LFM3-1.7B (Q4_K_M) β€” Mercan format

This repository repackages LiquidAI/LFM3-1.7B by Liquid AI as a single .mercan file for the Mercan runtime and the Mercan iOS app. It is not an official Liquid AI release and is not affiliated with or endorsed by Liquid AI.

What was changed

model.mercan is the official LFM3-1.7B-Q4_K_M.gguf from Liquid AI with only metadata added. The tensors and all original metadata are copied byte-for-byte; no weights were modified, re-quantized, or fine-tuned.

Added keys:

Key Value
mercan.format mercan
mercan.format_version 1
mercan.runtime_abi 1
mercan.model_family LFM3
mercan.tokenizer.type gguf-vocab
mercan.source_kind gguf
mercan.source_checkpoint LFM3-1.7B-Q4_K_M.gguf
mercan.source_sha256 3793e1e6e2266919d6e015287221b41100175a98eddaba92e111aed3b05c93c4

The conversion tool is scripts/gguf_to_mercan.py; model.mercan.json lists the source and output SHA-256 values.

Usage

Requires a Mercan runtime with the lfm3 architecture provider (branch ios-lfm or later).

mercan run MercanAI/LFM3-1.7B-mercan -p "What is the capital of France?"

In the Mercan iOS app, paste this URL into the model field:

https://huggingface.co/MercanAI/LFM3-1.7B-mercan/resolve/main/model.mercan

LFM3 is a reasoning model. The runtime opens a <think> block and hides the reasoning, so answers can take a while on longer questions.

License

The model is licensed by Liquid AI under the LFM Open License v1.0 (see LICENSE). Among other terms, commercial use is not licensed to entities with annual revenue of USD 10,000,000 or more. Please read the full license before use.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for MercanAI/LFM3-1.7B-mercan

Quantized
(114)
this model