-
I observed that the eval callbacks definition had not recieved any input parameter which contains batch item(s) from the dataloader. Moreover the ChannelPruner API doesnt perform finetuning on its own…
-
### Search before asking
- [X] I have searched the YOLOv5 [issues](https://github.com/ultralytics/yolov5/issues) and [discussions](https://github.com/ultralytics/yolov5/discussions) and found no simi…
-
Hey all,
I'm trying to use angr for some concolic execution. From looking at some GitHub issues on here + the Driller source, my intuition is that to do this I must (1) get a basic block address tr…
-
Previous log #43
-
Hi,
I have a question in the source code logic. In the channel pruning method, why the depthwise module is excluded from compression by the class ConvNoDepthwiseLayerSelector, thanks.
-
Hello to all contributors,
The last POT 0.6 release brought new features to the library and we have now 25 papers implemented in POT. It was discussed that before making the 1.0 release, we should …
-
Previous log #31
-
Hi!
I'm currently trying to prune a resnet50 using the `configs/pruning/mmpose/dcff/dcff_topdown_heatmap_resnet50_coco.py` config and I get the following error, even after applying the fix mentioned …
-
我扩充词表后,仅在模型权重目录下更新词表相关配置文件。
如果不改变其他代码和参数,预训练过程中能否对新添加的词元进行训练
-
Prior discussion:
https://github.com/keean/zenscript/issues/14#issuecomment-308722477
https://github.com/keean/zenscript/issues/33#issuecomment-362720879
https://github.com/keean/zenscript/is…