Instructions to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-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 AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL
# Run inference directly in the terminal:
llama cli -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL
# Run inference directly in the terminal:
llama cli -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

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 AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL
# Run inference directly in the terminal:
./llama-cli -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

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 AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL
# Run inference directly in the terminal:
./build/bin/llama-cli -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

Use Docker

docker model run hf.co/AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

LM Studio
Jan

vLLM

How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

Ollama
How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with Ollama:
```
ollama run hf.co/AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL
```
Unsloth Desktop

How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with Docker Model Runner:
```
docker model run hf.co/AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL
```

Lemonade

How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

Run and chat with the model

lemonade run user.Qwen3.8-Flash-Next-REAP-320-GGUF-UD-Q2_K_XL

List all available models

lemonade list

Hermes Agent

How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

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 AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL

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 "AnonimousA/Qwen3.8-Flash-Next-REAP-320-GGUF:UD-Q2_K_XL" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

bf16 and imatrix?

by minyor25 - opened 30 days ago

Discussion

minyor25

30 days ago

•

edited 30 days ago

Hello, thank you for your great work, Q3 quant works good for me in 48Gb of vram.
But I would love to play with it and create couple of alternative mixed precision quants myself, can you pls share 16bit model and imatrix?
Thanks

AnonimousA

Owner 29 days ago

Glad the Q3 is working on 48 GB.

There's no bf16 — it never existed here. The prune is a binary copy: surviving expert tensors get lifted straight out of Unsloth's already-quantized GGUF, the rest are dropped, zero requantization. No fp16 checkpoint is ever produced on my side. The imatrix I can't share either, it's built on private traffic.

But the thing you actually want is already in the repo: manifests/seleccion_mass_K320.json — the per-layer kept-expert list ({"<layer>": [ids...]}, stock ids 0-511). That's the entire prune. Apply it to any Unsloth tier and you get that tier at K=320, which is exactly how the Q2 and Q3 here were made: same manifest, different source quant. An IQ4 at K=320 costs you a file copy, not a quantization run.

For genuinely mixed per-tensor precision you need the real weights, and those are public: Qwen's safetensors -> convert_hf_to_gguf.py -> llama-quantize with your own imatrix and --tensor-type overrides, then apply the manifest. It's a binary copy, so either order composes.

Worth knowing before you start, since "imatrix" is doing two different jobs in this repo: mine only ranks experts for the REAP selection and never touches the bits — the bits are Unsloth's, untouched. So you don't want mine anyway, you want one calibrated on the traffic you serve. Recipe is in #5; changing only the corpus was worth +5.5 HumanEval points at identical everything else.

If you build something good, post it here.

minyor25

28 days ago

Thank you for the explanation, I am still very new to all this reap stuff :)
Will definitely try to play with the manifest.
Just curious, is it possible to similarly reap the unsloth's imatrix file?

minyor25

6 days ago

Hey, was experimenting with the reaping stuff, I think I managed to reproduce your pruning technique, created 2 scripts one for model and another for imatrix pruning.
https://github.com/minyor/ymq-compiler/blob/main/ymq_reap_gguf.py
https://github.com/minyor/ymq-compiler/blob/main/ymq_reap_imatrix_gguf.py

Reaped base Qwen3.8 Flash into a bf16, then quantized it with my YMQ Mixed Precision quantization script.
Expert Gradient: Q6_K ➔ IQ4_NL ➔ IQ3_XXS ➔ IQ2_XS (floor) · Armor: Q8_0 / Q5_K
The size turned out to be about ~2Gb smaller so I can comfortably run this in 48Gb of vram with --spec-draft-n-max 2 and mmproj
If you curios, you can try it here:
https://huggingface.co/zerodigest/Qwen3.8-Flash-Next-REAP-320-YMQ-GGUF/blob/main/Qwen3.8-Flash-Next-REAP-320-YMQ-M-TI.gguf

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