Instructions to use Trelis/deepseek-coder-33b-instruct-function-calling-v3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Trelis/deepseek-coder-33b-instruct-function-calling-v3 with Transformers:

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

pipe = pipeline("text-generation", model="Trelis/deepseek-coder-33b-instruct-function-calling-v3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Trelis/deepseek-coder-33b-instruct-function-calling-v3")
model = AutoModelForCausalLM.from_pretrained("Trelis/deepseek-coder-33b-instruct-function-calling-v3")
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 Trelis/deepseek-coder-33b-instruct-function-calling-v3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Trelis/deepseek-coder-33b-instruct-function-calling-v3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Trelis/deepseek-coder-33b-instruct-function-calling-v3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Trelis/deepseek-coder-33b-instruct-function-calling-v3

SGLang

How to use Trelis/deepseek-coder-33b-instruct-function-calling-v3 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 "Trelis/deepseek-coder-33b-instruct-function-calling-v3" \
    --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": "Trelis/deepseek-coder-33b-instruct-function-calling-v3",
		"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 "Trelis/deepseek-coder-33b-instruct-function-calling-v3" \
        --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": "Trelis/deepseek-coder-33b-instruct-function-calling-v3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Trelis/deepseek-coder-33b-instruct-function-calling-v3 with Docker Model Runner:
```
docker model run hf.co/Trelis/deepseek-coder-33b-instruct-function-calling-v3
```

Can this be dropped into openai compatible apis to support function calling / tools in langchain/langraph?

by rakataprime - opened Mar 13, 2024

Discussion

rakataprime

Mar 13, 2024

Can this be used with langchain / langchain tools through openai compatible apis? I would normally test this on my own, but since its paywalled I cannot. I am looking for an local model to run langchain/langgraph as an openai alternative, but most function calling models are not capable of function calling and coding in a compatible way.

RonanMcGovern

Trelis org Mar 13, 2024

Howdy @rakataprime , can you post one sample prompt (completely formatted, including bos, eos etc. tokens) here for me to see. That will help me tell you either a) this can work or b) what I need to do to adjust it to work.

rakataprime

Mar 14, 2024

here is a full end to end example that I have used for testing function calling with langgraph. https://github.com/rakataprime/local_llm_langgraph. I will use this to generate/format some sample prompts tomorrow including the tokens.

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