Source: https://seerepo.com/r/InternLM/lmdeploy/built
Markdown source: https://seerepo.com/r/InternLM/lmdeploy/built.md
Description: lmdeploy is a Python repo built with FastAPI, Ray, AIOHTTP, and uvicorn.

# What is lmdeploy built with?

lmdeploy is a Python repo built with FastAPI, Ray, AIOHTTP, and uvicorn. It uses AI through apache-tvm-ffi, flash-linear-attention, openai-harmony, partial-json-parser, qwen-vl-utils, and 18 more, makes its content with markdown, myst-parser, and pillow, validates its data with Pydantic, is tested with allure-pytest, pytest-cov, pytest-order, pytest-rerunfailures, pytest-sugar, and 3 more, runs in production with prometheus-client, and uses pyzmq. It has 299,565 written lines, with 336 test files.

Read on November 17, 2026, at commit fc33f69.

Repository: https://github.com/InternLM/lmdeploy

## Frontend

### Styles

- **sphinx-book-theme**

### Interface

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

## Backend

### The base

- **[FastAPI](https://seerepo.com/uses/fastapi)**: 24 files

### Data

- **[pyzmq](https://seerepo.com/uses/pyzmq)**: 3 files

### AI

- **[apache-tvm-ffi](https://seerepo.com/uses/apache-tvm-ffi)**
- **flash-linear-attention**
- **openai-harmony**
- **partial-json-parser**
- **[qwen-vl-utils](https://seerepo.com/uses/qwen-vl-utils)**
- **rdkit**
- **transformers-stream-generator**
- **[torch](https://seerepo.com/uses/torch)**: 426 files
- **[transformers](https://seerepo.com/uses/transformers)**: 123 files
- **[triton](https://seerepo.com/uses/triton)**: 50 files
- **xgrammar**: 8 files
- **[accelerate](https://seerepo.com/uses/accelerate)**: 7 files
- **[safetensors](https://seerepo.com/uses/safetensors)**: 4 files
- **[torchvision](https://seerepo.com/uses/torchvision)**: 4 files
- **[peft](https://seerepo.com/uses/peft)**: 2 files
- **torch-npu**: 2 files
- **[datasets](https://seerepo.com/uses/datasets)**: 1 file
- **[einops](https://seerepo.com/uses/einops)**: 1 file
- **[sentencepiece](https://seerepo.com/uses/sentencepiece)**: 1 file
- **[tiktoken](https://seerepo.com/uses/tiktoken)**: 1 file
- **[timm](https://seerepo.com/uses/timm)**: 1 file

### Services

- **[Anthropic](https://seerepo.com/uses/anthropic)**: 1 file · [autotest/utils/anthropic_messages.py, line 370](https://github.com/InternLM/lmdeploy/blob/fc33f69736ff48c10ad720593b021dcdaedcb80d/autotest/utils/anthropic_messages.py#L370)
- **[OpenAI](https://seerepo.com/uses/openai)**: 1 file · [autotest/utils/run_restful_chat.py, line 10](https://github.com/InternLM/lmdeploy/blob/fc33f69736ff48c10ad720593b021dcdaedcb80d/autotest/utils/run_restful_chat.py#L10)

## Support

### The base

- **[Ray](https://seerepo.com/uses/ray)**: 7 files
- **[AIOHTTP](https://seerepo.com/uses/aiohttp)**: 3 files
- **[uvicorn](https://seerepo.com/uses/uvicorn)**: 2 files

### Content

- **[markdown](https://seerepo.com/uses/markdown)**
- **[myst-parser](https://seerepo.com/uses/myst-parser)**
- **[pillow](https://seerepo.com/uses/pillow)**: 10 files

### Validation

- **[Pydantic](https://seerepo.com/uses/pydantic)**: 13 files

### Testing

- **allure-pytest**
- **pytest-cov**
- **pytest-order**
- **pytest-rerunfailures**
- **pytest-sugar**
- **pytest-xdist**
- **Pytest**: 260 files
- **[jsonschema](https://seerepo.com/uses/jsonschema)**: 6 files

### Deploy

- **[prometheus-client](https://seerepo.com/uses/prometheus-client)**: 2 files

### Tools

- **pybind11**
- **Setuptools**
- **Sphinx**
- **sphinx-autodoc-typehints**
- **sphinx-copybutton**
- **sphinx-tabs**
- **sphinxcontrib-mermaid**
- **sphinxcontrib-openapi**
- **fire**: 9 files
- **tilelang**: 8 files
- **coverage**: 1 file

### Other

- **nvidia-ml-py**
- **protobuf**
- **NumPy**: 55 files
- **shortuuid**: 11 files
- **pybase64**: 5 files
- **pyyaml**: 2 files
- **cloudpickle**: 1 file
- **opencv-python-headless**: 1 file
- **urllib3**: 1 file

## Other repos with a similar stack

- [vllm-project/vllm](https://seerepo.com/r/vllm-project/vllm/built): FastAPI, pyzmq, OpenAI, Ray, AIOHTTP, apache-tvm-ffi, openai-harmony, partial-json-parser, torch, transformers, triton, xgrammar, safetensors, torchvision, einops, sentencepiece, tiktoken, timm, markdown, pillow, Pydantic, prometheus-client · [Compare](https://seerepo.com/compare/InternLM/lmdeploy/vs/vllm-project/vllm)
- [ray-project/ray](https://seerepo.com/r/ray-project/ray/built): FastAPI, pyzmq, Ray, AIOHTTP, uvicorn, torch, transformers, triton, safetensors, torchvision, peft, datasets, einops, sentencepiece, tiktoken, timm, markdown, myst-parser, pillow, Pydantic, prometheus-client · [Compare](https://seerepo.com/compare/InternLM/lmdeploy/vs/ray-project/ray)
- [sgl-project/sglang](https://seerepo.com/r/sgl-project/sglang/built): FastAPI, pyzmq, Anthropic, OpenAI, AIOHTTP, uvicorn, apache-tvm-ffi, openai-harmony, partial-json-parser, torch, transformers, xgrammar, torchvision, einops, sentencepiece, tiktoken, timm, pillow, Pydantic, prometheus-client · [Compare](https://seerepo.com/compare/InternLM/lmdeploy/vs/sgl-project/sglang)
- [hpcaitech/ColossalAI](https://seerepo.com/r/hpcaitech/ColossalAI/built): FastAPI, OpenAI, Ray, uvicorn, matplotlib, torch, transformers, triton, accelerate, safetensors, torchvision, peft, datasets, einops, sentencepiece, tiktoken, pillow, Pydantic · [Compare](https://seerepo.com/compare/hpcaitech/ColossalAI/vs/InternLM/lmdeploy)
- [bojieli/ai-agent-book](https://seerepo.com/r/bojieli/ai-agent-book/built): FastAPI, Anthropic, OpenAI, AIOHTTP, uvicorn, matplotlib, torch, transformers, triton, accelerate, peft, datasets, sentencepiece, tiktoken, markdown, pillow, Pydantic · [Compare](https://seerepo.com/compare/bojieli/ai-agent-book/vs/InternLM/lmdeploy)
- [unslothai/unsloth](https://seerepo.com/r/unslothai/unsloth/built): FastAPI, OpenAI, uvicorn, matplotlib, torch, transformers, accelerate, safetensors, peft, datasets, einops, sentencepiece, tiktoken, timm, pillow, Pydantic · [Compare](https://seerepo.com/compare/InternLM/lmdeploy/vs/unslothai/unsloth)

## Frequently asked questions

### What language is lmdeploy written in?

lmdeploy is written in Python, with 299,565 written lines.

### What framework does lmdeploy use?

lmdeploy is built with FastAPI, Ray, AIOHTTP, and uvicorn.

### Does lmdeploy use AI?

Yes. lmdeploy uses AI through apache-tvm-ffi, flash-linear-attention, openai-harmony, partial-json-parser, qwen-vl-utils, rdkit, transformers-stream-generator, torch, transformers, triton, xgrammar, accelerate, safetensors, torchvision, peft, torch-npu, datasets, einops, sentencepiece, tiktoken, timm, Anthropic, and OpenAI.

### What services does lmdeploy use?

lmdeploy uses pyzmq.

### Does lmdeploy have tests?

Yes. lmdeploy has 336 test files and is tested with allure-pytest, pytest-cov, pytest-order, pytest-rerunfailures, pytest-sugar, pytest-xdist, Pytest, and jsonschema.

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

The ones that share most of its stack: vllm-project/vllm, ray-project/ray, and sgl-project/sglang.

## 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.
