Instructions to use giannisan/Hy3-ds4-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use giannisan/Hy3-ds4-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="giannisan/Hy3-ds4-gguf",
	filename="Hy3-ds4-IQ2XXS-AttnQ8-fromBF16.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use giannisan/Hy3-ds4-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 giannisan/Hy3-ds4-gguf:BF16
# Run inference directly in the terminal:
llama cli -hf giannisan/Hy3-ds4-gguf:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf giannisan/Hy3-ds4-gguf:BF16
# Run inference directly in the terminal:
llama cli -hf giannisan/Hy3-ds4-gguf:BF16

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 giannisan/Hy3-ds4-gguf:BF16
# Run inference directly in the terminal:
./llama-cli -hf giannisan/Hy3-ds4-gguf:BF16

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 giannisan/Hy3-ds4-gguf:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf giannisan/Hy3-ds4-gguf:BF16

Use Docker

docker model run hf.co/giannisan/Hy3-ds4-gguf:BF16

LM Studio
Jan
Ollama
How to use giannisan/Hy3-ds4-gguf with Ollama:
```
ollama run hf.co/giannisan/Hy3-ds4-gguf:BF16
```

Unsloth Studio

How to use giannisan/Hy3-ds4-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 giannisan/Hy3-ds4-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 giannisan/Hy3-ds4-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for giannisan/Hy3-ds4-gguf to start chatting

How to use giannisan/Hy3-ds4-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf giannisan/Hy3-ds4-gguf:BF16

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": "giannisan/Hy3-ds4-gguf:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use giannisan/Hy3-ds4-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 giannisan/Hy3-ds4-gguf:BF16

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 giannisan/Hy3-ds4-gguf:BF16

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use giannisan/Hy3-ds4-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf giannisan/Hy3-ds4-gguf:BF16

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 "giannisan/Hy3-ds4-gguf:BF16" \
  --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 giannisan/Hy3-ds4-gguf with Docker Model Runner:
```
docker model run hf.co/giannisan/Hy3-ds4-gguf:BF16
```

Lemonade

How to use giannisan/Hy3-ds4-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull giannisan/Hy3-ds4-gguf:BF16

Run and chat with the model

lemonade run user.Hy3-ds4-gguf-BF16

List all available models

lemonade list

Hy3 (295B) GGUF for pulsar / ds4 / NeutronStar (SSD streaming, CUDA)

Mixed-precision GGUF of tencent/Hy3 (295B total / 21B active MoE, Apache 2.0) built for SSD-streaming inference engines: routed experts live on disk and stream per token, so the model runs on GPUs that cannot hold it.

Runs on pulsar (Rust + CUDA, recommended) and the NeutronStar hy3 branch (C, a CUDA port of antirez/ds4).

Measured decode, greedy, warm cache:

engine	hardware	tok/s
pulsar	RTX 5060 Ti 16GB + RTX 4060 Ti 16GB, Gen5 NVMe	7.2
pulsar	RTX 4060 Ti 16GB, Gen4 NVMe	2.6
NeutronStar/ds4	RTX 4060 Ti 16GB, Gen4 NVMe	0.6-1.8

Per token only 8 of 192 experts per layer are read (~3GB/token at this quant); attention, shared experts, and the router stay resident.

Files

file	provenance	recommendation
`Hy3-ds4-IQ2XXS-AttnQ8-fromBF16.gguf`	single quantization straight from the BF16 checkpoint	use this one
`Hy3-ds4-IQ2XXS-AttnQ8.gguf`	requantized from an IQ4 intermediate (see below)	kept for continuity

Both use the identical recipe and the same importance matrix; they differ only in what the quantizer saw as input.

fromBF16 (new): tencent/Hy3 BF16 (598GB) converted to a q8_0 intermediate with the AngelSlim llama.cpp patches (proper hy_v3 architecture support), then quantized with the imatrix in one step. q8_0 is effectively lossless as an intermediate, so the routed experts see exactly one lossy quantization.

original: built before the BF16 pipeline existed here, from the IQ4-UD edition of YanissAmz/Hy3-295B-A21B-GGUF. The tensors this recipe keeps at Q8_0 passed through essentially lossless, but the routed experts went through two lossy steps (IQ4_XS/IQ3_S to IQ2_XXS). At a 2-bit target the 2-bit noise dominates, so the gap is small, but the fromBF16 file removes it entirely.

Recipe

The layout targets a streaming expert cache: routed experts must be uniform fixed-size slabs, and everything that makes decisions stays high precision. Same design as antirez's GLM-5.2 ds4 build, including the MTP layer riding at Q2_K because importance matrices never cover the draft layer (imatrix generation runs normal forwards, which skip it).

Tensors	Type	Why
routed experts, layers 1-79 (gate/up/down)	IQ2_XXS (imatrix)	streamed from disk per token; uniform slabs
routed experts, layer 80 (MTP)	Q2_K	no imatrix coverage exists for the draft layer
attention q/k/v/output, all layers	Q8_0	resident, paid once
shared expert + dense layer 0 FFN	Q8_0	resident
nextn.eh_proj (MTP glue)	Q8_0	tiny, no imatrix coverage
token embeddings, output head	Q8_0	ds4 embed kernel contract
router (ffn_gate_inp), expert bias, all norms	F32	decision makers stay exact

imatrix: the 125-chunk general-purpose matrix published with the source repo of the original build, reused for the fromBF16 build.

Usage

pulsar (recommended)

git clone https://github.com/giannisanni/pulsar
cd pulsar
CXX=g++-12 cargo build --release -p engine

./target/release/pulsar-cli -m Hy3-ds4-IQ2XXS-AttnQ8-fromBF16.gguf \
    -p "The capital of France is" -n 64
# or interactive chat:      --chat
# or OpenAI-compatible API: cargo build --release -p serve

Zero-config multi-GPU: pulsar measures each card's PCIe bandwidth at startup, streams experts over the fastest link, and fills spare GPUs with resident hot experts. First run is cold; a .warm sidecar makes every later run start hot. Main knob: PULSAR_CACHE_GB (host expert cache, defaults to measured free RAM minus a reserve).

NeutronStar / ds4

git clone -b hy3 https://github.com/giannisanni/neutronstar
cd neutronstar && make ds4
./ds4 -m Hy3-ds4-IQ2XXS-AttnQ8-fromBF16.gguf --cuda --ssd-streaming \
  --ssd-streaming-cache-experts 64 --ctx 4096 --nothink

Useful knobs: DS4_CUDA_HOST_EXPERT_CACHE_GB=16 (host expert cache, the main speed lever) and DS4_CUDA_PARALLEL_FETCH_THREADS=16.

MTP note

blk.80 (the MTP draft layer) is present in both files. pulsar wires it behind PULSAR_MTP=1 (opt-in): drafts verify correctly at 44 percent acceptance, but measurements show speculative decoding only pays when the verify pass is fully cache-resident, which a 30GB-RAM box cannot deliver. Tencent's own AngelSlim deployment guidance reaches the same conclusion, enabling MTP only from 2x H20 (192GB) upward. Keep it off for speed; it is there for bigger boxes.