[SDXL] Vanilla cache dataset报错 - Githubissues

mindspore-lab / mindone

one for all, Optimal generator with No Exception

https://mindspore-lab.github.io/mindone/

Apache License 2.0

332 stars 63 forks source link

[SDXL] Vanilla cache dataset报错 #370

Closed daiyuxin0511 closed 4 months ago

daiyuxin0511 commented 4 months ago

Hardware Environment | 硬件环境

please tell us what kind of hardware can reproduce your error? 请告诉我们您报错的后端类型
- [ ] Ascend

Software Environment | 软件环境

MindSpore version: 请告诉我们您正在使用的MindSpore版本：
- [ ] 2.2.10
Python version(e.g., 3.7.5):py3.7
OS(e.g., Linux Ubuntu 16.04)
GCC/Compiler version:

Describe the current behavior | 目前输出

Describe the expected behavior | 期望输出

please describe expected outputs or functions you want to have: 请告诉我们您期望得到的结果或功能

生成cache_data

Steps to reproduce the issue | 复现报错的步骤

code url | 代码链接: https://github.com/mindspore-lab/mindone/blob/master/examples/stable_diffusion_xl/vanilla_finetune.md#31-cache-dataset
command that can reproduce your error | 可以复现报错的命令添加--max_device_memory 59GB参数，跑1p和8p都报错

Related log / screenshot | 完整日志

https://roma.huawei.com/csb/#/hec/eiwizard/training/com.csi.dpsl.atlasB.mindspore/trainingjobv2?thirdType=detail&id=5fbdcfeb-2a68-4c2e-9b5d-62ad212ecbda&trackId=009689d2-ab93-4b6b-a640-a5939b055d55 https://roma.huawei.com/csb/#/hec/eiwizard/training/com.csi.dpsl.atlasB.mindspore/trainingjobv2?thirdType=detail&id=2d9a6d1a-b965-4481-962e-db6b924d4664&trackId=009689d2-ab93-4b6b-a640-a5939b055d55

Special notes for this issue | 其他信息