Instructions to use m-a-p/OpenCodeInterpreter-DS-33B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use m-a-p/OpenCodeInterpreter-DS-33B with Transformers:

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

pipe = pipeline("text-generation", model="m-a-p/OpenCodeInterpreter-DS-33B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("m-a-p/OpenCodeInterpreter-DS-33B")
model = AutoModelForCausalLM.from_pretrained("m-a-p/OpenCodeInterpreter-DS-33B")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use m-a-p/OpenCodeInterpreter-DS-33B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "m-a-p/OpenCodeInterpreter-DS-33B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "m-a-p/OpenCodeInterpreter-DS-33B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/m-a-p/OpenCodeInterpreter-DS-33B

SGLang

How to use m-a-p/OpenCodeInterpreter-DS-33B 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 "m-a-p/OpenCodeInterpreter-DS-33B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "m-a-p/OpenCodeInterpreter-DS-33B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "m-a-p/OpenCodeInterpreter-DS-33B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "m-a-p/OpenCodeInterpreter-DS-33B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use m-a-p/OpenCodeInterpreter-DS-33B with Docker Model Runner:
```
docker model run hf.co/m-a-p/OpenCodeInterpreter-DS-33B
```

Can someone point to the chat template used for this model? Would be awesome.

by HriDal - opened Mar 24, 2024

Discussion

HriDal

Mar 24, 2024

When I use the .apply_chat_template, I get this warning: No chat template is defined for this tokenizer - using the default template for the LlamaTokenizerFast class. If the default is not appropriate for your model, please set tokenizer.chat_template to an appropriate template. See https://huggingface.co/docs/transformers/main/chat_templating for more information.

NoneGG

May 11, 2024

If you use tokenizer.apply_chat_template and there is no additional "chat_template" in the model tokenizer_config.json, a default template of torch package is used.
My problem is why the chat templated is provided inm-a-p/OpenCodeInterpreter-DS-6.7B but not provided in m-a-p/OpenCodeInterpreter-DS-33B.
And if i apply the chat template of 6.7B for 33B, 33B will return response in wierd format, which weaken its accuracy in HumanEval test.

HriDal

May 13, 2024

Thanks I figured it out later, but was confused as it did not seem to have a system prompt at that time..

NoneGG

Jun 8, 2024

How did you solve that problem? Just make a Jinja template for it?

HriDal

Jun 17, 2024

I just utilized a normal template of you are a smart assistant and helpful, etc. I applied it manually by giving this template to the tokenizer and then just applying it using tokenizer.apply_template

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