Instructions to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf
# Run inference directly in the terminal:
llama cli -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf
# Run inference directly in the terminal:
llama cli -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf
# Run inference directly in the terminal:
./llama-cli -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

Use Docker

docker model run hf.co/Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

LM Studio
Jan

vLLM

How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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": "Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

Ollama
How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf with Ollama:
```
ollama run hf.co/Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf
```

Unsloth Studio

How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf to start chatting

How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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": "Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 "Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-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 Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf with Docker Model Runner:
```
docker model run hf.co/Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf
```

Lemonade

How to use Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

Run and chat with the model

lemonade run user.DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf-{{QUANT_TAG}}

List all available models

lemonade list

DeepSeek-V4-Flash-0731 — 2-bit imatrix GGUF for DwarfStar

A 2-bit, importance-matrix-guided GGUF quantization of deepseek-ai/DeepSeek-V4-Flash-0731, built with the DwarfStar (ds4) toolchain.

Intended for 96 GB and 128 GB machines that cannot hold a 4-bit build.


File	`DeepSeek-V4-Flash-0731-IQ2XXS-w2Q2K-AProjQ8-SExpQ8-OutQ8-imatrix.gguf`
Size	86,720,111,520 bytes (80.8 GiB)
Resident model	80.76 GiB
Context overhead	~0.86 GiB at `--ctx 32768`
Architecture	43 layers, 256 routed experts (top-6), 1 shared expert

Quantization recipe

Tensor group	Type
Routed experts, gate (`w1`) / up (`w3`)	`IQ2_XXS`
Routed experts, down (`w2`)	`Q2_K`
Attention projections	`Q8_0`
Shared experts	`Q8_0`
Output	`Q8_0`
HC / compressor / indexer	`F16`

This matches the established DwarfStar Flash q2-imatrix recipe, applied to the 0731 checkpoint. The 0731 weights ship as fp4 routed experts (I8 packed E2M1 with F8_E8M0 block-32 scales) and fp8 dense tensors; deepseek4-quantize dequantizes both to f32 before requantizing.

Importance matrix

Activation statistics were collected with the ds4 runtime itself, hooking the layer-major Metal prefill graph, so the counts come from the real DS4 inference graph rather than a proxy implementation.

Dataset: gguf-tools/imatrix/dataset/rendered_prompts.txt (DS4-rendered chat prompts: source review, long context, agent/tool-call DSML, prose tasks, and benchmark prompts, in both thinking and non-thinking forms)
Budget: 2,720 prompts / 1,496,249 tokens / 386,032,242 routed-expert observations
Output: 129 entries (43 layers × 3 routed tensors), one packed vector per expert

For gate/up the collector records squared FFN-normalized input activation; for down, the squared route-weighted SwiGLU row.

Measured effect of the imatrix

Scored against 100 official DeepSeek V4 Flash continuations by target-token negative log likelihood, comparing this build to an otherwise identical build of the same checkpoint quantized without an imatrix:

Build	avg NLL
0731 Q2, no imatrix	0.579809
0731 Q2, imatrix (this file)	0.571552
Relative change	-1.42%

Per-case wins: 54 imatrix / 46 no-imatrix.

Both builds are the same checkpoint, so this isolates the imatrix contribution. Note that the reference continuations were collected from an earlier Flash checkpoint, which raises the absolute NLL for both builds equally; only the relative difference is meaningful here.

Usage

git clone https://github.com/antirez/ds4
cd ds4 && make
./ds4 -m DeepSeek-V4-Flash-0731-IQ2XXS-w2Q2K-AProjQ8-SExpQ8-OutQ8-imatrix.gguf \
      -p "Write a C function that reverses a string in place."

Or serve it:

./ds4-server -m DeepSeek-V4-Flash-0731-IQ2XXS-w2Q2K-AProjQ8-SExpQ8-OutQ8-imatrix.gguf --ctx 100000

Observed on an Apple M5 Max (128 GB, Metal): prefill 86.9 t/s, generation 29.6 t/s at short context. Throughput falls as context grows.

Caveats

2-bit routed experts are a real quality tradeoff. If you have the RAM for a 4-bit or mixed q2/q4 build, prefer it.
Only the routed expert tensors are 2-bit; attention, shared experts, and the output head stay at Q8_0, and the HC/compressor/indexer path stays at F16.
This file requires DwarfStar. It is not a llama.cpp-compatible GGUF: the DS4 architecture, compressed attention, and packed routed-expert layout are specific to the ds4 runtime.

License

MIT, inherited from the base model.

Downloads last month: -

GGUF

Model size

284B params

Architecture

deepseek4

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for Rednalreden/DeepSeek-V4-Flash-0731-dwarfstar-q2-gguf

Base model

deepseek-ai/DeepSeek-V4-Flash-0731

Quantized

(42)

this model