, makes its content with jinja2, markupsafe, and pillow, validates its data with Pydantic, is tested with Lit, runs in production with psutil, and uses pickleshare, pyarrow, pyzmq, and Hugging Face. It has 35,146 written lines, with 22 test files.

Read on November 17, 2026, at commit e9b1da0.

Repository: https://github.com/OptimalScale/LMFlow

## Frontend

### Styles

- **[cycler](https://seerepo.com/uses/cycler)**
- **[fonttools](https://seerepo.com/uses/fonttools)**

### Interface

- **[contourpy](https://seerepo.com/uses/contourpy)**
- **[matplotlib-inline](https://seerepo.com/uses/matplotlib-inline)**
- **[prompt-toolkit](https://seerepo.com/uses/prompt-toolkit)**
- **[matplotlib](https://seerepo.com/uses/matplotlib)**: 1 file

## Backend

### Data

- **pickleshare**
- **[pyarrow](https://seerepo.com/uses/pyarrow)**
- **[pyzmq](https://seerepo.com/uses/pyzmq)**

### AI

- **cpm-kernels**
- **[einops](https://seerepo.com/uses/einops)**
- **[NVIDIA](https://seerepo.com/uses/nvidia)**
- **open-clip-torch**
- **[sentencepiece](https://seerepo.com/uses/sentencepiece)**
- **[timm](https://seerepo.com/uses/timm)**
- **[tokenizers](https://seerepo.com/uses/tokenizers)**
- **[triton](https://seerepo.com/uses/triton)**
- **[torch](https://seerepo.com/uses/torch)**: 53 files
- **[transformers](https://seerepo.com/uses/transformers)**: 45 files
- **[datasets](https://seerepo.com/uses/datasets)**: 12 files
- **[accelerate](https://seerepo.com/uses/accelerate)**: 11 files
- **[diffusers](https://seerepo.com/uses/diffusers)**: 8 files
- **[peft](https://seerepo.com/uses/peft)**: 8 files
- **[bitsandbytes](https://seerepo.com/uses/bitsandbytes)**: 7 files
- **[torchvision](https://seerepo.com/uses/torchvision)**: 7 files
- **xformers**: 6 files
- **evaluate**: 1 file
- **tensordict**: 1 file

### Services

- **[Hugging Face](https://seerepo.com/uses/hugging-face)**: 6 files · [src/lmflow/pipeline/utils/raft_trainer.py, line 51](https://github.com/OptimalScale/LMFlow/blob/e9b1da012f08d17fadad698278f505800d30a8af/src/lmflow/pipeline/utils/raft_trainer.py#L51)

## Support

### The base

- **[jupyter-core](https://seerepo.com/uses/jupyter-core)**
- **[tornado](https://seerepo.com/uses/tornado)**

### Content

- **[jinja2](https://seerepo.com/uses/jinja2)**
- **[markupsafe](https://seerepo.com/uses/markupsafe)**
- **[pillow](https://seerepo.com/uses/pillow)**: 6 files

### Validation

- **[Pydantic](https://seerepo.com/uses/pydantic)**

### Testing

- **[Lit](https://seerepo.com/uses/lit)**

### Deploy

- **[psutil](https://seerepo.com/uses/psutil)**

### Tools

- **asttokens**
- **clip**
- **CMake**
- **debugpy**
- **executing**
- **ipykernel**
- **ipython**
- **jedi**
- **Jupyter**
- **mypy-extensions**
- **parso**
- **pip**
- **pure-eval**
- **pygments**
- **pyre-extensions**
- **Setuptools**
- **stack-data**
- **typing-extensions**
- **typing-inspect**
- **wheel**
- **packaging**: 9 files

### Other

- **backcall**
- **certifi**
- **charset-normalizer**
- **comm**
- **decorator**
- **filelock**
- **ftfy**
- **idna**
- **importlib-metadata**
- **importlib-resources**
- **kiwisolver**
- **mpmath**
- **nest-asyncio**
- **networkx**
- **pandas**
- **pexpect**
- **platformdirs**
- **protobuf**
- **ptyprocess**
- **pyparsing**
- **python-dateutil**
- **pytz**
- **pyyaml**
- **regex**
- **six**
- **SymPy**
- **traitlets**
- **tzdata**
- **urllib3**
- **wcwidth**
- **zipp**
- **NumPy**: 20 files
- **tqdm**: 14 files
- **wandb**: 10 files
- **requests**: 1 file

## Other repos with a similar stack

- [ray-project/ray](https://seerepo.com/r/ray-project/ray/built): pickleshare, pyarrow, pyzmq, Hugging Face, jupyter-core, tornado, cycler, fonttools, contourpy, matplotlib-inline, prompt-toolkit, einops, sentencepiece, timm, tokenizers, triton, torch, transformers, datasets, diffusers, peft, bitsandbytes, torchvision, jinja2, markupsafe, pillow, Pydantic, psutil · [Compare](https://seerepo.com/compare/OptimalScale/LMFlow/vs/ray-project/ray)
- [run-llama/llama_index](https://seerepo.com/r/run-llama/llama_index/built): pickleshare, pyarrow, pyzmq, jupyter-core, tornado, matplotlib-inline, prompt-toolkit, matplotlib, einops, open-clip-torch, sentencepiece, tokenizers, torch, transformers, datasets, peft, jinja2, markupsafe, pillow, Pydantic, psutil · [Compare](https://seerepo.com/compare/OptimalScale/LMFlow/vs/run-llama/llama_index)
- [onyx-dot-app/onyx](https://seerepo.com/r/onyx-dot-app/onyx/built): pyzmq, Hugging Face, jupyter-core, tornado, cycler, fonttools, contourpy, matplotlib-inline, prompt-toolkit, einops, NVIDIA, tokenizers, triton, torch, transformers, accelerate, jinja2, markupsafe, pillow, Pydantic, psutil · [Compare](https://seerepo.com/compare/onyx-dot-app/onyx/vs/OptimalScale/LMFlow)
- [mudler/LocalAI](https://seerepo.com/r/mudler/LocalAI/built): pyarrow, Hugging Face, matplotlib, einops, NVIDIA, sentencepiece, tokenizers, torch, transformers, datasets, accelerate, diffusers, peft, bitsandbytes, torchvision, jinja2, markupsafe, pillow · [Compare](https://seerepo.com/compare/mudler/LocalAI/vs/OptimalScale/LMFlow)
- [hpcaitech/ColossalAI](https://seerepo.com/r/hpcaitech/ColossalAI/built): Hugging Face, matplotlib, einops, sentencepiece, triton, torch, transformers, datasets, accelerate, diffusers, peft, bitsandbytes, torchvision, jinja2, markupsafe, pillow, Pydantic, psutil · [Compare](https://seerepo.com/compare/hpcaitech/ColossalAI/vs/OptimalScale/LMFlow)
- [harvard-edge/cs249r_book](https://seerepo.com/r/harvard-edge/cs249r_book/built): pyarrow, pyzmq, jupyter-core, tornado, cycler, fonttools, contourpy, matplotlib-inline, prompt-toolkit, matplotlib, torch, transformers, torchvision, jinja2, markupsafe, pillow, Pydantic, psutil · [Compare](https://seerepo.com/compare/harvard-edge/cs249r_book/vs/OptimalScale/LMFlow)

## Frequently asked questions

### What language is LMFlow written in?

LMFlow is written in Python, with 35,146 written lines.

### What framework does LMFlow use?

LMFlow is built with jupyter-core and tornado.

### Does LMFlow use AI?

Yes. LMFlow uses AI through cpm-kernels, einops, NVIDIA, open-clip-torch, sentencepiece, timm, tokenizers, triton, torch, transformers, datasets, accelerate, diffusers, peft, bitsandbytes, torchvision, xformers, evaluate, and tensordict.

### What services does LMFlow use?

LMFlow uses pickleshare, pyarrow, pyzmq, and Hugging Face.

### Does LMFlow have tests?

Yes. LMFlow has 22 test files and is tested with Lit.

### Which repos have a stack like LMFlow's?

The ones that share most of its stack: ray-project/ray, run-llama/llama_index, and onyx-dot-app/onyx.

## What is your repo built with?

Any public repo, with no account: https://seerepo.com/r/owner/repo/built, with the owner and name it has on GitHub.
