Instructions to use jrad123777/effect-qwen36-35b-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jrad123777/effect-qwen36-35b-mlx 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("jrad123777/effect-qwen36-35b-mlx")

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 jrad123777/effect-qwen36-35b-mlx with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "jrad123777/effect-qwen36-35b-mlx"

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": "jrad123777/effect-qwen36-35b-mlx"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use jrad123777/effect-qwen36-35b-mlx with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "jrad123777/effect-qwen36-35b-mlx"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "jrad123777/effect-qwen36-35b-mlx"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "jrad123777/effect-qwen36-35b-mlx",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use jrad123777/effect-qwen36-35b-mlx 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 "jrad123777/effect-qwen36-35b-mlx"

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 jrad123777/effect-qwen36-35b-mlx

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use jrad123777/effect-qwen36-35b-mlx with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "jrad123777/effect-qwen36-35b-mlx"

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 "jrad123777/effect-qwen36-35b-mlx" \
  --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"

jrad123777 commited on Jun 16

Commit

a701fc0

verified ·

1 Parent(s): dc307e1

Update format note: working GGUF now available

Browse files

Files changed (1) hide show

README.md +5 -6

README.md CHANGED Viewed

@@ -41,11 +41,10 @@ This is the exact artifact our serving stack runs.
 - **This repo:** the champion LoRA **fused into the 4-bit base**, in native MLX format (~18 GB). Runs
   directly with `mlx_lm` on Apple Silicon — no conversion, no dequantization.
-> **Format note (why MLX, not GGUF):** the model was trained and is served in MLX. A llama.cpp GGUF build
-> is in progress; `convert_hf_to_gguf` currently mistranslates the MLX tensor layout for this hybrid
-> GatedDeltaNet-MoE architecture (the bare-weights numerics are correct — verified — but the GGUF layer
-> ordering is not yet right). Until that is resolved, **this MLX repo is the canonical, verified-working
-> release.**
 ## Eval (real `tsc --strict`, frozen 24-task held-out benchmark)
 **Raw single-greedy + RAG** — honest, same-harness, multi-seed flat mean (never a cherry-picked run).
@@ -97,7 +96,7 @@ first that compiles under `tsc`.
 - Single-greedy compile rate is ~⅓–½ of hard tasks; pair with RAG + best-of-N + a `tsc` gate.
 - `effect@4.0.0-beta.80` only; later betas may shift APIs.
 - Reasoning/thinking is disabled (`enable_thinking=False`) — it's a direct code generator.
-- MLX format → Apple Silicon. A portable GGUF is in progress (see format note above).
 Supersedes `jrad123777/effect-qwen36-35b-v3-gguf` (an earlier, weaker SFT checkpoint).

 - **This repo:** the champion LoRA **fused into the 4-bit base**, in native MLX format (~18 GB). Runs
   directly with `mlx_lm` on Apple Silicon — no conversion, no dequantization.
+> **Format note (MLX vs GGUF):** the model was trained and is served in MLX, so this repo is the native,
+> canonical release. A portable **llama.cpp GGUF build is now also available** (Q4_K_M + Q8_0,
+> smoke-verified) at [`jrad123777/effect-qwen36-35b-gguf`](https://huggingface.co/jrad123777/effect-qwen36-35b-gguf)
+> — use that for CPU / non-Apple inference.
 ## Eval (real `tsc --strict`, frozen 24-task held-out benchmark)
 **Raw single-greedy + RAG** — honest, same-harness, multi-seed flat mean (never a cherry-picked run).
 - Single-greedy compile rate is ~⅓–½ of hard tasks; pair with RAG + best-of-N + a `tsc` gate.
 - `effect@4.0.0-beta.80` only; later betas may shift APIs.
 - Reasoning/thinking is disabled (`enable_thinking=False`) — it's a direct code generator.
+- MLX format → Apple Silicon. For CPU / non-Apple, use the portable GGUF (see format note above).
 Supersedes `jrad123777/effect-qwen36-35b-v3-gguf` (an earlier, weaker SFT checkpoint).