Instructions to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-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 pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-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 pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-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 pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-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 pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M

Use Docker

docker model run hf.co/pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M

SGLang

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with Ollama:
```
ollama run hf.co/pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M
```
Unsloth Desktop

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with Pi:

Start the llama.cpp server

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

Run Pi

# Start Pi in your project directory:
pi

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

Lemonade

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen-35B-A3B-SignOfFour-Coder-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-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 pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-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 pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use pragmaticcs/Qwen-35B-A3B-SignOfFour-Coder-GGUF with OpenClaw:

Start the llama.cpp server

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

pragmaticcs commited on Sep 4

Commit

fdf2cc9

1 Parent(s): 7815aa8

Update README.md

Browse files

Files changed (1) hide show

README.md +27 -15

README.md CHANGED Viewed

@@ -87,43 +87,55 @@ A four-way MoE merge of the Qwen 35B-A3B architecture, fusing task vectors from
 ## Merge Methodology
-For each floating-point parameter, a task delta is computed per donor $k$:
-$$\Delta_k = D_k - W_0$$
-**DARE pruning.** A Bernoulli mask at retention density $p$ zeroes out low-magnitude updates; surviving values are rescaled by $p^{-1}$:
-$$\tilde{\Delta}_k = \frac{1}{p} \left(\Delta_k \odot M_k\right), \quad M_k \sim \text{Bernoulli}(p)$$
-**TIES sign election.** A consensus sign $\Gamma$ is computed via weighted vote across donors, and any donor update conflicting with it is dropped before averaging:
-$$\Gamma = \text{sgn}\left(\sum_{k=1}^K \alpha_k \tilde{\Delta}_k\right)$$
-$$\Delta_{\text{TIES}} = \frac{\sum_{k=1}^K \alpha_k \tilde{\Delta}_k \odot \mathbb{I}\left(\text{sgn}(\tilde{\Delta}_k) = \Gamma\right)}{\sum_{k=1}^K \alpha_k \cdot \mathbb{I}\left(\text{sgn}(\tilde{\Delta}_k) = \Gamma\right) + \epsilon}$$
 **Depth-scaled reconstruction.** The merged weight is reconstructed as:
-$$W_{\text{final}} = W_0 + \lambda(l) \cdot \Delta_{\text{TIES}}$$
-where the layer scaling factor $\lambda(l)$ across decoder layer index $l \in [0, 39]$ is defined as:
-$$\lambda(l) = \beta \cdot \left(0.5 + 0.5 \sin\left(\pi \frac{l}{39}\right)\right)$$
-This keeps input/output projections closer to the base and applies the strongest task transfer to middle layers ($l \in [12, 28]$).
 ---
 ## Layer-Stratified Policies
-| Parameter Group | Match Substring | Policy | Density ($p$) | Base Scale ($\beta$) |
 | :--- | :--- | :---: | :---: | :---: |
 | **Embeddings / LM head** | `embed_tokens`, `lm_head` | Linear | — | 1.00 |
 | **Norms / biases** | `norm`, `bias`, 1D tensors | Linear | — | 1.00 |
 | **DeltaNet recurrent state** | `a_log`, `dt_bias`, `conv1d` | Linear | — | 1.00 |
 | **MoE router gate** | `mlp.gate.weight`, `block_sparse_moe.gate` | Linear | — | 1.00 |
-| **MoE shared expert** | `shared_expert` | DARE-TIES | 0.70 | 0.60 |
-| **Attention projections** | `attn`, `rotary`, `in_proj`, `out_proj`, `x_proj` | DARE-TIES | 0.75 | 0.60 |
-| **Routed experts (×256)** | `experts`, `mlp` | DARE-TIES | 0.65 | 0.55 |
 - **Router protection:** Gate weights use linear interpolation (~57% base, ~43% donors) rather than DARE to avoid destabilizing expert routing.
 - **DeltaNet stability:** Recurrent state kernels are excluded from DARE to prevent divergence in the linear-attention state space.

 ## Merge Methodology
+For each floating-point parameter, a task delta is computed per donor \\(k\\):
+$$
+\Delta_k = D_k - W_0
+$$
+**DARE pruning.** A Bernoulli mask at retention density \\(p\\) zeroes out low-magnitude updates; surviving values are rescaled by \\(p^{-1}\\):
+$$
+\tilde{\Delta}_k = \frac{1}{p} \left(\Delta_k \odot M_k\right), \quad M_k \sim \text{Bernoulli}(p)
+$$
+**TIES sign election.** A consensus sign \\(\Gamma\\) is computed via weighted vote across donors, and any donor update conflicting with it is dropped before averaging:
+$$
+\Gamma = \operatorname{sgn}\left(\sum_{k=1}^K \alpha_k \tilde{\Delta}_k\right)
+$$
+$$
+\Delta_{\text{TIES}} = \frac{\sum_{k=1}^K \alpha_k \tilde{\Delta}_k \odot \mathbb{I}\left(\operatorname{sgn}(\tilde{\Delta}_k) = \Gamma\right)}{\sum_{k=1}^K \alpha_k \cdot \mathbb{I}\left(\operatorname{sgn}(\tilde{\Delta}_k) = \Gamma\right) + \epsilon}
+$$
 **Depth-scaled reconstruction.** The merged weight is reconstructed as:
+$$
+W_{\text{final}} = W_0 + \lambda(l) \cdot \Delta_{\text{TIES}}
+$$
+where the layer scaling factor \\(\lambda(l)\\) across decoder layer index \\(l \in [0, 39]\\) is defined as:
+$$
+\lambda(l) = \beta \cdot \left(0.5 + 0.5 \sin\left(\pi \frac{l}{39}\right)\right)
+$$
+This keeps input/output projections closer to the base and applies the strongest task transfer to middle layers \\((l \in [12, 28])\\).
 ---
 ## Layer-Stratified Policies
+| Parameter Group | Match Substring | Policy | Density (p) | Base Scale (β) |
 | :--- | :--- | :---: | :---: | :---: |
 | **Embeddings / LM head** | `embed_tokens`, `lm_head` | Linear | — | 1.00 |
 | **Norms / biases** | `norm`, `bias`, 1D tensors | Linear | — | 1.00 |
 | **DeltaNet recurrent state** | `a_log`, `dt_bias`, `conv1d` | Linear | — | 1.00 |
 | **MoE router gate** | `mlp.gate.weight`, `block_sparse_moe.gate` | Linear | — | 1.00 |
+| **MoE shared expert** | `shared_expert` | DARE&#8209;TIES | 0.70 | 0.60 |
+| **Attention projections** | `attn`, `rotary`, `in_proj`, `out_proj`, `x_proj` | DARE&#8209;TIES | 0.75 | 0.60 |
+| **Routed experts (×256)** | `experts`, `mlp` | DARE&#8209;TIES | 0.65 | 0.55 |
 - **Router protection:** Gate weights use linear interpolation (~57% base, ~43% donors) rather than DARE to avoid destabilizing expert routing.
 - **DeltaNet stability:** Recurrent state kernels are excluded from DARE to prevent divergence in the linear-attention state space.