Instructions to use ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ with Transformers:

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

pipe = pipeline("text-generation", model="ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ")
model = AutoModelForCausalLM.from_pretrained("ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ

SGLang

How to use ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ 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 "ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ" \
    --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": "ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ",
		"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 "ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ" \
        --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": "ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ with Docker Model Runner:
```
docker model run hf.co/ArthurGprog/Codestral-22B-v0.1-FIM-Fix-GPTQ
```

This model is a GPTQ-quantized version of the Codestral-22B-v0.1 model with the FIM fix, designed to provide improved performance and efficiency. The original Codestral-22B-v0.1 model had a tokenization error for the FIM use case, which was addressed by the legraphista/Codestral-22B-v0.1-hf-FIM-fix model. This GPTQ-quantized model builds upon the fixed model, using default AutoGPTQ arguments and 4 bits per weight.

Compatible with vllm

Alternative to dan-kwiat/Codestral-22B-v0.1-hf-FIM-fix-awq

Tips for Code Completion

When using this model or any other Codestral-22B-v0.1 model for code completion in VSCode on Windows, make sure to use LF (Line Feed) instead of CRLF (Carriage Return Line Feed) for better results. This will prevent unrelated '\' characters from appearing in completions.

Downloads last month: 1

Safetensors

Model size

22B params

Tensor type

I32

F16