Text-embedding encoder used by the Kimodo Animation Maker Draft Mode (ARDY, NVIDIA's realtime motion model): LLM2Vec MNTP adaptation of Meta-Llama-3-8B-Instruct with the McGill-NLP MNTP adapters merged into the base weights for direct local loading. The supervised LLM2Vec adapter (McGill-NLP/LLM2Vec-Meta-Llama-3-8B-Instruct-mntp-supervised, MIT) is applied on top at runtime. Produces one pooled 4096-dim embedding per prompt.

This repository exists so the Kimodo installer can fetch the encoder automatically. It is not a chat model.

Meta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright © Meta Platforms, Inc. All rights reserved. Use is subject to that license and Meta's Acceptable Use Policy. This model is a derivative of meta-llama/Meta-Llama-3-8B-Instruct; per the license, the model name begins with "Llama".

LLM2Vec adapters: © McGill NLP, MIT license.