Clappy for MLOArchitect (Experimental)

Clappy is the optional local helper inside MLOArchitect, the FiveM MLO builder by AJTheDev. He answers questions about the app and makes small changes to your build when you ask, running on your own graphics card. Nothing is sent anywhere.

Install it from inside MLOArchitect 1.4 or newer: Setup > Optional downloads > Clappy. MLOArchitect downloads clappy-1.0.0.zip from here, checks its size and SHA-256, and unpacks it. You don't need to download anything from this page by hand.

File	What it is
`clappy-1.0.0.zip` (2.1 GB)	The pack: llama.cpp b11972 (Windows x64, Vulkan), LFM3-Audio-2B (Q4_K_M) with its audio projector for the voice, Clappy's LoRA, EmbeddingGemma 2 (Q8_0) for search, and `clappy.json`
`THIRD-PARTY-LICENCES.txt`	Every licence in the pack, in full

SHA-256 of clappy-1.0.0.zip: adc3bef25b4f08a0811c749f56d118a81ed3b89e3142d1cff72000bbf3ec0367

Licences

LFM3-Audio-2B by Liquid AI, and Clappy's LoRA trained on it by AJTheDev: LFM Open License v1.0. The model was quantised to Q4_K_M by AJTheDev.
EmbeddingGemma 2 by Google DeepMind: Apache License 2.0. Converted to GGUF (text model only) and quantised to Q8_0 by AJTheDev. Use must follow the Gemma Prohibited Use Policy.
llama.cpp by the ggml authors: MIT. Unmodified release build.

Full texts: THIRD-PARTY-LICENCES.txt (and inside the zip).

Limits

Clappy is Experimental: a 2-billion-parameter model fine-tuned on a few thousand examples. He answers in English from MLOArchitect's own knowledge base, says so when something isn't in it, and can be wrong. Every change he makes to a build is one undo step, and he asks before removing anything.

Support: the AJTheDev Discord.

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. 🙋 Ask for provider support

Model tree for AlexJamesDean/MLOArchitect-Clappy

Base model

LiquidAI/LFM3-Audio-2B

Adapter

(1)

this model