add qwen3-asr

This commit is contained in:
jhqxxx
2026-02-05 00:43:49 +08:00
parent 073723447f
commit 1c507a7b12
21 changed files with 1079 additions and 971 deletions
+2 -2
View File
@@ -2,9 +2,9 @@
use std::time::Instant;
use aha::utils::{audio_utils::create_hann_window, tensor_utils::interpolate_nearest_1d};
use aha::utils::{tensor_utils::interpolate_nearest_1d};
use anyhow::Result;
use candle_core::{DType, Tensor};
use candle_core::{Tensor};
// use symphonia::core::io::MediaSourceStream;
#[test]
+52 -56
View File
@@ -1,12 +1,57 @@
use std::{pin::pin, time::Instant};
use aha::models::{GenerateModel, fun_asr_nano::generate::FunAsrNanoGenerateModel, qwen3_asr::generate::Qwen3AsrGenerateModel};
use aha::models::{GenerateModel, qwen3_asr::generate::Qwen3AsrGenerateModel};
use aha_openai_dive::v1::resources::chat::ChatCompletionParameters;
use anyhow::Result;
use rocket::futures::StreamExt;
#[test]
fn qwen3_asr_generate() -> Result<()> {
// RUST_BACKTRACE=1 cargo test -F cuda qwen3_asr_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!("{}/Qwen/Qwen3-ASR-0.6B/", save_dir); //Qwen/Qwen3-ASR-1.7B
let message = r#"
{
"model": "qwen3-asr",
"messages": [
{
"role": "user",
"content": [
{
"type": "audio",
"audio_url":
{
"url": "https://package-release.coderbox.cn/aiway/test/other/%E5%93%AA%E5%90%92.wav"
}
}
]
}
]
}
"#;
// "metadata": {"language": "Chinese"}
let mes: ChatCompletionParameters = serde_json::from_str(message)?;
let i_start = Instant::now();
let mut model = Qwen3AsrGenerateModel::init(&model_path, None, None)?;
let i_duration = i_start.elapsed();
println!("Time elapsed in load model is: {:?}", i_duration);
let i_start = Instant::now();
let res = model.generate(mes)?;
let i_duration = i_start.elapsed();
println!("generate: \n {:?}", res);
if res.usage.is_some() {
let num_token = res.usage.as_ref().unwrap().total_tokens;
let duration_secs = i_duration.as_secs_f64();
let tps = num_token as f64 / duration_secs;
println!("Tokens per second (TPS): {:.2}", tps);
}
println!("Time elapsed in generate is: {:?}", i_duration);
Ok(())
}
#[tokio::test]
async fn qwen3_asr_stream() -> Result<()> {
// RUST_BACKTRACE=1 cargo test -F cuda qwen3_asr_stream -r -- --nocapture
let save_dir =
aha::utils::get_default_save_dir().ok_or(anyhow::anyhow!("Failed to get save dir"))?;
let model_path = format!("{}/Qwen/Qwen3-ASR-0.6B/", save_dir);
@@ -26,69 +71,20 @@ fn qwen3_asr_generate() -> Result<()> {
}
]
}
],
"metadata": {"language": "Chinese"}
]
}
"#;
let mes: ChatCompletionParameters = serde_json::from_str(message)?;
let i_start = Instant::now();
let mut model = Qwen3AsrGenerateModel::init(&model_path, None, None)?;
let mut fun_asr_model = Qwen3AsrGenerateModel::init(&model_path, None, None)?;
let i_duration = i_start.elapsed();
println!("Time elapsed in load model is: {:?}", i_duration);
let i_start = Instant::now();
let res = model.generate(mes)?;
let mut stream = pin!(fun_asr_model.generate_stream(mes)?);
while let Some(item) = stream.next().await {
println!("generate: \n {:?}", item);
}
let i_duration = i_start.elapsed();
println!("generate: \n {:?}", res);
// if res.usage.is_some() {
// let num_token = res.usage.as_ref().unwrap().total_tokens;
// let duration_secs = i_duration.as_secs_f64();
// let tps = num_token as f64 / duration_secs;
// println!("Tokens per second (TPS): {:.2}", tps);
// }
println!("Time elapsed in generate is: {:?}", i_duration);
Ok(())
}
// #[tokio::test]
// async fn qwen3_asr_stream() -> Result<()> {
// // RUST_BACKTRACE=1 cargo test -F cuda fun_asr_nano_stream -r -- --nocapture
// let save_dir =
// aha::utils::get_default_save_dir().ok_or(anyhow::anyhow!("Failed to get save dir"))?;
// let model_path = format!("{}/Qwen/Qwen3-ASR-0.6B/", save_dir);
// let message = r#"
// {
// "model": "qwen3-asr",
// "messages": [
// {
// "role": "user",
// "content": [
// {
// "type": "audio",
// "audio_url":
// {
// "url": "https://package-release.coderbox.cn/aiway/test/other/%E5%93%AA%E5%90%92.wav"
// }
// },
// {
// "type": "text",
// "text": "语音转写:"
// }
// ]
// }
// ]
// }
// "#;
// let mes: ChatCompletionParameters = serde_json::from_str(message)?;
// let i_start = Instant::now();
// let mut fun_asr_model = FunAsrNanoGenerateModel::init(&model_path, None, None)?;
// let i_duration = i_start.elapsed();
// println!("Time elapsed in load model is: {:?}", i_duration);
// let i_start = Instant::now();
// let mut stream = pin!(fun_asr_model.generate_stream(mes)?);
// while let Some(item) = stream.next().await {
// println!("generate: \n {:?}", item);
// }
// let i_duration = i_start.elapsed();
// println!("Time elapsed in generate is: {:?}", i_duration);
// Ok(())
// }