Today
-
Shorten AX650 note and recommend LavaSR
-
Add AX650 board comparison note
5 days ago
-
docs: 精简 badges/tags，移除 license 元数据
-
docs: badges + tags + highlights
-
Create config.json
-
README: cell spectrogram width 360
-
add per-audio spectrograms
-
README: spectrogram below each audio in the same cell
-
add speech 8k/16k split spectrograms
-
README: spectrograms inside audio table
-
add noisy samples + spectrograms
-
README: 精简结果页 + 带噪示例
-
重构试听页：8k/16k Open Speech + Samplelib，LSD/频谱图；删除 AudioSR/GT
-
add 16k reference audio (input/lavasr/gt)
-
README: add 16k reference audio section
-
add LavaSR FP32 outputs
-
README: FP32 comparison + vertical layout
-
README: absolute audio URLs + AudioSR-method LSD + 16k reference
-
add AudioSR 8k comparison audio
-
README: AudioSR 8k comparison
-
C++ RTF 0.061/0.039，精度不变
-
C++ RTF 0.061/0.039，精度不变
-
C++ RTF 0.061/0.039，精度不变
6 days ago
-
C++ OpenMP 优化：RTF 0.141 / 0.106
-
performance: clarify C++ full pipeline vs BWE-only
-
增加超分前后试听样例；C++ SDK 优化（RTF 0.238）
-
avoid __pycache__ on NFS
-
avoid __pycache__ on NFS
-
LavaSR v2 AX650: clean __pycache__
-
LavaSR v2 AX650: Pulsar2 8.0 量化模型 + Python/C++ SDK + 精度/性能报告
-
initial commit
