jhqxxx
51 lines
1.6 KiB
Rust
51 lines
1.6 KiB
Rust
use std::time::Instant;
|
|
|
|
use aha::models::{GenerateModel, qwen2_5vl::generate::Qwen2_5VLGenerateModel};
|
|
use aha_openai_dive::v1::resources::chat::ChatCompletionParameters;
|
|
use anyhow::Result;
|
|
|
|
#[test]
|
|
fn robo_brain_generate() -> Result<()> {
|
|
// test with cuda: RUST_BACKTRACE=1 cargo test -F cuda robo_brain_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!("{}/BAAI/RoboBrain2.0-3B/", save_dir);
|
|
|
|
let message = r#"
|
|
{
|
|
"model": "qwen2.5vl",
|
|
"messages": [
|
|
{
|
|
"role": "user",
|
|
"content": [
|
|
{
|
|
"type": "text",
|
|
"text": "hello RoboBrain"
|
|
}
|
|
]
|
|
}
|
|
]
|
|
}
|
|
"#;
|
|
let mes: ChatCompletionParameters = serde_json::from_str(message)?;
|
|
let i_start = Instant::now();
|
|
let mut model = Qwen2_5VLGenerateModel::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 result = model.generate(mes)?;
|
|
let i_duration = i_start.elapsed();
|
|
println!("generate: \n {:?}", result);
|
|
if result.usage.is_some() {
|
|
let num_token = result.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(())
|
|
}
|