Instructions to use inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i 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("inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i")

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 Settings
LM Studio

How to use inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i 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 "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i"

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 inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "inferencerlabs/DeepSeek-V4-Flash-MLX-Q2i" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

inferencerlabs commited on May 22

Commit

1d2ae77

verified ·

1 Parent(s): 82db41e

Update README.md

Browse files

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -15,9 +15,9 @@ pipeline_tag: text-generation
 - Text inference: ~30 tokens/s @ 2000 tokens ~95.49 GiB (debug build)
 <p style="margin-bottom:0px;">
-<strong>Q2.8-EXP is an experimental build of DeepSeek-V4-Flash</strong>
-This build compresses the model to operate within the memory constraints of a 128 GiB RAM system while maintaining response coherence. However, overall accuracy is degraded due to the level of compression. Stay tuned for updates.
 </p>
 ![Screenshot](https://cdn-uploads.huggingface.co/production/uploads/688479d616f1ec82fa645019/TphT4hr5CABuKO1MGccqh.jpeg)

 - Text inference: ~30 tokens/s @ 2000 tokens ~95.49 GiB (debug build)
 <p style="margin-bottom:0px;">
+<strong>Q2.8-INF uses the data-agnostic INF method tuned to yield maximum general accuracy within a 128 GiB memory budget</strong>
+This build compresses the model to operate within the memory constraints of a 128 GiB RAM system while maintaining response coherence. However, overall accuracy is degraded due to the level of compression.
 </p>
 ![Screenshot](https://cdn-uploads.huggingface.co/production/uploads/688479d616f1ec82fa645019/TphT4hr5CABuKO1MGccqh.jpeg)