【Inference】Fix qwen2.5-vl and deepseek-vl2 high-performance inference (#1175)

Co-authored-by: nifeng <nemonameless@qq.com>

Author: cheng221

deploy/deepseek_vl2/README.md

@@ -6,31 +6,29 @@
 |             Model               |
 |---------------------------------|
 | deepseek-ai/deepseek-vl2-small  |
-| deepseek-ai/deepseek-vl2        |
 
 ## 环境安装
 [安装PaddlePaddle](https://github.com/PaddlePaddle/PaddleMIX?tab=readme-ov-file#3-%EF%B8%8F%E5%AE%89%E8%A3%85paddlepaddle)
 - **python >= 3.10**
 - **paddlepaddle-gpu 要求develop版本**
 ```bash
-# Develop 版本安装示例，请确保使用的Paddle版本为develop版本
+# Develop 版本安装示例
 python -m pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu123/
 ```
 
 2） [安装PaddleMIX环境依赖包](https://github.com/PaddlePaddle/PaddleMIX?tab=readme-ov-file#3-%EF%B8%8F%E5%AE%89%E8%A3%85paddlepaddle)
 ```bash
 # pip 安装示例，安装paddlemix、ppdiffusers、项目依赖
-python -m pip install -e . 
+python -m pip install -e .
 python -m pip install -e ppdiffusers
 python -m pip install -r requirements.txt
 
 # 安装PaddleNLP
-pip uninstall paddlenlp && rm -rf PaddleNLP
-git clone https://github.com/PaddlePaddle/PaddleNLP.git
+pip uninstall -y paddlenlp && rm -rf PaddleNLP
+git clone -b release/3.0-beta4-new --depth=1 https://github.com/PaddlePaddle/PaddleNLP.git
 cd PaddleNLP
 pip install -e .
-cd csrc
-python setup_cuda.py install
+pip install https://paddlenlp.bj.bcebos.com/ops/cu118/paddlenlp_ops-3.0.0b4-py3-none-any.whl
 ```
 
 ## 3 高性能推理
@@ -39,56 +37,72 @@ python setup_cuda.py install
 
 ```
 export CUDA_VISIBLE_DEVICES=0
-export FLAGS_cascade_attention_max_partition_size=163840
-export FLAGS_mla_use_tensorcore=1
+export FLAGS_mla_use_tensorcore=0
+export FLAGS_cascade_attention_max_partition_size=128
+export FLAGS_cascade_attention_deal_each_time=16
 python deploy/deepseek_vl2/deepseek_vl2_infer.py \
     --model_name_or_path deepseek-ai/deepseek-vl2-small \
     --question "Describe this image." \
     --image_file paddlemix/demo_images/examples_image1.jpg \
     --min_length 128 \
     --max_length 128 \
+    --inference_model True \
+    --append_attn True \
+    --mode dynamic \
+    --dtype bfloat16 \
     --top_k 1 \
     --top_p 0.001 \
     --temperature 0.1 \
     --repetition_penalty 1.05 \
-    --block_attn True \
+    --benchmark
+
+# 多图推理
+python deploy/deepseek_vl2/deepseek_vl2_infer_multi_image.py \
+    --model_name_or_path deepseek-ai/deepseek-vl2-small \
+    --question "What are in these images." \
+    --image_file_1 paddlemix/demo_images/examples_image1.jpg \
+    --image_file_2 paddlemix/demo_images/examples_image2.jpg \
+    --image_file_3 paddlemix/demo_images/examples_image1.jpg \
+    --min_length 128 \
+    --max_length 128 \
     --inference_model True \
     --append_attn True \
     --mode dynamic \
     --dtype bfloat16 \
-    --mla_use_matrix_absorption
+    --top_k 1 \
+    --top_p 0.001 \
+    --temperature 0.1 \
+    --repetition_penalty 1.05 \
+    --benchmark
 ```
 
 ### b. wint8 高性能推理
 ```
 export CUDA_VISIBLE_DEVICES=0
-export FLAGS_cascade_attention_max_partition_size=163840
-export FLAGS_mla_use_tensorcore=1
+export FLAGS_mla_use_tensorcore=0
+export FLAGS_cascade_attention_max_partition_size=128
+export FLAGS_cascade_attention_deal_each_time=16
 python deploy/deepseek_vl2/deepseek_vl2_infer.py \
     --model_name_or_path deepseek-ai/deepseek-vl2-small \
     --question "Describe this image." \
     --image_file paddlemix/demo_images/examples_image1.jpg \
     --min_length 128 \
     --max_length 128 \
-    --top_k 1 \
-    --top_p 0.001 \
-    --temperature 0.1 \
-    --repetition_penalty 1.05 \
-    --block_attn True \
     --inference_model True \
     --append_attn True \
     --mode dynamic \
     --dtype bfloat16 \
-    --mla_use_matrix_absorption \
-    --quant_type "weight_only_int8" 
+    --top_k 1 \
+    --top_p 0.001 \
+    --temperature 0.1 \
+    --repetition_penalty 1.05 \
+    --quant_type "weight_only_int8" \
+    --benchmark True
 ```
 
 ## 4 一键推理 & 推理说明
-进入PaddleMIX目录运行
-```bash
 cd PaddleMIX
 sh deploy/deepseek_vl2/shell/run.sh
-```
 #### 参数设定
 |     parameter      |      Value     |
 | ------------------ | -------------- |
@@ -102,10 +116,3 @@ sh deploy/deepseek_vl2/shell/run.sh
 | ------------------ | -------------- |
 |     min_length     |       128      |
 |     min_length     |       128      |
-
-以下为单张图片的测速情况
-
-|             model              |    Paddle高性能推理    |    Paddle     |
-| ------------------------------ | ---------------------| ------------- |
-| deepseek-ai/deepseek-vl2-small |          9.3 s       |     12.8 s    |
-| deepseek-ai/deepseek-vl2       |           -          |     17.2 s    |