You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
C++/GGML based Fish Audio S2 Pro in audio.cpp. 3x faster than real time #1308
I have thoroughly reviewed the project documentation (installation, training, inference) but couldn't find any relevant information that meets my needs. English中文日本語Portuguese (Brazil)
I have searched for existing issues search for existing issues, including closed ones.
I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
[FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:)
Please do not modify this template :) and fill in all the required fields.
1. Is this request related to a challenge you're experiencing? Tell us your story.
Measured on RTX 5090:
Warmed requests run about 3.1x-3.4x faster than real time. Longform (6000+ chars) runs about 3.3x faster than real time.
changed the title [-]C++/GGML based Fish Audio S2 Pro in audio.cpp[/-][+]C++/GGML based Fish Audio S2 Pro in audio.cpp. 3x faster than real time[/+]on Jul 23, 2026
I did some follow-up AMD/HIP profiling and optimization work on the audio.cpp Fish S2 Pro path using a Radeon RX 7900 XTX with ROCm 7.2.1. With identical benchmark input/settings, the baseline warm mean was 4.291 s (RTF 0.97), GPU-side sampling reduced it to 3.667 s (RTF 0.83, ~14.5% wall-time reduction), and an experimental chained Fast-AR path reached 3.557 s (RTF ~0.81, ~17.1%). The compared WAV output was bit-identical for the documented test case. Patch set, benchmark details, and implementation notes: https://github.com/reezex0-ux/fish-s2-amd-hip . I also opened an audio.cpp discussion for upstreaming the useful pieces against current main: 0xShug0/audio.cpp#317 . So far this is only benchmarked on RX 7900 XTX; results from other AMD hardware would be useful.
Activity
[-]C++/GGML based Fish Audio S2 Pro in audio.cpp[/-][+]C++/GGML based Fish Audio S2 Pro in audio.cpp. 3x faster than real time[/+]github-actions commentedon Aug 23, 2026
This issue is stale because it has been open for 30 days with no activity.
reezex0-ux commentedon Aug 27, 2026
I did some follow-up AMD/HIP profiling and optimization work on the audio.cpp Fish S2 Pro path using a Radeon RX 7900 XTX with ROCm 7.2.1. With identical benchmark input/settings, the baseline warm mean was 4.291 s (RTF 0.97), GPU-side sampling reduced it to 3.667 s (RTF 0.83, ~14.5% wall-time reduction), and an experimental chained Fast-AR path reached 3.557 s (RTF ~0.81, ~17.1%). The compared WAV output was bit-identical for the documented test case. Patch set, benchmark details, and implementation notes: https://github.com/reezex0-ux/fish-s2-amd-hip . I also opened an audio.cpp discussion for upstreaming the useful pieces against current main: 0xShug0/audio.cpp#317 . So far this is only benchmarked on RX 7900 XTX; results from other AMD hardware would be useful.
github-actions commentedon Sep 28, 2026
This issue is stale because it has been open for 30 days with no activity.
github-actions commentedon Oct 12, 2026
This issue was closed because it has been inactive for 14 days since being marked as stale.