update for mr100

Dockerfile

@@ -1,5 +0,0 @@
-FROM corex:4.3.0
-
-WORKDIR /workspace
-RUN pip install diffusers==0.34.0 sentencepiece transformers==4.55.2
-COPY main.py dataset.json /workspace/

Dockerfile

@@ -0,0 +1 @@
+Dockerfile.mrv100

Dockerfile.a100

@@ -0,0 +1,5 @@
+FROM harbor-contest.4pd.io/zhangyiqun/public/pytorch:2.6.0-cuda12.4-cudnn9-devel
+WORKDIR /workspace
+# ENV PT_SDPA_ENABLE_HEAD_DIM_PADDING=1
+RUN pip install diffusers transformers sentencepiece -i https://nexus.4pd.io/repository/pypi-all/simple
+COPY main.py test.sh dataset.json /workspace/

Dockerfile.bi100

@@ -0,0 +1,6 @@
+FROM git.modelhub.org.cn:980/enginex-iluvatar/bi100-3.2.1-x86-ubuntu20.04-py3.10-poc-llm-infer:v1.2.2
+
+WORKDIR /workspace
+ENV PT_SDPA_ENABLE_HEAD_DIM_PADDING=1
+RUN pip install diffusers==0.34.0
+COPY main.py test.sh dataset.json /workspace/

Dockerfile.mrv100

@@ -0,0 +1,7 @@
+FROM git.modelhub.org.cn:980/enginex-iluvatar/mr100_corex:4.3.0
+
+WORKDIR /workspace
+COPY whls-mrv100 /packages
+RUN pip install diffusers==0.34.0 sentencepiece transformers==4.55.2
+# RUN pip install /packages/*.whl
+COPY main.py test.sh dataset.json /workspace/

README.md

@@ -1,40 +1,28 @@
+## Installation
+参考Dockerfile，构建运行镜像
+
 ## Quickstart
 
-### 构建镜像
-```bash
-docker build -t diffusers:v0.1 .
-```
-其中，基础镜像 corex:4.3.0 通过联系天数智芯智铠100厂商技术支持可获取
-
-### 模型下载
-模型地址：https://modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5
-并放到目录：`/mnt/contest_ceph/zhanghao/models/stable-diffusion-v1-5`（如更改目录，请修改后面的执行脚本中的模型路径）
-
 ### 测试程序
-1. 准备输入数据集，可以参考示例`dataset.json`
+1. 下载模型：https://modelscope.cn/models/AI-ModelScope/stable-diffusion-v1-5
-2. 在docker镜像里运行测试程序，会根据`dataset.json`内容，在`output`目录下生成图片文件。
+
+2. 运行测试程序
+
+修改测试程序`test.py`里面的模型路径，直接执行即可
+
 ```bash
-./run_in_docker.sh
+python3 test.py
+```
+
+### 批量测试程序
+1. 准备输入数据集`dataset.json`，可以参考示例`dataset.json`
+2. 运行测试程序
+```bash
+python3 main.py --model "/mnt/contest_ceph/zhanghao/models/stable-diffusion-v1-5" --json  "dataset.json" --results "results.json" --outdir "output" --device cuda --dtype fp16
 ```
 
 ## 测试结果
+|      | A100 平均生成时间（秒） | 智铠100 平均生成时间（秒） |
+|------|-------------------------|----------------------------|
+| 时间 | 1.5                 | 4.1                     |
 
-| 模型名称 | A100 平均生成时间（秒） | 智铠100 平均生成时间（秒） | 备注 |
-| ----- | ----- | ----- | ----- |
-| AI-ModelScope/stable-diffusion-3.5-medium | 5.3253 | 56.6237 | |
-| AI-ModelScope/stable-diffusion-v1-5 | 1.3944 | 3.9280 | |
-| Cleaner/lo_dress_gothic_style2_v2 | 1.8821 | 2.5121 | |
-| Cleaner/Tyndalleffect_Light | 1.7133 | 2.5138 | | 
-| HanYixuan1/Retro_style | 1.6764 | 2.4901 | |
-| JeffVan/Oil_Paint_Style_LORA | 1.6673 | 2.7132 | |
-| lljjcc/IndianSarres | 1.7252 | 2.5209 | |                                                           
-| lljjcc/outdoor | 1.6684 | 2.5230 | |      
-| muse/flux_vae | 1.0812 | 2.5350 | |      
-| MusePublic/489_ckpt_FLUX_1/base_model | 1.2286 | 2.5382 | |
-| rewfueranro/cartoon_lora | 1.6880 | 2.6043 | |
-| YorickHe/fairy_lora | 1.7765 | 2.5139 | |
-| YorickHe/JK_uniform_lora | 1.7844 | 2.5219 | |                                                      
-| YorickHe/outdoor_photo_lora | 1.6441 | 2.4851 | | 
-| YorickHe/polaroid_lora | 1.6272 | 2.5117 | |  
-| YorickHe/Winter_hanfu_lora | 1.5962 | 2.4892 | |
-| ZackWang123/filmvelvia_lora | 1.8465 | 2.5157 | |

run_in_docker.sh

@@ -1,3 +1,3 @@
 #! /usr/bin/env bash
-image=diffusers:v0.1
+image=harbor-contest.4pd.io/zhanghao/diffusers:bi100-0.2
-docker run -v `pwd`:/workspace -v /mnt/contest_ceph/zhanghao/models/stable-diffusion-v1-5:/workspace/stable-diffusion-v1-5 --device=dev/iluvatar1:/dev/iluvatar0 $image python3 main.py --model "./stable-diffusion-v1-5" --json  "dataset.json" --results "results.json" --outdir "output" --device cuda --dtype fp16
+docker run -it -v /root/zhanghao:/workspace -v /mnt:/mnt --device=dev/iluvatar1:/dev/iluvatar0 $image bash

run_in_docker_a100.sh

@@ -0,0 +1,3 @@
+#! /usr/bin/env bash
+image=harbor-contest.4pd.io/zhanghao/diffusers:a100-0.2
+docker run -it -v /home/zhanghao/workspace:/workspace -v /mnt:/mnt $image bash

run_in_docker_mrv100.sh

@@ -0,0 +1,3 @@
+#! /usr/bin/env bash
+image=harbor-contest.4pd.io/zhanghao/diffusers:mrv100-0.2
+docker run -it -v /root/zhanghao:/workspace -v /mnt:/mnt --device=dev/iluvatar0:/dev/iluvatar0 $image bash

test.py

@@ -0,0 +1,13 @@
+from diffusers import DiffusionPipeline
+import torch
+import time
+
+model_path = "/mnt/contest_ceph/zhanghao/models/stable-diffusion-v1-5"
+# model_path = "/mnt/contest_ceph/zhanghao/models/stable-diffusion-3.5-medium"
+pipeline = DiffusionPipeline.from_pretrained(model_path, torch_dtype=torch.float16)
+pipeline.to("cuda")
+start = time.time()
+image = pipeline("An image of a squirrel in Picasso style").images[0]
+end = time.time()
+print(f"elapsed: {end - start}")
+image.save("squirrel_picasso.png")

test.sh

@@ -0,0 +1 @@
+python3 main.py --model "/mnt/contest_ceph/zhanghao/models/stable-diffusion-v1-5" --json  "dataset.json" --results "results.json" --outdir "output" --device cuda --dtype fp16