Hi3Dv1.0-mlx · FP16

← Project & inference runtime · Results

Recommended starting point. Original mixed-precision transformer checkpoints: FP16 torso with sensitive tensors retained in FP32. The runtime defaults to an FP32 VAE.

Selective download	Transformer(s) + shared VAE
512³	3.77 GiB
1024³ / 1536³	7.20 GiB

These are checkpoint sizes, not runtime memory. Encoders and Python dependencies are additional. The shared components are stored in Hi3Dv1.0-mlx-shared.

Run it

Follow the main installation guide, then:

hi3d-mlx refine --image reference.png --coarse-mesh coarse.glb \
  --preset fp16 --resolution 512 --output refined.glb

The CLI downloads only this variant, the shared VAE/configs and the pinned image encoders. No conversion step is required. For an offline setup, run hi3d-mlx download --variant fp16 --resolution 512 once, then add --local-files-only to inference. Generic mlx-lm loaders do not implement this geometry pipeline.

Evidence and limitations

See the benchmark methodology and complete recorded results. The historical cohort used Apple M5, 32 GiB, MLX 0.32.4. Results are geometry only and require both an image and an existing coarse mesh. The 1536³ grid reuses transformer_1024; it is not a fixed face count.

Format and attribution

Original mixed FP16/FP32 checkpoints reused by the MLX runtime; shared FP32 VAE. No duplicated FP32 or FP16 model copy.

Transformer checkpoint bytes are unchanged from the original Hi3D release. All applicable upstream notices and the Hi3D Open License Agreement v1.0 are retained. No blanket MIT or Apache license is asserted for the model.

All index paths are relative to the assembled model root. shared_components are resolved from the shared repository, not this transformer repository.