Instructions to use unsloth/DeepSeek-V3-bf16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use unsloth/DeepSeek-V3-bf16 with Transformers:

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

pipe = pipeline("text-generation", model="unsloth/DeepSeek-V3-bf16", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("unsloth/DeepSeek-V3-bf16", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("unsloth/DeepSeek-V3-bf16", trust_remote_code=True)
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use unsloth/DeepSeek-V3-bf16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "unsloth/DeepSeek-V3-bf16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "unsloth/DeepSeek-V3-bf16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/unsloth/DeepSeek-V3-bf16

SGLang

How to use unsloth/DeepSeek-V3-bf16 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 "unsloth/DeepSeek-V3-bf16" \
    --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": "unsloth/DeepSeek-V3-bf16",
		"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 "unsloth/DeepSeek-V3-bf16" \
        --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": "unsloth/DeepSeek-V3-bf16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio new

How to use unsloth/DeepSeek-V3-bf16 with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for unsloth/DeepSeek-V3-bf16 to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for unsloth/DeepSeek-V3-bf16 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for unsloth/DeepSeek-V3-bf16 to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="unsloth/DeepSeek-V3-bf16",
    max_seq_length=2048,
)

Docker Model Runner
How to use unsloth/DeepSeek-V3-bf16 with Docker Model Runner:
```
docker model run hf.co/unsloth/DeepSeek-V3-bf16
```

Encountering Unknown quantization type, got fp8 - supported types are: XXXXX

by ivanmanu - opened Jan 14, 2025

Discussion

ivanmanu

Jan 14, 2025

Dear developers and community users,

I tried to load this unsloth/DeepSeek-V3-bf16 via AutoModelForCausalLM.from_pretrained(" unsloth/DeepSeek-V3-bf16 ", trust_remote_code=True)

But encountered the below error:

Unknown quantization type, got fp8 - supported types are: ['awq', 'bitsandbytes_4bit', 'bitsandbytes_8bit', 'gptq', 'aqlm', 'quanto', 'eetq', 'hqq', 'compressed-tensors', 'fbgemm_fp8', 'torchao', 'bitnet']

I tried with different transformers versions from 4.33.1, to 4.55, but none worked.

The config file of your model showed

https://huggingface.co/unsloth/DeepSeek-V3-bf16/blob/main/config.json

"torch_dtype": "bfloat16",
"transformers_version": "4.33.1",

Please advise of any fixes? Thanks

shimmyshimmer

Unsloth AI org Jan 15, 2025

Dear developers and community users,

I tried to load this unsloth/DeepSeek-V3-bf16 via AutoModelForCausalLM.from_pretrained(" unsloth/DeepSeek-V3-bf16 ", trust_remote_code=True)

But encountered the below error:

Unknown quantization type, got fp8 - supported types are: ['awq', 'bitsandbytes_4bit', 'bitsandbytes_8bit', 'gptq', 'aqlm', 'quanto', 'eetq', 'hqq', 'compressed-tensors', 'fbgemm_fp8', 'torchao', 'bitnet']

I tried with different transformers versions from 4.33.1, to 4.55, but none worked.

The config file of your model showed

https://huggingface.co/unsloth/DeepSeek-V3-bf16/blob/main/config.json

"torch_dtype": "bfloat16",
"transformers_version": "4.33.1",

Please advise of any fixes? Thanks

ohhh im not sure if hugging face implemented the support for it. :(

ivanmanu

Jan 15, 2025

Dear developers and community users,

I tried to load this unsloth/DeepSeek-V3-bf16 via AutoModelForCausalLM.from_pretrained(" unsloth/DeepSeek-V3-bf16 ", trust_remote_code=True)

But encountered the below error:

Unknown quantization type, got fp8 - supported types are: ['awq', 'bitsandbytes_4bit', 'bitsandbytes_8bit', 'gptq', 'aqlm', 'quanto', 'eetq', 'hqq', 'compressed-tensors', 'fbgemm_fp8', 'torchao', 'bitnet']

I tried with different transformers versions from 4.33.1, to 4.55, but none worked.

The config file of your model showed

https://huggingface.co/unsloth/DeepSeek-V3-bf16/blob/main/config.json

"torch_dtype": "bfloat16",
"transformers_version": "4.33.1",

Please advise of any fixes? Thanks

ohhh im not sure if hugging face implemented the support for it. :(

Could you please advise how I can load the model and run it? If HF's AutoModelForCausalLM.from_pretrained() doesn't support it?
thanks.

fsaudm

Jan 21, 2025

•

edited Jan 21, 2025

@ivanmanu

Try with vLLM or llama.cpp. Amazing tools overall!

I find vllm to be easier to start with: https://docs.vllm.ai/en/latest/

Though, llama.cpp has been great as well. I was able to fully run DeepSeek-V3-Q2_K_XS/DeepSeek-V3-Q2_K_XS.gguf on CPU exclusively. If you have the GPUs, you could go brrr.

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