Instructions to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-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 MichaelAnthony/qwen-3.5-0.8b-Hedgehog-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 MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF: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 MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF: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 MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

Use Docker

docker model run hf.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF with Ollama:
```
ollama run hf.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M
```
Unsloth Desktop

How to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

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": "MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF with Docker Model Runner:
```
docker model run hf.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M
```

Lemonade

How to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.qwen-3.5-0.8b-Hedgehog-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-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 MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

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 MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M

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 "MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M" \
  --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"

MichaelAnthony commited on Aug 13

Commit

ebb7bb2

verified ·

1 Parent(s): 8d58e48

release-r3-complex-implicit

Browse files

Files changed (7) hide show

.gitattributes +1 -0
Modelfile +3 -7
README.md +43 -42
qwen-3.5-0.8b-Hedgehog.F16.gguf +1 -1
qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf +1 -1
qwen-3.5-0.8b-Hedgehog.Q6_K.gguf +1 -1
qwen-3.5-0.8b-Hedgehog.Q8_0.gguf +1 -1

.gitattributes CHANGED Viewed

@@ -33,6 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 qwen-3.5-0.8b-Hedgehog.F16.gguf filter=lfs diff=lfs merge=lfs -text
 qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
 qwen-3.5-0.8b-Hedgehog.Q6_K.gguf filter=lfs diff=lfs merge=lfs -text

 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text
 qwen-3.5-0.8b-Hedgehog.F16.gguf filter=lfs diff=lfs merge=lfs -text
 qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf filter=lfs diff=lfs merge=lfs -text
 qwen-3.5-0.8b-Hedgehog.Q6_K.gguf filter=lfs diff=lfs merge=lfs -text

Modelfile CHANGED Viewed

@@ -1,7 +1,3 @@
-FROM qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf
-TEMPLATE """<|im_start|>system
-{{ .System }}<|im_end|>
-<|im_start|>user
-{{ .Prompt }}<|im_end|>
-<|im_start|>assistant
-"""


1	+ FROM ./qwen-3.5-0.8b-Hedgehog.Q8_0.gguf
2	+ PARAMETER temperature 0
3	+

README.md CHANGED Viewed

@@ -1,42 +1,43 @@
----
-license: apache-2.0
-language:
-- en
-pipeline_tag: text-generation
-tags:
-- qwen
-- gguf
-- llama-cpp
-- ollama
-- hedgehog
-base_model: Qwen/Qwen3.5-0.8B
----
-# Qwen 3.5 0.8B Hedgehog (GGUF Quantizations)
-Official GGUF quantizations for **Qwen 3.5 0.8B Hedgehog** for fast local inference using `llama.cpp` and `Ollama`.
-## 📦 Quantizations Included
-- `qwen-3.5-0.8b-Hedgehog.F16.gguf` (1.41 GiB) - 16-bit Float precision
-- `qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf` (504.8 MiB) - 4-bit medium (Recommended)
-- `qwen-3.5-0.8b-Hedgehog.Q6_K.gguf` (600.6 MiB) - 6-bit quantization
-- `qwen-3.5-0.8b-Hedgehog.Q8_0.gguf` (774.2 MiB) - 8-bit quantization
----
-## 🚀 Quickstart with Ollama
-```bash
-# Pull & Run 4-bit quantization directly
-ollama run hf.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q4_K_M "Extract key findings from: ..."
-# 8-bit quantization
-ollama run hf.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:Q8_0
-# 16-bit Float precision
-ollama run hf.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-GGUF:F16
-```
-## 📦 Related Repositories
-- Main PyTorch Model: [`MichaelAnthony/qwen-3.5-0.8b-Hedgehog`](https://huggingface.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog)
-- MLX Apple Silicon Model: [`MichaelAnthony/qwen-3.5-0.8b-Hedgehog-MLX`](https://huggingface.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog-MLX)

+---
+license: apache-2.0
+library_name: gguf
+language:
+- en
+pipeline_tag: text-generation
+base_model: MichaelAnthony/qwen-3.5-0.8b-Hedgehog
+tags:
+- gguf
+- qwen3.5
+- rag
+- document-extraction
+- json
+- simplerag
+- hedgehog
+---
+# qwen-3.5-0.8b-Hedgehog-GGUF
+R3 GGUF releases of
+[`MichaelAnthony/qwen-3.5-0.8b-Hedgehog`](https://huggingface.co/MichaelAnthony/qwen-3.5-0.8b-Hedgehog),
+a Qwen 3.5 0.8B SimpleRAG document extractor.
+| Format | File | Size | SHA-256 |
+| --- | --- | ---: | --- |
+| F16 | `qwen-3.5-0.8b-Hedgehog.F16.gguf` | 1.41 GiB | `08371f4b71e8c60aba4bbcf0bd2f524cce770f1c4c5a92113d38fda333340c85` |
+| Q4_K_M | `qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf` | 504.8 MiB | `ebb46a0d8d3e9933d4426fdb379dcfc4a61e404166ae06144dd25e9ed518ac1e` |
+| Q6_K | `qwen-3.5-0.8b-Hedgehog.Q6_K.gguf` | 600.6 MiB | `28d8b1fdfecd84b907ef01f271426abf769f7aa2ea7a2c7f0d3d14953fb3092e` |
+| Q8_0 | `qwen-3.5-0.8b-Hedgehog.Q8_0.gguf` | 774.2 MiB | `352f51d967af35b6c0e48a6c68243261542c4fbcf8541920a868439964365502` |
+All files report `general.name=qwen-3.5-0.8b-Hedgehog`,
+`general.architecture=qwen35`, 24 transformer blocks, 320 tensors, and no
+next-token-prediction head.
+The R3 Q8_0 file received a BenchLocal DataExtract-15 final score of **72/100**
+under temperature-0 generation, compared with **41/100** for the stock Qwen3.5
+0.8B Q8_0 in the same run. This benchmark was performed after training and its
+scenarios were not used as training examples. Only Q8_0 was used for that
+comparison; validate other quantizations independently.
+For strict extraction, use temperature 0, disable thinking, provide only the
+retrieved evidence, request a concrete JSON contract, and validate the response.

qwen-3.5-0.8b-Hedgehog.F16.gguf CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:27bd38b2a53819892023694f72db08c04ef163e20e7446666dd1317021d689c4
 size 1516744064

 version https://git-lfs.github.com/spec/v1
+oid sha256:08371f4b71e8c60aba4bbcf0bd2f524cce770f1c4c5a92113d38fda333340c85
 size 1516744064

qwen-3.5-0.8b-Hedgehog.Q4_K_M.gguf CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7da1f067072cb7d4d3a75064ab6ecea1fa996305eee0022ad633c5686cf3b65e
 size 529296768

 version https://git-lfs.github.com/spec/v1
+oid sha256:ebb46a0d8d3e9933d4426fdb379dcfc4a61e404166ae06144dd25e9ed518ac1e
 size 529296768

qwen-3.5-0.8b-Hedgehog.Q6_K.gguf CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:954f7cc7f09b977f91c5cc8754cdbd30110578fc3d06ab82745259d1231afb27
 size 629743488

 version https://git-lfs.github.com/spec/v1
+oid sha256:28d8b1fdfecd84b907ef01f271426abf769f7aa2ea7a2c7f0d3d14953fb3092e
 size 629743488

qwen-3.5-0.8b-Hedgehog.Q8_0.gguf CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a7d3983e360d76e91019f232e3f89921f5874c93cda4f00d4a5c4438e24ae65
 size 811842944

 version https://git-lfs.github.com/spec/v1
+oid sha256:352f51d967af35b6c0e48a6c68243261542c4fbcf8541920a868439964365502
 size 811842944