Instructions to use cortexso/codestral with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cortexso/codestral with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="cortexso/codestral",
	filename="codestral-22b-v0.1-q2_k.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use cortexso/codestral with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/codestral:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/codestral:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/codestral:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/codestral: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 cortexso/codestral:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf cortexso/codestral: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 cortexso/codestral:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf cortexso/codestral:Q4_K_M

Use Docker

docker model run hf.co/cortexso/codestral:Q4_K_M

LM Studio
Jan

vLLM

How to use cortexso/codestral with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/cortexso/codestral:Q4_K_M

Ollama
How to use cortexso/codestral with Ollama:
```
ollama run hf.co/cortexso/codestral:Q4_K_M
```

Unsloth Studio new

How to use cortexso/codestral 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 cortexso/codestral 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 cortexso/codestral to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for cortexso/codestral to start chatting

Docker Model Runner
How to use cortexso/codestral with Docker Model Runner:
```
docker model run hf.co/cortexso/codestral:Q4_K_M
```

Lemonade

How to use cortexso/codestral with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull cortexso/codestral:Q4_K_M

Run and chat with the model

lemonade run user.codestral-Q4_K_M

List all available models

lemonade list

jan-hq commited on Jun 19, 2024

Commit

937563f

verified ·

1 Parent(s): 3622cca

Update model.yml

Browse files

Files changed (1) hide show

model.yml +7 -14

model.yml CHANGED Viewed

@@ -1,23 +1,16 @@
 name: Codestreal 22B
 model: codestral:22B
 version: 1
 files:
-	- llama_model_path: model.gguf
-# Results Preferences
 top_p: 0.95
 temperature: 0.7
 frequency_penalty: 0
 presence_penalty: 0
-max_tokens: 32000 # Infer from base config.json -> max_position_embeddings
-stream: true # true | false
-# Engine / Model Settings
-ngl: 32 # Infer from base config.json -> num_attention_heads
-ctx_len: 32000 # Infer from base config.json -> max_position_embeddings
 engine: cortex.llamacpp
-prompt_template: "{system_message} [INST] {prompt} [/INST]"
-# Prompt template: Can only be retrieved from instruct model
-# - https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct/blob/main/tokenizer_config.json#L2053
-# - Requires jinja format parser

+---
 name: Codestreal 22B
 model: codestral:22B
 version: 1
 files:
+  - llama_model_path: model.gguf
 top_p: 0.95
 temperature: 0.7
 frequency_penalty: 0
 presence_penalty: 0
+max_tokens: 32000
+stream: true
+ngl: 32
+ctx_len: 32000
 engine: cortex.llamacpp
+prompt_template: "{system_message} [INST] {prompt} [/INST]"