Instructions to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-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 cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-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 cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF
# Run inference directly in the terminal:
llama cli -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF
# Run inference directly in the terminal:
llama cli -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

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 cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF
# Run inference directly in the terminal:
./llama-cli -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

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 cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF
# Run inference directly in the terminal:
./build/bin/llama-cli -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

Use Docker

docker model run hf.co/cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

LM Studio
Jan

vLLM

How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-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": "cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

Ollama
How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with Ollama:
```
ollama run hf.co/cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF
```

Unsloth Studio

How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF to start chatting

How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-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 cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

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 cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

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 "cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF" \
  --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"

Docker Model Runner
How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with Docker Model Runner:
```
docker model run hf.co/cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF
```

Lemonade

How to use cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF

Run and chat with the model

lemonade run user.qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF-{{QUANT_TAG}}

List all available models

lemonade list

cudabenchmarktest commited on Apr 9

Commit

a023ed8

verified ·

1 Parent(s): d8da57a

Expand card: 22-scenario hard tool stress test + image handling section + honest HX8 knowledge regression + hf download/ollama create run instructions

Browse files

Files changed (1) hide show

README.md +103 -0

README.md CHANGED Viewed

@@ -65,6 +65,109 @@ GSM8K samples are drawn from `openai/gsm8k` (test split) with a fixed seed
 (`1337`). The **same** 100 questions are used for base
 and tuned so the delta is directly comparable.
 ## How this adapter was trained
 ```bash

 (`1337`). The **same** 100 questions are used for base
 and tuned so the delta is directly comparable.
+### Stress test — 22 hard tool-calling scenarios (deterministic)
+Decoding: `temperature=0`, `top_p=1.0`, `top_k=1`, `seed=42`.
+Endpoint: http://localhost:11435/v1/chat/completions (Open Agents gateway).
+Scores: **21 / 22** base, **20 / 22** tuned.
+| # | Test | Base | Tuned |
+|---|---|:---:|:---:|
+| H1 | many-tool distraction (15 tools, 1 right answer) | ✅ | ✅ |
+| HX1 | HARD: four-op precedence (12² − (5+3)) / 4 | ✅ | ✅ |
+| H2 | nested object argument (recipient={name,address}) | ✅ | ✅ |
+| HX2 | HARD: C→F conversion inside enum-constrained call | ✅ | ✅ |
+| H3 | array argument (list of user_ids) | ✅ | ✅ |
+| HX3 | HARD: Gauss formula vs brute force (sum 1..100) | ✅ | ✅ |
+| H4 | enum normalization (user says 'cooling', enum is 'cool') | ✅ | ✅ |
+| HX4 | HARD: nested-object + array (multi-attendee meeting) | ✅ | ✅ |
+| H5 | numeric extraction from words ('fifty-two') | ✅ | ✅ |
+| HX5 | HARD: world knowledge + calculator (c − v_sound) | ✅ | ✅ |
+| H6 | negative + expression (calculator preserves parens) | ✅ | ✅ |
+| HX6 | HARD: 5-element array from varied prose | ✅ | ✅ |
+| H7 | chained sequential (convert + compute) | ✅ | ✅ |
+| HX7 | HARD: schema min/max refusal or clamp (roll 5000 d20) | ❌ | ❌ |
+| H8 | same tool, THREE different args (air_quality x 3) | ✅ | ✅ |
+| HX8 | HARD: domain knowledge → 7 parallel calls (G7 capitals) | ✅ | ❌ |
+| H9 | tool non-existence (no tool fits — should NOT hallucinate) | ✅ | ✅ |
+| HX9 | HARD: compound convert + chained compute (tax on EUR) | ✅ | ✅ |
+| H10 | disambiguation (wiki vs news — user wants breaking news) | ✅ | ✅ |
+| HX10 | HARD: deep-nested email with enum priority + body synthesis | ✅ | ✅ |
+| H11 | enum strictness (user asks 'freezing' — not in enum) | ✅ | ✅ |
+| H12 | reasoning-before-call (multi-step arithmetic → single call) | ✅ | ✅ |
+**Honest notes on the failures:**
+- **HX7**: Both models ignored JSON Schema `maximum: 1000` on count. Shared weakness.
+- **HX8**: Base correctly listed Washington as a G7 capital. Tuned substituted Brussels (wrong) — knowledge regression after distillation on 500 samples.
+- **HX3**: Initial test had overly narrow regex; accepts any algebraic form of Gauss formula now.
+## Image handling — NOT SUPPORTED
+This distilled model does **not** support image inputs. Attempting to pass
+`images` to Ollama's `/api/chat` returns **HTTP 500**:
+```
+Failed to create new sequence: failed to process inputs: this model is missing data required for image input
+```
+### Why
+Distilled GGUF was converted from Qwen3_5ForCausalLM (text subnet only). Vision encoder weights are not present in the merged safetensors, so llama.cpp's convert_hf_to_gguf produced a text-only GGUF. Ollama correctly detects the missing image tensors and returns HTTP 500 with a clean error message.
+### Verified with a small probe (3 rendered test images)
+| Model | Image-probe score | HTTP statuses |
+|---|:---:|:---:|
+| `qwen3.5:9b` (base) | 0 / 3 | 200 (serves images) |
+| `qwen3.5-9b-qwen3.6-distilled:q4km` (tuned) | 0 / 3 | [500, 500, 500] (no vision) |
+The base model *serves* image requests (Ollama reports `vision` in its
+Capabilities) but its text-in-image OCR on the three probes was weak
+(e.g. "HELLO" → "HELO", "42" → "44", "BANANA" → "barna"). Do not assume
+the base model is a reliable OCR tool just because image requests return
+HTTP 200.
+### Remedy
+For image+text workloads, use the base `qwen3.5:9b` (which has the vision tower). A multimodal-preserving distillation would require loading the full `Qwen3_5ForConditionalGeneration` checkpoint and keeping the vision layers frozen during LoRA training — out of scope for this text-only run.
+## Running this model with Ollama
+**Recommended path — download the Modelfile and build locally.**
+Direct `ollama pull hf.co/...` will auto-derive a Modelfile from GGUF
+metadata and may lose tool support. Using the Modelfile in this repo is
+the reliable way to preserve `tools` + `thinking` capabilities:
+```bash
+# Install hf CLI if needed (part of huggingface_hub):  pip install -U huggingface_hub
+hf download cudabenchmarktest/qwen3.5-9b-qwen3.6-reasoning-distilled-GGUF     Modelfile qwen3.5-9b-qwen3.6-reasoning-distilled.q4km.gguf     --local-dir ./qwen-distilled
+cd qwen-distilled
+ollama create qwen3.5-9b-qwen3.6-distilled:q4km -f Modelfile
+ollama show    qwen3.5-9b-qwen3.6-distilled:q4km
+#   Capabilities should include: tools, thinking
+# Tool-calling smoke test
+curl -s http://localhost:11434/api/chat -d '{
+  "model": "qwen3.5-9b-qwen3.6-distilled:q4km",
+  "messages": [{"role": "user", "content": "What is the weather in Paris?"}],
+  "tools": [{"type":"function","function":{"name":"get_weather","description":"Get current weather","parameters":{"type":"object","properties":{"city":{"type":"string"}},"required":["city"]}}}],
+  "stream": false
+}' | jq .message.tool_calls
+```
+To use a different quant, download the matching `.gguf` and edit the
+Modelfile's `FROM` line (or use one of the comment-suggested filenames):
+- `qwen3.5-9b-qwen3.6-reasoning-distilled.q4km.gguf` (~5.6 GB) — recommended
+- `qwen3.5-9b-qwen3.6-reasoning-distilled.q80.gguf`  (~9.5 GB) — higher fidelity
+- `qwen3.5-9b-qwen3.6-reasoning-distilled.f16.gguf`  (~17.9 GB) — full precision
 ## How this adapter was trained
 ```bash