Safetensors
GGUF
English
Russian
lfm3
reasoning
test-time-compute
inference-time-scaling
code-generation
llama.cpp
negative-results
conversational
Instructions to use HelioAI/LFM3-1.7B-inference-modules-experiments with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use HelioAI/LFM3-1.7B-inference-modules-experiments 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 HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M # Run inference directly in the terminal: llama cli -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M # Run inference directly in the terminal: llama cli -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M # Run inference directly in the terminal: ./build/bin/llama cli -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
Use Docker
docker model run hf.co/HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
- Llama new
- LM Studio
- Jan
- Ollama
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Ollama:
ollama run hf.co/HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
- Unsloth Desktop
- Pi
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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": "HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Docker Model Runner:
docker model run hf.co/HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
- Lemonade
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
Run and chat with the model
lemonade run user.LFM3-1.7B-inference-modules-experiments-Q5_K_M
List all available models
lemonade list
- Hermes Agent
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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 HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_K_M
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use HelioAI/LFM3-1.7B-inference-modules-experiments with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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 "HelioAI/LFM3-1.7B-inference-modules-experiments:Q5_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"
Download research/scripts/make_fresh_tasks.py from HelioAI/LFM3-1.7B-inference-modules-experiments: direct link, hf CLI and curl.
- Browser
- Download file 4.25 kB
-
https://huggingface.co/HelioAI/LFM3-1.7B-inference-modules-experiments/resolve/main/research/scripts/make_fresh_tasks.py
- Command line
-
hf download hf://HelioAI/LFM3-1.7B-inference-modules-experiments/research/scripts/make_fresh_tasks.py
-
curl -L -o make_fresh_tasks.py https://huggingface.co/HelioAI/LFM3-1.7B-inference-modules-experiments/resolve/main/research/scripts/make_fresh_tasks.py
4.25 kB
| """Fresh held-out visual tasks (4 per category), written and frozen BEFORE the mentor training and never used for the | |
| curriculum or the advice. The curriculum was designed after seeing the student's weak skills on the 32 benchmark | |
| tasks; this set checks that what it learned carries over to subjects nobody looked at. | |
| """ | |
| import json | |
| from make_mentor_tasks import HTML_TAIL, LIB_NONE, LIB_THREE, SVG_TAIL # noqa: E402 (also rewrites mentor_tasks.jsonl) | |
| SVG = [ | |
| "Draw a cartoon penguin standing on ice: a black back and head, a white belly, an orange beak and feet, two flippers and two small eyes.", | |
| "Draw a radial menu: a ring divided into six equal segments with a label in each (Home, Mail, Music, Photos, Maps, Settings) and a round 'Menu' button in the centre.", | |
| "Draw a pyramid diagram: a triangle split into four horizontal layers labelled from bottom to top Basics, Practice, Projects, Mastery, each layer in a different colour.", | |
| "Draw a side view of a steam locomotive on rails: a long boiler, a cab with a window, a chimney with a puff of smoke, and three wheels joined by a connecting rod.", | |
| ] | |
| PAGES = [ | |
| "Build a 'Contact us' page with two columns: a contact form (name, email, subject, message, send button) on the left and the office address, phone, email and opening hours on the right.", | |
| "Build a social media profile header: a cover image placeholder (a gradient), a round avatar overlapping its bottom edge, the name, a short bio, follower and following counts and a Follow button.", | |
| "Build an invoice page: a company header, a bill-to address, a table of four line items with quantity, unit price and amount, the subtotal, tax and total, and payment terms at the bottom.", | |
| "Build a weekly habit tracker: a table of five habits by seven days with cells that are checked by clicking, and a completion percentage for each habit that updates.", | |
| ] | |
| CANVAS = [ | |
| ("Make a constellation animation on a canvas: 80 slowly drifting points, joined by lines when they are closer than 120 pixels, with the line opacity fading with distance.", True), | |
| ("Make an aquarium animation on a canvas: fish of different colours swimming left and right at different speeds and depths, rising bubbles and swaying seaweed.", True), | |
| ("Make a tic-tac-toe game on a canvas: clicking places X and O in turn, a win or a draw is detected and shown, and a restart button starts a new game.", False), | |
| ("Make a kaleidoscope animation on a canvas: a pattern of shapes mirrored into 12 symmetric segments around the centre, slowly rotating and changing colour.", True), | |
| ] | |
| THREE = [ | |
| ("Make a three.js carousel: a round platform with a cone roof on a central pole and six horses represented by colourful boxes that move up and down while the carousel rotates.", True), | |
| ("Make a three.js bowling scene: a wooden lane, ten white pins arranged in a triangle and a ball rolling towards them.", True), | |
| ("Make a three.js lava lamp: a transparent glass capsule on a dark base with several orange blobs (spheres) slowly rising and sinking inside it.", True), | |
| ("Make a three.js Ferris wheel: a large ring on two supports with eight gondolas (boxes) hanging from it that stay upright while the wheel turns.", True), | |
| ] | |
| tasks = [{"id": f"fresh/s{i}", "cat": "svg", "kind": "svg", "animated": False, "prompt": p + SVG_TAIL} | |
| for i, p in enumerate(SVG, 1)] | |
| tasks += [{"id": f"fresh/p{i}", "cat": "html", "kind": "html", "animated": False, | |
| "prompt": p + HTML_TAIL.format(lib=LIB_NONE)} for i, p in enumerate(PAGES, 1)] | |
| tasks += [{"id": f"fresh/a{i}", "cat": "canvas", "kind": "html", "animated": an, | |
| "prompt": p + HTML_TAIL.format(lib=LIB_NONE)} for i, (p, an) in enumerate(CANVAS, 1)] | |
| tasks += [{"id": f"fresh/d{i}", "cat": "three", "kind": "html", "animated": an, | |
| "prompt": p + HTML_TAIL.format(lib=LIB_THREE)} for i, (p, an) in enumerate(THREE, 1)] | |
| with open("data/visual/tasks_fresh.jsonl", "w", encoding="utf-8") as f: | |
| for t in tasks: | |
| f.write(json.dumps({"id": t["id"], "task": "visual", **{k: v for k, v in t.items() if k != "id"}}, | |
| ensure_ascii=False) + "\n") | |
| print(len(tasks), "fresh held-out tasks -> data/visual/tasks_fresh.jsonl") | |