pixtral

Model type	pixtral
Class	LLaVAModel
Task	vision-language
Source	models/llava.py

Description

LLaVA vision-language model (3-model split).

Builds three separate ONNX models:

• decoder: text decoder taking inputs_embeds

• vision_encoder: CLIP/SigLIP + MLP projector

• embedding: token embedding + image feature fusion

Usage

mobius build --model <MODEL_ID> output_dir/

from mobius import build

model = build("<MODEL_ID>")