index
:
TaSTT-Whisper.git
master
High-performance GPGPU inference of OpenAI's Whisper automatic speech recognition (ASR) model
summary
refs
log
tree
commit
diff
stats
log msg
author
committer
range
Commit message (
Collapse
)
Author
Age
...
*
Updated performance summary
Konstantin
2023-01-24
|
*
Minor, performance summary tool
Konstantin
2023-01-24
|
*
Performance results for version 1.5
Konstantin
2023-01-24
|
*
Release notes
Konstantin
2023-01-24
|
*
Version 1.5
Konstantin
2023-01-24
|
*
Performance tuning on AMD iGPU
Konstantin
2023-01-24
|
*
GPU performance, matrix multiplication shaders
Konstantin
2023-01-24
|
*
Minor, cleanup unused
Konstantin
2023-01-24
|
*
Comments
Konstantin
2023-01-24
|
*
GPU performance, mulMatTiled shader
Konstantin
2023-01-24
|
*
mulMatByRowTiled shader, further performance optimizations
Konstantin
2023-01-24
|
*
GPU performance, mulMatByRowTiled shader
Konstantin
2023-01-23
|
*
Minor, micro-optimization
Konstantin
2023-01-23
|
*
Performance improvement, no longer destroying temporary buffers in ↵
Konstantin
2023-01-23
|
|
|
|
`encode()` method
*
Improved VRAM memory management, both speed and memory usage
Konstantin
2023-01-23
|
*
Minor, performance and VRAM use
Konstantin
2023-01-23
|
*
Minor, micro-optimization
Konstantin
2023-01-23
|
*
Performance improvement, `softMax` shader
Konstantin
2023-01-23
|
*
Minor, profiler tags
Konstantin
2023-01-23
|
*
VAD CPU performance, slightly better code generation
Konstantin
2023-01-23
|
*
GPU performance, optimized away a few shader dispatches
Konstantin
2023-01-22
|
*
Experimental, alternative busy wait implementation
Konstantin
2023-01-21
|
|
|
|
|
|
Disabled with a `constexpr` flag because on a desktop with discrete GPU this slowed down by about 20%. But the CPU load is about zero. Need to test on iGPUs, thermal shenanigans might make a difference there.
*
Minor, CPU performance
Konstantin
2023-01-21
|
*
CPU performance, SSE vectorization for MEL spectrogram
Konstantin
2023-01-21
|
*
Readme
Konstantin
2023-01-21
|
*
Minor
Konstantin
2023-01-21
|
*
Version 1.4
Konstantin
2023-01-20
|
*
Updated a screenshot
Konstantin
2023-01-20
|
*
UX bugfix, weird output path when a copy-pasted input path ends with a space
Konstantin
2023-01-20
|
*
Minor
Konstantin
2023-01-20
|
*
Minor, UX enhancements
Konstantin
2023-01-20
|
*
UX improvement, populate output path
Konstantin
2023-01-20
|
*
Bugfix, C# language projection
Konstantin
2023-01-20
|
*
Minor, documentation
Konstantin
2023-01-20
|
*
A tool to summarize performance data into one table
Konstantin
2023-01-20
|
*
Readme
Konstantin
2023-01-20
|
*
BIOS setting, enabled XMP for the memory
Konstantin
2023-01-20
|
*
Performance data on Ryzen 7 5700G iGPU
Konstantin
2023-01-20
|
*
Minor, error handling
Konstantin
2023-01-20
|
*
Release automation script
Konstantin
2023-01-19
|
*
Version 1.3
Konstantin
2023-01-19
|
*
Workaround for the Microsoft’s bug in their MP3 decoder MFT
Konstantin
2023-01-19
|
*
Updated screenshots
Konstantin
2023-01-18
|
*
Version 1.2
Konstantin
2023-01-18
|
*
Minor, logging and UX
Konstantin
2023-01-18
|
*
GUI to force specific version of the compute shaders
Konstantin
2023-01-18
|
*
Optional startup flags to override performance-related defaults for the ↵
Konstantin
2023-01-18
|
|
|
|
compute shaders
*
Consistent cancellation API across the library: S_OK = continue, S_FALSE = stop
Konstantin
2023-01-18
|
*
UX bugfix, pending visual state of the two main dialogs
Konstantin
2023-01-18
|
*
Added GeForce 1650 results to performance measures
Konstantin
2023-01-17
|
[prev]
[next]