Instructions to use SEBK4C/LEANSTRAL-WHITEDWARF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SEBK4C/LEANSTRAL-WHITEDWARF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="SEBK4C/LEANSTRAL-WHITEDWARF",
	filename="Leanstral-WhiteDwarf-cmid-52G.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 SEBK4C/LEANSTRAL-WHITEDWARF 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 SEBK4C/LEANSTRAL-WHITEDWARF
# Run inference directly in the terminal:
llama cli -hf SEBK4C/LEANSTRAL-WHITEDWARF

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf SEBK4C/LEANSTRAL-WHITEDWARF
# Run inference directly in the terminal:
llama cli -hf SEBK4C/LEANSTRAL-WHITEDWARF

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 SEBK4C/LEANSTRAL-WHITEDWARF
# Run inference directly in the terminal:
./llama-cli -hf SEBK4C/LEANSTRAL-WHITEDWARF

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 SEBK4C/LEANSTRAL-WHITEDWARF
# Run inference directly in the terminal:
./build/bin/llama-cli -hf SEBK4C/LEANSTRAL-WHITEDWARF

Use Docker

docker model run hf.co/SEBK4C/LEANSTRAL-WHITEDWARF

LM Studio
Jan
Ollama
How to use SEBK4C/LEANSTRAL-WHITEDWARF with Ollama:
```
ollama run hf.co/SEBK4C/LEANSTRAL-WHITEDWARF
```

Unsloth Studio

How to use SEBK4C/LEANSTRAL-WHITEDWARF 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 SEBK4C/LEANSTRAL-WHITEDWARF 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 SEBK4C/LEANSTRAL-WHITEDWARF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for SEBK4C/LEANSTRAL-WHITEDWARF to start chatting

How to use SEBK4C/LEANSTRAL-WHITEDWARF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SEBK4C/LEANSTRAL-WHITEDWARF

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": "SEBK4C/LEANSTRAL-WHITEDWARF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use SEBK4C/LEANSTRAL-WHITEDWARF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SEBK4C/LEANSTRAL-WHITEDWARF

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 SEBK4C/LEANSTRAL-WHITEDWARF

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use SEBK4C/LEANSTRAL-WHITEDWARF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf SEBK4C/LEANSTRAL-WHITEDWARF

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 "SEBK4C/LEANSTRAL-WHITEDWARF" \
  --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 SEBK4C/LEANSTRAL-WHITEDWARF with Docker Model Runner:
```
docker model run hf.co/SEBK4C/LEANSTRAL-WHITEDWARF
```

Lemonade

How to use SEBK4C/LEANSTRAL-WHITEDWARF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull SEBK4C/LEANSTRAL-WHITEDWARF

Run and chat with the model

lemonade run user.LEANSTRAL-WHITEDWARF-{{QUANT_TAG}}

List all available models

lemonade list

LEANSTRAL-WHITEDWARF / README.md

SEBK4C

Upload README.md with huggingface_hub

4603806 verified 11 days ago

preview code

Raw

History Blame Contribute Delete

3.91 kB

	---
	license: apache-2.0
	base_model:
	- mistralai/Leanstral-1.5-119B-A6B
	base_model_relation: quantized
	tags:
	- gguf
	- llama-cpp
	- lean4
	- moe
	- theorem-proving
	- autoresearch
	- work-in-progress
	---

	# Leanstral-WhiteDwarf — seed-v3 (EARLY CHECKPOINT, run not complete)

	**Status: this is an early-stopped checkpoint of an in-progress autoresearch
	quantization run — the ratchet loop has NOT completed.** The trajectory is
	promising (see numbers), the eval harness is frozen, and the run is designed
	to be resumed. Steps to continue are at the bottom.

	47 GB GGUF of [mistralai/Leanstral-1.5-119B-A6B](https://huggingface.co/mistralai/Leanstral-1.5-119B-A6B)
	(119B MoE Lean 4 prover, 128 experts / 4 active, MLA) — **39.5% of the FP8
	weights, ~3.2 bits/weight effective**, targeted at 2×24 GB + 91 GB RAM boxes.

	## Measured quality (vs the same model's FP8 golden run)

	\| Metric \| FP8 golden \| this checkpoint \| retention \|
	\|---\|---\|---\|---\|
	\| miniF2F-test slice, compiler-verified (4 attempts) \| 38/120 \| 29/120 \| 76% \|
	\| PutnamBench slice \| 1/60 \| 0/60 (partial: 0/10 scored at stop) \| — \|
	\| Teacher-forced top-1 agreement (135×32 tok, Lean) \| 1.0 \| 0.789 \| — \|
	\| Repetition battery (5-turn proof convs) \| pass \| pass \| no degeneration \|
	\| Decode on 2×RTX 4090 + DDR5 (experts on CPU) \| — \| 33 t/s \| — \|

	Verification = Lean 4 compiler in loop (mathlib-pinned), statement anti-cheat,
	no sorry/admit/native_decide. Golden reference data:
	[SEBK4C/Leanstral-WhiteDwarf-golden](https://huggingface.co/datasets/SEBK4C/Leanstral-WhiteDwarf-golden).

	## Recipe (the interesting part)

	Routed experts: `ffn_gate_exps`/`ffn_up_exps` → IQ2_XXS everywhere (verified
	safe). `ffn_down_exps` → **Q6_K on layers {0–3, every 3rd, 30–35}, Q4_K
	elsewhere** — this promotion pattern is load-bearing: flattening it to ≤Q4_K
	produces hard NaN on real Lean input while remaining coherent on casual English
	(seven falsified recipe variants; test on your target domain, not chit-chat).
	Everything else (attention/MLA, router gates, shared experts, embeddings,
	output) Q8_0. imatrix: 1.3M tokens of Lean-domain text (mathlib/STP/LeanDojo),
	included as `imatrix-lean-v1.dat`.

	Lineage: the asymmetric experts-2bit/rest-Q8 split follows
	[antirez/ds4](https://github.com/antirez/ds4)'s DeepSeek recipe; built with
	[llama.cpp](https://github.com/ggml-org/llama.cpp) (quantized at master
	bec4772). The last-layers-only protection prior from ds4/DeepSeek does NOT
	transfer to this arch — early layers have the flattest expert routing and
	need the most down-projection precision.

	## Serving (llama.cpp ≥ bec4772 recommended)

	```bash
	llama-server -m Leanstral-WhiteDwarf-seed-v3.gguf \
	--host 127.0.0.1 --port 8080 -np 4 -c 131072 \
	-b 2048 -ub 512 -fa on --jinja \
	--chat-template-file chat_template.jinja \
	-ngl 999 --override-tensor 'ffn_._exps.=CPU'
	```

	- Keep expert tensors OFF CUDA (`--override-tensor 'ffn_._exps.=CPU'`):
	2-bit expert tensors on the CUDA backend produced NaN on the builds tested.
	Non-expert layers + KV on GPU, experts in RAM: ~33 t/s on 2×4090.
	- The GGUF embeds no chat template (tekken v15): use the bundled
	`chat_template.jinja` (from Leanstral-2603, community consensus for 1.5).
	- Sampling per model card: temperature 1.0, top_p 0.95.

	## Continuing the run

	The full harness (program spec, recipes, tiered evals with Lean
	compiler-in-loop, golden anchors, resume journals) lives in the project repo
	`SEBK4C/Leanstral-WhiteDwarf`. Remaining route: finish t2 noise replicas →
	`golden/noise.json` → gate check (anchor 0.2167, local verifier) → ratchet
	loop (first phase: promote `ffn_down_exps` floor Q4_K→Q5_K until the gate
	passes, then shrink from above — expected gate-passing size 51–55 GB, then
	descend). `notes/CONTINUE-RUN.md` has exact commands for both local
	(2×4090, ~2 days) and rented-GPU (1×H100-80G, ~hours) execution.