How to use from
OpenClaw
Start the llama.cpp server
# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ITLL/LFM-3-Coder-Claude-Distill:Q4_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://127.0.0.1:9931/v1 \
  --custom-model-id "ITLL/LFM-3-Coder-Claude-Distill:Q4_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"
Quick Links

🚀 LFM 3 1.7B Claude Distill

An enhanced, lightweight code-reasoning model built on top of LiquidAI's LFM-3 1.7B. Fine-tuned to improve code performance and reasoning.


📌 Quick Summary

Feature Details
Developed by Schnuckade
Base Model LiquidAI/LFM3-1.7B
Fine-Tuning Framework Unsloth 🦥
Datasets Used Various datasets distilled from Claude
License Apache-2.0
Language English

💡 About the Model

LFM-3-Coder-Claude-Distill is a fine-tune designed to improve the model's reasoning and coding abilities. It was distilled on traces generated by Sonnet, Opus and Fable to better mimic their reasoning and answer style.

  • 🧠 Reasoning Focus: Helps the model think through problems step-by-step.
  • ⚡ Lightweight & Efficient: Reaches 200+ tokens/s on an RTX 4060.
  • 📦 Multi-Format: Includes full FP16/BF16 safetensors alongside quantized GGUF weights (Q4_K_M) for easy local deployment.

🛠️ Usage & Integration

Transformers / Unsloth

from unsloth import FastLanguageModel

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "Schnuckade/LFM-3-Coder-Claude-Distill",
    max_seq_length = 4096,
    load_in_4bit = True,
)
Downloads last month
-
Safetensors
Model size
2B params
Tensor type
F16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for ITLL/LFM-3-Coder-Claude-Distill

Quantized
(158)
this model

Datasets used to train ITLL/LFM-3-Coder-Claude-Distill

Collection including ITLL/LFM-3-Coder-Claude-Distill