Gooddz1
Gooddz1
原本我是想直接升级到tensorrt8,但是tx2的tensorrt升级需要重新刷机,这太难了。跪求,在tensorrt7中添加ScatterND 的插件
> ScatterND一般是由inplace操作产生的,你可以考虑去掉inplace而去掉ScatterND 请问怎么去掉inplace,有教程吗?
> ScatterND一般是由inplace操作产生的,你可以考虑去掉inplace而去掉ScatterND 例如:attention_matrix[self.non_diag_inds] = attention.flatten() # ScatterND ,应该怎么处理这个?
请问,问题解决了吗
> I had a similar issue and solved it by putting the model in eval() and passing dummy input through it once, before doing pruning.我有一个类似的问题,并通过将模型放入eval()中并在进行修剪之前传递一次虚拟输入来解决它。 Can you share the implementation...
> Sorry, I am not really working on an identical case, I am working on tracing a yolov8 model and just mentioned the solution as the error message I got...
> may be you could try the pruned model Pruned model: https://modelscope.cn/models/bdsqlsz/AnyDoor-Pruned/summary How do I need to modify the configuration file to use Anytool's pruned version of the model when...
> Use ["Pruned model - 4.57 GB"](https://modelscope.cn/models/bdsqlsz/AnyDoor-Pruned/files) I am running it on a GTX 1070 8GB VRAM GPU (utilizing 2GB from Shared Memory - Total 10GB) How do I need...