add deepseek_ocr

This commit is contained in:
jhqxxx
2025-11-22 23:27:14 +08:00
parent c95b91ffc0
commit a14c35014a
20 changed files with 1715 additions and 151 deletions
+60 -7
View File
@@ -1,11 +1,13 @@
use aha::models::deepseek_ocr::{generate::DeepseekOCRGenerateModel, processor::DeepseekOCRProcessor};
use std::{pin::pin, time::Instant};
use aha::models::{GenerateModel, deepseek_ocr::generate::DeepseekOCRGenerateModel};
use aha_openai_dive::v1::resources::chat::ChatCompletionParameters;
use anyhow::Result;
use candle_core::{DType, Device, IndexOp, Tensor};
use rocket::futures::StreamExt;
#[test]
fn deepseek_ocr_test() -> Result<()> {
// RUST_BACKTRACE=1 cargo test -F cuda deepseek_ocr_test -- --nocapture
fn deepseek_ocr_generate() -> Result<()> {
// RUST_BACKTRACE=1 cargo test -F cuda deepseek_ocr_generate -r -- --nocapture
let message = r#"
{
"model": "deepseek-ocr",
@@ -35,9 +37,60 @@ fn deepseek_ocr_test() -> Result<()> {
"#;
let model_path = "/home/jhq/huggingface_model/deepseek-ai/DeepSeek-OCR/";
let mes: ChatCompletionParameters = serde_json::from_str(message)?;
let device = Device::cuda_if_available(0)?;
let dtype = DType::BF16;
let mut model = DeepseekOCRGenerateModel::init(model_path, Some(&device), Some(dtype))?;
let i_start = Instant::now();
let mut model = DeepseekOCRGenerateModel::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!("Time elapsed in generate is: {:?}", i_duration);
println!("generate: \n {:?}", res);
Ok(())
}
#[tokio::test]
async fn deepseek_ocr_stream() -> Result<()> {
// test with cuda: RUST_BACKTRACE=1 cargo test -F cuda deepseek_ocr_stream -r -- --nocapture
let message = r#"
{
"model": "deepseek-ocr",
"messages": [
{
"role": "user",
"content": [
{
"type": "image",
"image_url":
{
"url": "file://./assets/img/ocr_test1.png"
}
},
{
"type": "text",
"text": "<image>\n<|grounding|>Convert the document to markdown. "
}
]
},
{
"role": "assistant",
"content": ""
}
]
}
"#;
let model_path = "/home/jhq/huggingface_model/deepseek-ai/DeepSeek-OCR/";
let mes: ChatCompletionParameters = serde_json::from_str(message)?;
let i_start = Instant::now();
let mut model = DeepseekOCRGenerateModel::init(model_path, None, None)?;
let i_duration = i_start.elapsed();
println!("Time elapsed in load model is: {:?}", i_duration);
let mut stream = pin!(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(())
}