Downloads · 30 days
14
21% of all-time downloads
Qwen3-VL-2B-Instruct-GazeHeads-AllLayers-ONNX
AnonSubmi/Qwen3-VL-2B-Instruct-GazeHeads-AllLayers-ONNX
Qwen3-VL-2B-Instruct-GazeHeads-AllLayers-ONNX is a image-text-to-text model from AnonSubmi. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. It is set up for transformers.js. The card lists the license as apache-2.0.
Anonymous review artifact for WACV 2027 submission 759.
All-time downloads
68
Public
Repo size
1.2 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.onnx_data1.2 GB · 99%
At a glance
- Task
- Image-Text-to-Text
- Library
- transformers.js
- License
- apache-2.0
- Model type
- qwen3_vl
- Access
- Public
- Created
- Aug 20, 2026
- Updated
- Aug 20, 2026
- SHA
- 3e6a0d77
Try a prompt
Base models
Spaces
Model card
From the Hugging Face model README
Qwen3-VL-2B-Instruct — Gaze-Head Steering, All Layers (ONNX, WebGPU)
Anonymous review artifact for WACV 2027 submission 759.
ONNX export (q4f16) of Qwen3-VL-2B-Instruct for transformers.js/WebGPU whose decoder graph carries two extra inputs implementing the paper's attention-steering intervention:
gaze_head_maskof shape (28, 16): selects which attention heads to steer.gaze_signof shape (total_seq_len,): a per-KV-position additive pre-softmax attention bias (plus or minus delta) applied to the selected heads.
Every layer is decomposed, so all 448 heads are steerable; one model serves the demo's gaze, random-head, and all-head modes. The accompanying browser demo precomputes vision features offline, so the vision encoder file here is a stub and the vision tower never runs in the browser.
- Task
- Image-Text-to-Text
- Library
- transformers.js
- Type
- qwen3_vl
- License
- apache-2.0
- Created
- Aug 20, 2026
- Updated
- Aug 20, 2026