Release 2.6.3

Summary

  • Binary Floyd–Steinberg dithering and repeated ThinPlateSpline crops take substantially less CPU time in the measured pipelines.
  • Sparse SaltAndPepper updates sampled positions directly. Existing mask-based replay records still work.
  • CPU NumPy-to-PyTorch conversion now uses Albucore’s shared bridge and copies read-only arrays before wrapping them as tensors.

Breaking changes

Upgrade workers that consume new sparse replay records

SaltAndPepper now records flat positions and a salt count in ReplayCompose when noise affects up to 16% of the spatial positions. Dense records still use two masks, and 2.6.3 accepts mask-based records from 2.6.2. Upgrade replay consumers to 2.6.3 before they receive the new sparse records. Continue passing the whole record to A.ReplayCompose.replay(record, image=image). #660

Performance

Dithering, repeated geometry, and sparse noise

  • Dithering processes independent diagonals together for binary uint8 Floyd–Steinberg error diffusion. It keeps the original order of FP32 additions and uses the scalar loop on short diagonals. Binary random and ordered dithering also allocate fewer intermediate arrays. #646, #664
  • ThinPlateSpline reuses grid bases for the two most recent eligible map shapes and control-point grids. The cache retains up to 16 MiB per process. Each call still samples its own deformation. PixelSpread builds coordinate maps from one-dimensional axes and writes into the final arrays. #647, #651
  • SaltAndPepper applies sparse noise at the sampled positions without allocating and scanning two full masks. Dense noise still uses masks. The sampled positions and outputs matched bit for bit in the paired tests. #660

These are paired median times from the cited PRs on an Apple M4 Max with one CPU thread. Each row compares the implementation with its own PR baseline and covers the listed CPU workload.

WorkloadBeforeAfterEvidence
Resize RGB 480×640 to 224×224, then apply binary Floyd–Steinberg dithering29.64 ms/image2.96 ms/image#646
Crop RGB to 224×224, then apply warm-cache ThinPlateSpline with the default grid2.154 ms/image0.504 ms/image#647
Decoded RGB batch of 16 through 224×224 crop, sparse SaltAndPepper, ToTensorV2, and collation308.1 µs/image91.6 µs/image#660

On Linux x86-64, the Floyd–Steinberg PR measured 8.1–11.8× speedups across nine Compose cases. ThinPlateSpline benefits most when the same grids recur. The SaltAndPepper PR also timed the full path starting from JPEGs on Python 3.11. Median savings fell to 10.1% on the M4 Max and 6.1% on the Linux Intel Xeon runner when read and decode were included. Dithering results, ThinPlateSpline results, SaltAndPepper results

Faster CPU paths for conversion, color, and noise

  • ToTensorV2, ToTensor3D, and CPU Tensor fallback use the Albucore NumPy–Torch bridge. Terminal transforms produce contiguous tensors. Fallback can retain compatible strided views. #650
  • ChannelSwap reverses RGB channels with one OpenCV operation per supported image or batch, including FP16 inputs whose bit patterns are preserved. Identity RGB orders return an independent copy directly. #648, #649, #654
  • Equalize uses Albucore’s histogram kernels. Colorize uses a shared byte lookup for each output plane and reuses those planes across frames. #653, #655
  • GaussNoise samples directly into a float32 buffer while preserving tested seed streams. Shared uint8 RGB noise no longer needs an expanded field for each channel in GaussNoise and AdditiveNoise. RGBShift picks up Albucore’s saturated uint8 RGB addition path. #656, #657, #661
  • CLAHE equalizes the lightness plane in its work buffer. ColorJitter reuses the uint8 hue plane for lookup. ChromaticAberration and eligible multichannel JPEG ImageCompression assemble output with cv2.merge. #658, #659, #662
  • PlasmaBrightnessContrast and PlasmaShadow reuse Diamond–Square work buffers. Paired tests verified bitwise equality of generated maps and identical final RNG state. #663

Bug fixes

  • CPU NumPy-to-Tensor conversion copies read-only storage before wrapping it. This prevents a returned tensor from writing into its read-only source array. Compatible writable arrays can still share storage. #650

Dependencies

Commits

CommitPRDescription
beb6ab5#646Vectorize binary Floyd–Steinberg Dithering
062ed3b#647Cache ThinPlateSpline grid bases
dd12884#648Speed up RGB channel reversal in images
c7c1e50#649Reverse RGB channels in one batch operation
4d994b0#650Adopt the Albucore CPU NumPy–Torch bridge
9ede7cf#655Use StringZilla byte lookup for Colorize
6e9f133#654Copy RGB identity channel orders directly
855bd83#653Delegate Equalize histogram kernels to Albucore
97fc617#651Build PixelSpread maps without dense grids
d1b6d0d#656Fill Gaussian noise through a flat buffer
2fa2d18#657Apply shared uint8 noise through Albucore
bddb30d#658Reuse the CLAHE lightness plane
592c92e#659Merge aberration and JPEG channels with OpenCV
fdf3d43#660Apply sparse SaltAndPepper by index
8df83e8#661Pin Albucore to 0.3.9
b0664b3#662Update the HSV hue plane in place
1668e05#663Reuse Diamond–Square work buffers
2b760df#664Avoid binary dithering temporaries
e494aeb#665Bump to 2.6.3 and update the release inventory

Full changelog for 2.6.2 to 2.6.3