Instructions to use facebook/KernelLLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use facebook/KernelLLM with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="facebook/KernelLLM")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("facebook/KernelLLM")
model = AutoModelForCausalLM.from_pretrained("facebook/KernelLLM")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use facebook/KernelLLM with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "facebook/KernelLLM"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "facebook/KernelLLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/facebook/KernelLLM

SGLang

How to use facebook/KernelLLM with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "facebook/KernelLLM" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "facebook/KernelLLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "facebook/KernelLLM" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "facebook/KernelLLM",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use facebook/KernelLLM with Docker Model Runner:
```
docker model run hf.co/facebook/KernelLLM
```

Zacharias030 commited on May 10, 2025

Commit

86d9bc4

verified ·

1 Parent(s): 1019e2b

Update README.md

Browse files

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -49,7 +49,7 @@ Our 8B parameter model achieves competitive or superior performance compared to
 The resulting model is competitive with state of the art LLMs despite its small size. We evaluate our model on KernelBench which is an open-source benchmark to evaluate the ability of LLMs to write efficient GPU kernels. It contains 250 selected PyTorch modules organized into difficulty levels, from single torch operators such as Conv2D or Swish (level 1), to full model architectures (level 3). The benchmark measures both correctness (by comparing against reference PyTorch outputs) and performance (by measuring speedup over baseline implementations). We implemented a new KernelBench-Triton variant that evaluates an LLMs ability to generate Triton kernels, making it an ideal benchmark for evaluating KernelLLM's capabilities. All our measurements were done on Nvidia H100 GPUs.
 ![pass at k analysis plot](media/kernelllm_pass_at_k_scaling.png)
 For more information, please see [Project Popcorn](https://gpu-mode.github.io/popcorn/).

 The resulting model is competitive with state of the art LLMs despite its small size. We evaluate our model on KernelBench which is an open-source benchmark to evaluate the ability of LLMs to write efficient GPU kernels. It contains 250 selected PyTorch modules organized into difficulty levels, from single torch operators such as Conv2D or Swish (level 1), to full model architectures (level 3). The benchmark measures both correctness (by comparing against reference PyTorch outputs) and performance (by measuring speedup over baseline implementations). We implemented a new KernelBench-Triton variant that evaluates an LLMs ability to generate Triton kernels, making it an ideal benchmark for evaluating KernelLLM's capabilities. All our measurements were done on Nvidia H100 GPUs.
 ![pass at k analysis plot](media/kernelllm_pass_at_k_scaling.png)
+*KernelLLM shows quasi log-linear scaling behavior during pass@k analysis.*
 For more information, please see [Project Popcorn](https://gpu-mode.github.io/popcorn/).