Oliver
Oliver
@LJoson 是的，我尝试了一下khadas自家提供的sdk，它们同样调用的是TIM-VX，没有发现这个问题，应该是Tengine这边出bug了 > 我看了一下，对应示例，我个人理解，原因可能是默认尺寸是相等的，对应的tensor以及构建gragh,是根据尺寸想等的情况下处理，所以不相等时可能出现异常，而且项目没有对这个异常进行处理（依稀记得虫叔那篇文章提过输入尺寸的规则），更具体的好像要研究一下对应参数，以及run gragh流程了
any updates?
> New Hyper-SD (**Trajectory** Segmented **Consistency** **Distillation**) from bytedance apparently runs atop of diffusers's TCDScheduler aswell. https://hyper-sd.github.io/ Yes we adopt the TCDScheduler on our unified LoRAs. And we find the...
@JettHu Nice done, friend! We have tested it on our Hyper-SD and it works perfect!!! We will soon update it on our [HF page](https://huggingface.co/ByteDance/Hyper-SD). Thanks for your great job!
@JettHu Yes, our 1-step unified lora is able to inference on 1-8 steps with the help of TCDScheduler within single checkpoint! And we found previous implementation of it behaves differently...
Same question about this.
@zRzRzRzRzRzRzR Any updates on this?