Instructions to use moonshotai/Kimi-Linear-48B-A3B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use moonshotai/Kimi-Linear-48B-A3B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="moonshotai/Kimi-Linear-48B-A3B-Instruct", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("moonshotai/Kimi-Linear-48B-A3B-Instruct", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("moonshotai/Kimi-Linear-48B-A3B-Instruct", trust_remote_code=True, device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use moonshotai/Kimi-Linear-48B-A3B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "moonshotai/Kimi-Linear-48B-A3B-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "moonshotai/Kimi-Linear-48B-A3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/moonshotai/Kimi-Linear-48B-A3B-Instruct

SGLang

How to use moonshotai/Kimi-Linear-48B-A3B-Instruct 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 "moonshotai/Kimi-Linear-48B-A3B-Instruct" \
    --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": "moonshotai/Kimi-Linear-48B-A3B-Instruct",
		"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 "moonshotai/Kimi-Linear-48B-A3B-Instruct" \
        --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": "moonshotai/Kimi-Linear-48B-A3B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use moonshotai/Kimi-Linear-48B-A3B-Instruct with Docker Model Runner:
```
docker model run hf.co/moonshotai/Kimi-Linear-48B-A3B-Instruct
```

Non-existant API Call - Request for Help

#21

by sjlee1234 - opened Feb 18

Discussion

sjlee1234

Feb 18

When this is called: transformers.AutoModelForCausalLM.from_pretrained("tiny-random/kimi-linear") to load a smaller, randomly initialized version of kimi linear,
It comes up with this error: ImportError: cannot import name 'OutputRecorder' from 'transformers.utils.generic' (/opt/homebrew/lib/python3.14/site-packages/transformers/utils/generic.py) which originates from modeling_kimi.py from moonshotai/Kimi-Linear-48B-A3B-Instruct.
I am currently using transformers==5.2.0 and would like a workaround or a fix.
Thank you.

Full error message:
File /opt/homebrew/lib/python3.14/site-packages/transformers/models/auto/auto_factory.py:356, in _BaseAutoModelClass.from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs)
353 kwargs["adapter_kwargs"] = adapter_kwargs
355 if has_remote_code and trust_remote_code:
--> 356 model_class = get_class_from_dynamic_module(
357 class_ref, pretrained_model_name_or_path, code_revision=code_revision, **hub_kwargs, **kwargs
358 )
359 _ = hub_kwargs.pop("code_revision", None)
360 # This block handles the case where the user is loading a model with trust_remote_code=True
361 # but a library model exists with the same name. We don't want to override the autoclass
362 # mappings in this case, or all future loads of that model will be the remote code model.

File /opt/homebrew/lib/python3.14/site-packages/transformers/dynamic_module_utils.py:583, in get_class_from_dynamic_module(class_reference, pretrained_model_name_or_path, cache_dir, force_download, proxies, token, revision, local_files_only, repo_type, code_revision, **kwargs)
571 # And lastly we get the class inside our newly created module
572 final_module = get_cached_module_file(
573 repo_id,
574 module_file + ".py",
(...) 581 repo_type=repo_type,
582 )
--> 583 return get_class_in_module(class_name, final_module, force_reload=force_download)
...
---> 21 from transformers.utils.generic import OutputRecorder, check_model_inputs
23 try:
24 from fla.modules import FusedRMSNormGated, ShortConvolution

ImportError: cannot import name 'OutputRecorder' from 'transformers.utils.generic' (/opt/homebrew/lib/python3.14/site-packages/transformers/utils/generic.py)

sjlee1234 changed discussion title from Non-existant API Call - Request for Update to Non-existant API Call - Request for Help Feb 18

sjlee1234 changed discussion status to closed Feb 19

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment