Instructions to use cmarkea/CodeLlama-70b-hf-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cmarkea/CodeLlama-70b-hf-4bit with Transformers:

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

pipe = pipeline("text-generation", model="cmarkea/CodeLlama-70b-hf-4bit")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cmarkea/CodeLlama-70b-hf-4bit")
model = AutoModelForCausalLM.from_pretrained("cmarkea/CodeLlama-70b-hf-4bit", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use cmarkea/CodeLlama-70b-hf-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cmarkea/CodeLlama-70b-hf-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cmarkea/CodeLlama-70b-hf-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/cmarkea/CodeLlama-70b-hf-4bit

SGLang

How to use cmarkea/CodeLlama-70b-hf-4bit 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 "cmarkea/CodeLlama-70b-hf-4bit" \
    --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": "cmarkea/CodeLlama-70b-hf-4bit",
		"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 "cmarkea/CodeLlama-70b-hf-4bit" \
        --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": "cmarkea/CodeLlama-70b-hf-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use cmarkea/CodeLlama-70b-hf-4bit with Docker Model Runner:
```
docker model run hf.co/cmarkea/CodeLlama-70b-hf-4bit
```

Converted version of CodeLlama-70b to 4-bit using bitsandbytes. For more information about the model, refer to the model's page.

Impact on performance

In the following figure, we can see the impact on the performance of a set of models relative to the required RAM space. It is noticeable that the quantized models have equivalent performance while providing a significant gain in RAM usage.

Downloads last month: 18

Safetensors

Model size

71B params

Tensor type

F32

BF16

Collection including cmarkea/CodeLlama-70b-hf-4bit

Quantized 4-bit models

Collection

Large model quantized with post-quantization performance very close to the original models, allowing it to run on reasonable infrastructure. • 9 items • Updated Mar 2 • 1