Instructions to use TheBloke/CodeLlama-7B-Instruct-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TheBloke/CodeLlama-7B-Instruct-GGUF with Transformers:

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

pipe = pipeline("text-generation", model="TheBloke/CodeLlama-7B-Instruct-GGUF")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("TheBloke/CodeLlama-7B-Instruct-GGUF", dtype="auto")

llama-cpp-python

How to use TheBloke/CodeLlama-7B-Instruct-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="TheBloke/CodeLlama-7B-Instruct-GGUF",
	filename="codellama-7b-instruct.Q2_K.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use TheBloke/CodeLlama-7B-Instruct-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

Use Docker

docker model run hf.co/TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use TheBloke/CodeLlama-7B-Instruct-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TheBloke/CodeLlama-7B-Instruct-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TheBloke/CodeLlama-7B-Instruct-GGUF",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

SGLang

How to use TheBloke/CodeLlama-7B-Instruct-GGUF 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 "TheBloke/CodeLlama-7B-Instruct-GGUF" \
    --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": "TheBloke/CodeLlama-7B-Instruct-GGUF",
		"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 "TheBloke/CodeLlama-7B-Instruct-GGUF" \
        --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": "TheBloke/CodeLlama-7B-Instruct-GGUF",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use TheBloke/CodeLlama-7B-Instruct-GGUF with Ollama:
```
ollama run hf.co/TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M
```

Unsloth Studio

How to use TheBloke/CodeLlama-7B-Instruct-GGUF 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 TheBloke/CodeLlama-7B-Instruct-GGUF 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 TheBloke/CodeLlama-7B-Instruct-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for TheBloke/CodeLlama-7B-Instruct-GGUF to start chatting

Atomic Chat new
Docker Model Runner
How to use TheBloke/CodeLlama-7B-Instruct-GGUF with Docker Model Runner:
```
docker model run hf.co/TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M
```

Lemonade

How to use TheBloke/CodeLlama-7B-Instruct-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull TheBloke/CodeLlama-7B-Instruct-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.CodeLlama-7B-Instruct-GGUF-Q4_K_M

List all available models

lemonade list

Very slow response on LM Studio with these settings

by yassersharaf - opened Apr 8, 2024

Discussion

yassersharaf

Apr 8, 2024

•

edited Apr 8, 2024

Hello,

This is my first time to try running local models using LM Studio and I downloaded this model:

{
  "name": "codellama_codellama-7b-instruct-hf",
  "arch": "llama",
  "quant": "Q4_K_S",
  "context_length": 16384,
  "embedding_length": 4096,
  "num_layers": 32,
  "rope": {
    "freq_base": 1000000,
    "dimension_count": 128
  },
  "head_count": 32,
  "head_count_kv": 32,
  "parameters": "7B"
}

I have a Windows 11 laptop running Intel i7 10th gen with 24GB RAM and a dedicated Nvidia Geforce GTX 1650 Ti 4GB display memory. And I have these settings for the model in LM Studio:

n_gpu_layers (GPU offload): 4
use_mlock (Keep entire model in RAM) set to true
n_threads (CPU Threads): 6
n_batch (Prompt eval batch size): 512
n_ctx (Context Length): 2048

But it takes so long to return the first token and it's slow also in writing the answers.
Is there anything wrong with these settings? Is there any other setting I should modify? I tried to disable n_gpu_layers by setting it to 0 and also tried increasing the value but it didn't solve the problem.

I reverted to the default settings below and it takes about 40 seconds to start responsing:

{
  "name": "Config for Chat ID 1712524642526",
  "load_params": {
    "n_ctx": 2048,
    "n_batch": 512,
    "rope_freq_base": 0,
    "rope_freq_scale": 0,
    "n_gpu_layers": 10,
    "use_mlock": true,
    "main_gpu": 0,
    "tensor_split": [
      0
    ],
    "seed": -1,
    "f16_kv": true,
    "use_mmap": true,
    "no_kv_offload": false,
    "num_experts_used": 0
  },
  "inference_params": {
    "n_threads": 4,
    "n_predict": -1,
    "top_k": 40,
    "min_p": 0.05,
    "top_p": 0.95,
    "temp": 0.8,
    "repeat_penalty": 1.1,
    "input_prefix": "\n### Instruction:\n",
    "input_suffix": "\n### Response:\n",
    "antiprompt": [
      "### Instruction:"
    ],
    "pre_prompt": "Below is an instruction that describes a task. Write a response that appropriately completes the request.",
    "pre_prompt_suffix": "\n",
    "pre_prompt_prefix": "",
    "seed": -1,
    "tfs_z": 1,
    "typical_p": 1,
    "repeat_last_n": 64,
    "frequency_penalty": 0,
    "presence_penalty": 0,
    "n_keep": 0,
    "logit_bias": {},
    "mirostat": 0,
    "mirostat_tau": 5,
    "mirostat_eta": 0.1,
    "memory_f16": true,
    "multiline_input": false,
    "penalize_nl": true
  }
}

Assist, please.

YaTharThShaRma999

Apr 8, 2024

@yassersharaf try offloading more layers, maybe 10? That might help(4 layers does not equal 4gb vram). Also set mlock to false.

yassersharaf

Apr 8, 2024

@YaTharThShaRma999 that helped it reduces response time to 15 seconds. how to find the best settings to decrease it to 2 or 3 seconds only if that's possible locally? Thanks a lot.

YaTharThShaRma999

Apr 8, 2024

@yassersharaf offload even more. Maybe 20? Try the max you can go before you get out of memory.

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