tckm.yan/aha
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix FireRedVad detect_frame bug

Browse Source
This commit is contained in:
jhqxxx
2026-04-20 15:15:13 +08:00
parent deb8027a82
commit 0017bcd924
4 changed files with 33 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_fire_red_vad.rs
+1 -1
View File
@@ -23,7 +23,7 @@ fn stream_vad() -> Result<()> {
let device = aha::Device::Cpu;
let save_dir =
aha::utils::get_default_save_dir().ok_or(anyhow::anyhow!("Failed to get save dir"))?;
let model_path = format!("{}/xukaituo/FireRedVAD/Stream-VAD/", save_dir);
let model_path = format!("{}/jiangjiangaha/FireRedVAD-Stream-VAD/", save_dir);
let vad = FireRedVad::init(&model_path, Some(&device), None)?;
let res = vad.detect_file(audio_path)?;
tests/test_voxcpm.rs
+26 -25
View File
@@ -60,7 +60,7 @@ fn voxcpm_use_message_generate() -> Result<()> {
#[test]
fn voxcpm_generate() -> Result<()> {
// RUST_BACKTRACE=1 cargo test -F cuda voxcpm_generate -r -- --nocapture
// RUST_BACKTRACE=1 cargo test -F cuda --test test_voxcpm voxcpm_generate -r -- --nocapture
let save_dir =
aha::utils::get_default_save_dir().ok_or(anyhow::anyhow!("Failed to get save dir"))?;
let model_path = format!("{}/OpenBMB/VoxCPM-0.5B/", save_dir);
@@ -70,38 +70,39 @@ fn voxcpm_generate() -> Result<()> {
let i_duration = i_start.elapsed();
println!("Time elapsed in load model is: {:?}", i_duration);
let i_start = Instant::now();
// let i_start = Instant::now();
// let generate = voxcpm_generate.generate_simple("太阳当空照，花儿对我笑，小鸟说早早早".to_string())?;
let generate = voxcpm_generate.inference(
"VoxCPM is an innovative end-to-end TTS model from ModelBest, designed to generate highly realistic speech.".to_string(),
Some("啥子小师叔，打狗还要看主人，你再要继续，我，就是你的对手".to_string()),
Some("file://./assets/audio/voice_01.wav".to_string()),
// Some("一定被灰太狼给吃了，我已经为他准备好了花圈了".to_string()),
// Some("file://./assets/audio/voice_05.wav".to_string()),
2,
100,
10,
2.0,
// false,
6.0,
)?;
// 创建prompt_cache
// let _ = voxcpm_generate.build_prompt_cache(
// "啥子小师叔，打狗还要看主人，你再要继续，我，就是你的对手".to_string(),
// "file://./assets/audio/voice_01.wav".to_string(),
// )?;
// // 使用prompt_cache生成语音
// let generate = voxcpm_generate.generate_use_prompt_cache(
// "太阳当空照，花儿对我笑，小鸟说早早早".to_string(),
// let generate = voxcpm_generate.inference(
// "VoxCPM is an innovative end-to-end TTS model from ModelBest, designed to generate highly realistic speech.".to_string(),
// Some("啥子小师叔，打狗还要看主人，你再要继续，我，就是你的对手".to_string()),
// Some("file://./assets/audio/voice_01.wav".to_string()),
// // Some("一定被灰太狼给吃了，我已经为他准备好了花圈了".to_string()),
// // Some("file://./assets/audio/voice_05.wav".to_string()),
// 2,
// 100,
// 10,
// 2.0,
// false,
// // false,
// 6.0,
// )?;
// 创建prompt_cache
let _ = voxcpm_generate.build_prompt_cache(
"啥子小师叔，打狗还要看主人，你再要继续，我，就是你的对手".to_string(),
"file://./assets/audio/voice_01.wav".to_string(),
)?;
// 使用prompt_cache生成语音
let i_start = Instant::now();
let generate = voxcpm_generate.generate_use_prompt_cache(
"太阳当空照，花儿对我笑，小鸟说早早早".to_string(),
2,
100,
10,
2.0,
false,
6.0,
)?;
let i_duration = i_start.elapsed();
println!("Time elapsed in generate is: {:?}", i_duration);
save_wav(&generate, "voxcpm.wav", 16000)?;