poxularandCursor 4965f27ba4 Add full GPU pipeline with ComfyUI nodes and phase profiling. ...

Ports scene extract, face ONNX preprocess, and net reference resize to Torch/CUDA with minimal CPU sync, while keeping the CPU parity path unchanged.

Co-authored-by: Cursor <cursoragent@cursor.com>

2026-06-02 01:01:46 +02:00

..

__init__.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00

color_torch.py

Fix GPU color matmul dtype mismatch (float64 matrices).

2026-06-02 00:46:20 +02:00

develop_torch.py

Add force_color_look support to GPU develop stack.

2026-06-02 00:49:13 +02:00

device.py

Add Torch GPU pipeline for CUDA ComfyUI nodes.

2026-06-02 00:43:01 +02:00

face_preprocess_torch.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00

luminance_torch.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00

lut_torch.py

Add Torch GPU pipeline for CUDA ComfyUI nodes.

2026-06-02 00:43:01 +02:00

pipeline_full_gpu.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00

pipeline.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00

reference_prep_torch.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00

reference_torch.py

Add Torch GPU pipeline for CUDA ComfyUI nodes.

2026-06-02 00:43:01 +02:00

resize_torch.py

Speed up GPU path: fused one-shot, GPU net resize, vectorized luminance.

2026-06-02 00:53:08 +02:00

scene_extract_torch.py

Add full GPU pipeline with ComfyUI nodes and phase profiling.

2026-06-02 01:01:46 +02:00