gpu_tensor 1.7.0

A comprehensive GPU tensor library for Dart and Flutter, designed to leverage the power of webgpu across platforms for efficient tensor operations.

gpu_tensor

1.7.0

• Version bump to keep the minigpu 1.7.0 family aligned; inherits the persistent shader cache from minigpu_ffi 1.7.0. Every WGSL kernel this package compiles is now cached on disk and reused on the next process launch, so tensor kernels pay compilation once per machine rather than once per run. No API change in this package.

1.6.1

• released 08/13/26 - MR

1.6.0

• Version bump to keep the minigpu 1.6.0 family aligned; inherits the serialized, reference-counted native context from minigpu_ffi 1.6.0. No API change in this package.

1.5.9

• Version lockstep with minigpu 1.5.9; no API change.

1.5.8

1.5.7

1.5.6

1.5.5

1.5.4

• fix release version pins

1.5.2

• Add buffer copy

1.5.1

• fix frame wait and timeouts

1.5.0

• Fix handle issue, release 1.5.0

1.4.15

• fix fallback paths

1.4.14

• fix Tier C

1.4.12

• fix texture path

1.4.11

• fixing build hook

1.4.9

• tryfix central dawn location

1.4.8

• fixes texture path on windows, fixes texture view on web

1.4.7

• Fixes logger characters

1.4.6

• fix garbled adapter name in logs: WGPUStringView.data is not null-terminated, copy to std::string before passing to snprintf
• fix garbled adapter name in logs: WGPUStringView.data is not null-terminated, copy to std::string before passing to snprintf

1.4.5

• fix Dawn built inside pub-cache instead of system dir: pass DAWN_DIR from Dart hook as cmake -D define so cmake subprocess inherits the correct path regardless of env; fix FETCHCONTENT_BASE_DIR pointing to pub-cache (now uses cmake binary dir)
• add Minigpu.setLogCallback / setLogLevel: routes native Dawn/GPU log lines through a Dart callback (NativeCallable.listener); mgpuSetLogCallback + mgpuSetLogLevel exported from C layer; all stderr calls in minigpu_external.cpp replaced with structured LOG_ERROR/INFO/WARN/DEBUG macros

1.4.4

• fix FormatException on non-UTF-8 bytes in setLogCallback: use Utf8Decoder(allowMalformed: true) instead of toDartString()

1.4.3

• improve adapter selection to prefer discrete GPU using dawn native EnumerateAdapters; fixes incorrect adapter picked on Optimus laptops

1.4.2

• fixes dawn library not being found

1.4.1

• added bindings observer

1.4.0

• adds minigpu_view, gpu_pipeline libraries

1.3.0

• Adds Texture Sharing

1.2.4-WIP

• working on texture imports

1.2.3

• adds VRAM API
• fixes memleaks and broken tests

1.2.2

• fixes memleaks and broken tests

1.2.1

• Fix: fresh builds need dawn find off
• Change: 1.2.0 migrates minigpu to direct webgpu usage
• Breaking: Changed setData and references to .write
• Fix: broken compute shader and buffer finalizers

1.2.0

1.1.9

• fix pubspec version issue

1.1.8

• fixed concurrent buffer op crash

1.1.7

• fixed fft operations for 1D tensors
• fixed memory leaks

1.1.6

• fixed problem with audio input capture providing raw data

1.0.1

• Fix: pub.dev needs flutter dependency in pubspec.yaml

1.0.0

• Initial version includes:
  • gpu_tensor package with GPU tensor operations.
  • Basic tensor operations: addition, subtraction, multiplication, and division.
  • GPU accelerated FFT 1D and 2D operations.
  • Tensor data operations such as slice, reshape, getElement, head, and tail.
  • Unit tests for all tensor operations.
  • Example usage in the README file.