qwen3_5

Model type

qwen3_5

Class

Qwen35VL3ModelCausalLMModel

Task

hybrid-qwen-vl

Source

models/qwen35.py

Description

Qwen3.5-VL vision-language model (3-model split).

Builds three separate ONNX models for onnxruntime-genai:

  • decoder: text decoder taking inputs_embeds (interleaved MRoPE)

  • vision_encoder: packed-attention ViT outputting merged features

  • embedding: token embedding + image feature fusion

The vision encoder is identical to Qwen3-VL’s

class:

Qwen3VLVisionModel. The text decoder uses

class:

Qwen35TextModel (hybrid linear/full attention).

Usage

mobius build --model <MODEL_ID> output_dir/

from mobius import build

model = build("<MODEL_ID>")