Nicolas Granger

In a torch -> onnx -> onnxruntime[TensorRTExecutionProvider] pipeline, pytorch casts values to uint64 to satisfy [type constrains on the BitShift operator](https://onnx.ai/onnx/operators/onnx__BitShift.html#type-constraints).  However, this causes an error in the onnxruntime...

I cannot reproduce it here. Forticlient is running with EMS connected, but only forticlient can connect to the VPN, not openforticlient.

If I'm to go manually with CLI I might as well rm the checkpoints. A separate button or checkbox to delete artifacts was more what I had in mind.

In case it can help anyone, I have written a library which can load any given data sample from Waymo (and also KITTI, NuScenes or ZOD). As we discussed in...

My 2ct contribution for this discussion: - Gnome works to integrate apps so as to avoid the need for trays. i3 doesn't provide integration for volume control, network, battery status,...

I'll post it here if I find one. I just though it could help with readability and when we need to squeeze a table within the line width.

Here is an MRE of this issue: `a.py`: ```py class A: __slots__ = ["a"] def __init__(self): self.a = 0 ``` `slots.ipynb`: ```py # %% %load_ext autoreload %autoreload 2 # %%...

Triton just raises an assertion error when trying to index a local tensor. I suppose it is related to this issue. Are there any workarounds?

> https://github.com/facebookresearch/xformers/blob/main/xformers/ops/_triton/k_index_select_cat.py > > There’s this in xformers seems similar to indexing into a sparse tensor Yes but it goes through global memory which is slow as mentioned by @Jokeren.

Actually I'm pretty sure this is a driver configuration issue so I've submitted a bug to https://bugzilla.kernel.org/show_bug.cgi?id=218928