Add direct-manipulation UI: drag on the 3D head to drive the controls (#1)

Restyle Expression tab keyword-preset control to match custom UI aesthetic

Upload gnm/shape/data/semantic_sampler/expression_decoder_model.h5 with huggingface_hub

Upload gnm/shape/semantic_sampler.py with huggingface_hub

Upload README.md with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload app.py with huggingface_hub

Neck/Head side-by-side layout; raise csv field limit for Examples caching

Upload app.py with huggingface_hub

Tighten vertical slider layout: reduce slider gaps, eyes side-by-side, tighter title spacing

Replace native sliders with custom vertical sliders (side-by-side per group)

Clamp preset values to expression slider range [-3,3]

Add per-group Reset buttons and grouped vertical slider layout

Clamp preset values to expression slider range [-3,3]

Replace example presets: open-mouth+tongue, surprised, wink

Reduce viewer light intensity/exposure and simplify description text

Docs: reflect 'mouth' control label and in-place morphing viewer

Label lower-face expression group as 'mouth' (matches ipynb); replace gr.Model3D with persistent Three.js viewer that morphs vertices in place (no camera reset / no flicker)

Update Controls docs: precise naming + front-facing/persistent camera

Precise control labels, front-facing default camera, persist camera + no flicker on update

Upload folder using huggingface_hub

initial commit