Instructions to use NimbleAINinja/Beepo-22B-mlx-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NimbleAINinja/Beepo-22B-mlx-4bit with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("NimbleAINinja/Beepo-22B-mlx-4bit")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps
LM Studio

Pi new

How to use NimbleAINinja/Beepo-22B-mlx-4bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "NimbleAINinja/Beepo-22B-mlx-4bit"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "NimbleAINinja/Beepo-22B-mlx-4bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use NimbleAINinja/Beepo-22B-mlx-4bit with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "NimbleAINinja/Beepo-22B-mlx-4bit"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default NimbleAINinja/Beepo-22B-mlx-4bit

Run Hermes

hermes

MLX LM

How to use NimbleAINinja/Beepo-22B-mlx-4bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "NimbleAINinja/Beepo-22B-mlx-4bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "NimbleAINinja/Beepo-22B-mlx-4bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "NimbleAINinja/Beepo-22B-mlx-4bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Beepo-22B-GGUF

This is the MLX 4bit quantization of https://huggingface.co/concedo/Beepo-22B, which was originally finetuned on top of the https://huggingface.co/mistralai/Mistral-Small-Instruct-2409 model.

You can use LMStudio to run this model.

Key Features:

Retains Intelligence - LR was kept low and dataset heavily pruned to avoid losing too much of the original model's intelligence.
Instruct prompt format supports Alpaca - Honestly, I don't know why more models don't use it. If you are an Alpaca format lover like me, this should help. The original Mistral instruct format can still be used, but is not recommended.
Instruct Decensoring Applied - You should not need a jailbreak for a model to obey the user. The model should always do what you tell it to. No need for weird "Sure, I will" or kitten-murdering-threat tricks. No abliteration was done, only finetuning. This model is not evil. It does not judge or moralize. Like a good tool, it simply obeys.

Prompt template: Alpaca

### Instruction:
{prompt}

### Response:

Please leave any feedback or issues that you may have.

Downloads last month: 9

Safetensors

Model size

3B params

Tensor type

F16

U32

MLX

Hardware compatibility

4-bit

Model tree for NimbleAINinja/Beepo-22B-mlx-4bit

Base model

mistralai/Mistral-Small-Instruct-2409

Finetuned

concedo/Beepo-22B

Quantized

(9)

this model