flash-infer

flash-infer

CONTAINERS IMAGES RUN BUILD

CONTAINERS

flashinfer:0.2.1.post2
Requires	L4T ['>=35']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn python numpy cmake onnx pytorch triton
Dependants	vllm:0.8.4
Dockerfile	Dockerfile

flashinfer:0.2.2
Requires	L4T ['>=35']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn python numpy cmake onnx pytorch triton
Dockerfile	Dockerfile

flashinfer:0.2.2.post1
Requires	L4T ['>=35']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn python numpy cmake onnx pytorch triton
Dependants	vllm:v0.8.5.post1
Dockerfile	Dockerfile

flashinfer:0.2.6.post1
Requires	L4T ['>=35']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn python numpy cmake onnx pytorch triton
Dependants	mlc:0.20.0 sglang:0.4.6
Dockerfile	Dockerfile

flashinfer:0.2.7
Requires	L4T ['>=35']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn python numpy cmake onnx pytorch triton
Dockerfile	Dockerfile

flashinfer:0.2.8
Aliases	flashinfer
Requires	L4T ['>=35']
Dependencies	build-essential pip_cache:cu126 cuda:12.6 cudnn python numpy cmake onnx pytorch triton
Dependants	cosmos-reason1 dynamo:0.3.2 genai-bench:0.1.0 l4t-dynamo l4t-text-generation llama-factory local_llm mlc:0.21.0 nano_llm:24.4 nano_llm:24.4-foxy nano_llm:24.4-galactic nano_llm:24.4-humble nano_llm:24.4-iron nano_llm:24.4.1 nano_llm:24.4.1-foxy nano_llm:24.4.1-galactic nano_llm:24.4.1-humble nano_llm:24.4.1-iron nano_llm:24.5 nano_llm:24.5-foxy nano_llm:24.5-galactic nano_llm:24.5-humble nano_llm:24.5-iron nano_llm:24.5.1 nano_llm:24.5.1-foxy nano_llm:24.5.1-galactic nano_llm:24.5.1-humble nano_llm:24.5.1-iron nano_llm:24.6 nano_llm:24.6-foxy nano_llm:24.6-galactic nano_llm:24.6-humble nano_llm:24.6-iron nano_llm:24.7 nano_llm:24.7-foxy nano_llm:24.7-galactic nano_llm:24.7-humble nano_llm:24.7-iron nano_llm:main nano_llm:main-foxy nano_llm:main-galactic nano_llm:main-humble nano_llm:main-iron radial-attention:0.1.0 sgl-kernel:0.2.3 sglang:0.4.4 sglang:0.4.9 vllm:0.7.4 vllm:0.9.0 vllm:0.9.2 vllm:0.9.3
Dockerfile	Dockerfile

RUN CONTAINER

To start the container, you can use jetson-containers run and autotag, or manually put together a docker run command:

# automatically pull or build a compatible container image
jetson-containers run $(autotag flash-infer)

# or if using 'docker run' (specify image and mounts/ect)
sudo docker run --runtime nvidia -it --rm --network=host flash-infer:36.4.0

^{jetson-containers run forwards arguments to docker run with some defaults added (like --runtime nvidia, mounts a /data cache, and detects devices)
autotag finds a container image that's compatible with your version of JetPack/L4T - either locally, pulled from a registry, or by building it.}

To mount your own directories into the container, use the -v or --volume flags:

jetson-containers run -v /path/on/host:/path/in/container $(autotag flash-infer)

To launch the container running a command, as opposed to an interactive shell:

jetson-containers run $(autotag flash-infer) my_app --abc xyz

You can pass any options to it that you would to docker run, and it'll print out the full command that it constructs before executing it.

BUILD CONTAINER

If you use autotag as shown above, it'll ask to build the container for you if needed. To manually build it, first do the system setup, then run:

jetson-containers build flash-infer

The dependencies from above will be built into the container, and it'll be tested during. Run it with --help for build options.