Instructions to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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": "DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M

Ollama
How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF with Ollama:
```
ollama run hf.co/DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M
```

Unsloth Studio

How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF to start chatting

How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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": "DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF with Docker Model Runner:
```
docker model run hf.co/DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M
```

Lemonade

How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen.Qwen-AgentWorld-35B-A3B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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 "DevQuasar/Qwen.Qwen-AgentWorld-35B-A3B-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"

llama.cpp crash: missing tensor 'blk.40.attn_norm.weight' (Incorrect GGUF Metadata)

by freykun - opened Jun 24

Discussion

freykun

Jun 24

Hi, thanks for providing these quants!

I ran into an issue where loading the GGUF files from this repository in llama.cpp (and standard backends like Ollama/LM Studio) causes a crash on load:

E llama_model_load: error loading model: missing tensor 'blk.40.attn_norm.weight'
The Cause
The quantization script used to generate these files seems to have exported the incorrect metadata for the layer count. The model metadata claims qwen35moe.block_count: 41 and qwen35moe.nextn_predict_layers: 1 (presumably due to the MTP layers). However, the actual tensors for layer 41 (blk.40) and the MTP projection (blk.39.nextn.eh_proj.weight) were not included in the GGUF file. When llama.cpp tries to load them, it hits a missing tensor error and crashes.

Temporary Workaround for Users
For anyone else running into this, you can bypass the crash by forcing llama.cpp to ignore the broken metadata and expect the standard 40 blocks with no MTP layers using the --override-kv flag:

bash
llama-server -m Qwen.Qwen-AgentWorld-35B-A3B.f16.gguf.Q4_K_M.gguf --override-kv "qwen35moe.block_count=int:40,qwen35moe.nextn_predict_layers=int:0"

To the uploader: A re-quantization with the latest convert_hf_to_gguf.py script that either correctly exports the MTP tensors, or correctly strips them out and sets the metadata to 40 blocks, should permanently fix this for all users.

csabakecskemeti

DevQuasar org Jun 25

@freykun
GGUFs have been patched -> you don't need that hack

csabakecskemeti

DevQuasar org Jun 25

final fix

csabakecskemeti changed discussion status to closed Jun 25

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