Instructions to use FermionResearch/Neutrino-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use FermionResearch/Neutrino-8B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="FermionResearch/Neutrino-8B") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("FermionResearch/Neutrino-8B", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use FermionResearch/Neutrino-8B 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 FermionResearch/Neutrino-8B # Run inference directly in the terminal: llama cli -hf FermionResearch/Neutrino-8B
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf FermionResearch/Neutrino-8B # Run inference directly in the terminal: llama cli -hf FermionResearch/Neutrino-8B
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 FermionResearch/Neutrino-8B # Run inference directly in the terminal: ./llama-cli -hf FermionResearch/Neutrino-8B
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 FermionResearch/Neutrino-8B # Run inference directly in the terminal: ./build/bin/llama-cli -hf FermionResearch/Neutrino-8B
Use Docker
docker model run hf.co/FermionResearch/Neutrino-8B
- LM Studio
- Jan
- vLLM
How to use FermionResearch/Neutrino-8B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "FermionResearch/Neutrino-8B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "FermionResearch/Neutrino-8B", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/FermionResearch/Neutrino-8B
- SGLang
How to use FermionResearch/Neutrino-8B 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 "FermionResearch/Neutrino-8B" \ --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": "FermionResearch/Neutrino-8B", "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 "FermionResearch/Neutrino-8B" \ --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": "FermionResearch/Neutrino-8B", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Ollama
How to use FermionResearch/Neutrino-8B with Ollama:
ollama run hf.co/FermionResearch/Neutrino-8B
- Unsloth Studio
How to use FermionResearch/Neutrino-8B 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 FermionResearch/Neutrino-8B 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 FermionResearch/Neutrino-8B to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for FermionResearch/Neutrino-8B to start chatting
- Pi
How to use FermionResearch/Neutrino-8B with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf FermionResearch/Neutrino-8B
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": "FermionResearch/Neutrino-8B" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Hermes Agent new
How to use FermionResearch/Neutrino-8B with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf FermionResearch/Neutrino-8B
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 FermionResearch/Neutrino-8B
Run Hermes
hermes
- Atomic Chat new
- OpenClaw new
How to use FermionResearch/Neutrino-8B with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf FermionResearch/Neutrino-8B
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 "FermionResearch/Neutrino-8B" \ --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 FermionResearch/Neutrino-8B with Docker Model Runner:
docker model run hf.co/FermionResearch/Neutrino-8B
- Lemonade
How to use FermionResearch/Neutrino-8B with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull FermionResearch/Neutrino-8B
Run and chat with the model
lemonade run user.Neutrino-8B-{{QUANT_TAG}}List all available models
lemonade list
| { | |
| "note": "MLX pack provenance: file -> bytes/sha256/source. Weights entry is declarative (not duplicated in this subfolder). fermion_mlx/ replaced 2026-07-31 by the v10 rebuild (adds in-process draft + batched-verify + MMA/SDPA paths).", | |
| "generated_utc": "2026-07-31T00:00:00Z", | |
| "weights_container": { | |
| "file": "neutrino-8b_v4.bin", | |
| "status": "GATED ON THE RELEASE CONTAINER: pathgate + refgate both PASS 2026-07-27 (fused-vs-mxops 32-token greedy IDENTICAL; TF rel max 0.000%, argmax 9/9, greedy 24/24, 0 flips) \u2014 receipts pathgate_8b_shipbrain.json / refgate_8b_shipbrain.json.", | |
| "gated_on_file": "neutrino-8b_v4.bin", | |
| "gated_on_bytes": 3875404812, | |
| "gated_on_sha256": "016c6f362ad237a8cf4043601efbc1e766574f76ea58691e78fdef51b4155fa0", | |
| "gated_on_source": "Fermion Research release archive" | |
| }, | |
| "files": { | |
| "README.md": { | |
| "bytes": 6000, | |
| "sha256": "3197e7892bb9d63339a03e4352aa1358b1b0ff8795876650a5b6a5a3752564a3", | |
| "source": "pack file" | |
| }, | |
| "fermion_mlx/MANIFEST.json": { | |
| "bytes": 1077, | |
| "sha256": "6f2f11bf17214c7f450f231b377a9b47c446904e86ea670cdce9b328cf472cb8", | |
| "source": "v10 pack inner manifest (file -> sha256 of this directory)" | |
| }, | |
| "fermion_mlx/__init__.py": { | |
| "bytes": 6626, | |
| "sha256": "999f003da4e49311413876df3b4011eae515fbf15684a96a4794af8f946d5185", | |
| "source": "v10 rebuild 2026-07-31: load()/generate()/chat() API" | |
| }, | |
| "fermion_mlx/__main__.py": { | |
| "bytes": 13847, | |
| "sha256": "8ba4104366a29737b5fecc042ecfac88eb0106122464af29c5822b2ee81ba7d9", | |
| "source": "v10 rebuild 2026-07-31: CLI (gen/chat/gate/cachegate/pathgate/bench + spec/specgate/specbench + draft modes)" | |
| }, | |
| "fermion_mlx/arch3.py": { | |
| "bytes": 8206, | |
| "sha256": "84a15da8566acdfbf6acef60e09099113db150c0f53deefb04491365ea7efef5", | |
| "source": "v10 rebuild 2026-07-31: fused per-head QK-RMSNorm+rope+KV-append kernel + int8 unembed" | |
| }, | |
| "fermion_mlx/container.py": { | |
| "bytes": 6059, | |
| "sha256": "c5b352184ba9ddf9a3e424731226cb7dcb9901d3a1516c67d4557953b0101f2f", | |
| "source": "TRTC v4 arch-3 reader (unchanged from 2026-07-25 pack)" | |
| }, | |
| "fermion_mlx/drafter.py": { | |
| "bytes": 6892, | |
| "sha256": "a0c5ace9e4dc94926f46e6495407e9463676d25345884a6eef626dcb92a9de14", | |
| "source": "NEW v10 2026-07-31: in-process draft walker for speculative decode" | |
| }, | |
| "fermion_mlx/flash_sdpa.py": { | |
| "bytes": 11367, | |
| "sha256": "6ce98b11049ee902db14259b3db320f7df32b9be4a1fb13d86a2a76577a30cb1", | |
| "source": "NEW v10 2026-07-31: fused SDPA path" | |
| }, | |
| "fermion_mlx/gates.py": { | |
| "bytes": 7411, | |
| "sha256": "c032e35944e4ec3c0b991e246114697a9538fa9bfcfff183d39b066d3d180349", | |
| "source": "v10 rebuild 2026-07-31: correctness gates" | |
| }, | |
| "fermion_mlx/heads.py": { | |
| "bytes": 9239, | |
| "sha256": "a9353d3e9bd93130a1533379986b11c7baefa0e11af555c36a94d9f9d0e97226", | |
| "source": "NEW v10 2026-07-31: fast-head margin path (rederived, no closed-stack code)" | |
| }, | |
| "fermion_mlx/mbatch.py": { | |
| "bytes": 20068, | |
| "sha256": "5736d2cf0ef3d7215b92beebc2c4dc3fbdc09f147effd5f7e3b8b696cdfb71a1", | |
| "source": "v10 rebuild 2026-07-31: batched-verify M-token dispatch" | |
| }, | |
| "fermion_mlx/mma.py": { | |
| "bytes": 14725, | |
| "sha256": "82905730ea79240d9647e63a8131c076781e72979efe2a05999c911206db8c96", | |
| "source": "NEW v10 2026-07-31: simdgroup-MMA batched matmul" | |
| }, | |
| "fermion_mlx/model.py": { | |
| "bytes": 26179, | |
| "sha256": "7d48f63b7043c8e8508a937f0d2e2e4a17a82c02800a082671a7dbae163d0d72", | |
| "source": "v10 rebuild 2026-07-31: model graph" | |
| }, | |
| "fermion_mlx/spec.py": { | |
| "bytes": 15875, | |
| "sha256": "cb92be27e752580f3e9852191d50dc6b421cd5e33d18b7f53b528d4a832c7895", | |
| "source": "v10 rebuild 2026-07-31: speculative decode driver" | |
| }, | |
| "fermion_mlx/trit_kernels.py": { | |
| "bytes": 44235, | |
| "sha256": "ba1c32e18585d8a1e6202c739ba6e4cbcdaefcc23f0dd7d665ffbfcfb7bbd1d4", | |
| "source": "v10 rebuild 2026-07-31: FV5 trit kernels" | |
| }, | |
| "merges.txt": { | |
| "bytes": 1671853, | |
| "sha256": "8831e4f1a044471340f7c0a83d7bd71306a5b867e95fd870f74d0c5308a904d5", | |
| "source": "donor Qwen/Qwen3-8B (same files as the parent tree root EXCEPT tokenizer_config.json (see R4_PARSER_AUDIT.md P4))" | |
| }, | |
| "receipts/specbench_8b_0p6b_m5.json": { | |
| "bytes": 826, | |
| "sha256": "c36ef09659d8c7690d53c23d065943103a95417890f8defbd01e629d7b5feeca", | |
| "source": "spec gate/bench receipt (2026-07-25)" | |
| }, | |
| "receipts/specgate_8b_0p6b.json": { | |
| "bytes": 705, | |
| "sha256": "64bcb90fe965835bc78887d873d3cfd0677b76c6d72cda04a84e295efee7c10c", | |
| "source": "spec gate/bench receipt (2026-07-25)" | |
| }, | |
| "requirements.txt": { | |
| "bytes": 191, | |
| "sha256": "1cf4d1fa581aaa5deae0568f88388c51166bbd69e7a89080b5a53488a7ef39b6", | |
| "source": "pack file" | |
| }, | |
| "tokenizer.json": { | |
| "bytes": 11422654, | |
| "sha256": "aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4", | |
| "source": "donor Qwen/Qwen3-8B (same files as the parent tree root EXCEPT tokenizer_config.json (see R4_PARSER_AUDIT.md P4))" | |
| }, | |
| "tokenizer_config.json": { | |
| "bytes": 9980, | |
| "sha256": "9f1e48d00384f25decc7f30dc7ec3a76fb3a98b37f5297b8494a134cef786102", | |
| "source": "donor Qwen/Qwen3-8B (same files as the parent tree root EXCEPT tokenizer_config.json (see R4_PARSER_AUDIT.md P4))" | |
| }, | |
| "vocab.json": { | |
| "bytes": 2776833, | |
| "sha256": "ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910", | |
| "source": "donor Qwen/Qwen3-8B (same files as the parent tree root EXCEPT tokenizer_config.json (see R4_PARSER_AUDIT.md P4))" | |
| }, | |
| "receipts/noflash_fp16_race_20260731.json": { | |
| "bytes": 1287, | |
| "sha256": "efd6ad55ee8d994ba509347542ad49c409d78372deb949dcd75f8fe8bac94a67", | |
| "source": "no-flash fallback race autopsy + ship constraint (2026-07-31)" | |
| } | |
| }, | |
| "addenda": [ | |
| "2026-07-25: EXPERIMENTAL --spec speculative decoding added (new modules mbatch.py/spec.py + CLI modes + README section). Exactness-gated: specgate 6/6 token-identical vs plain (receipts). Existing gate battery re-verified on the updated package." | |
| ] | |
| } |