Ryan Deng

Is it necessary to install Cuda8 or do I not have to install it if I use the CPU only Tensorflow? Also, this only occurs in demo 2. Demo 1...

Do you have any idea why the __del__ method in VideoFileClip.py causes all these threading errors? I am really stuck right now. I ran the program on Windows as well...

EDIT: I think the error is what a previous user said about using their own videos as input. Could you point to where you hard coded values? Sorry, I am...

@JunshengFu Hi, I don't want the lane departure warning. I just want to be able to detect vehicles in a video. Could you tell me what parts to modify/comment out?...

I narrowed down the issue to the setup phase, particularly line 206 of abysetup.cpp. I'm not sure why this part causes the error? Could it be related to the multi-threading?

Changing that parameter fundamentally changes the model > same problem It's similar to #4094 > > 1. I modify the `num_experts_per_tok` to 8 in `config.json` of mixtral-8x7b and everything works...

Does that mean you can call multiple asynchronous operations on the same stream like multiple sends or multiple receives before the prior calls finish? If so, how does MPI know...

``` #include #include #include #include #include #include #include #include #include constexpr static int NUM_ELEMENTS = 512; constexpr static int NUM_LEVELS = 4; constexpr static int NUM_STREAMS = 8; constexpr bool...

So changing the code such that the behavior above doesn't occur, by slightly modifying how stream indices are assigned to elements. In this code, I changed how `elem_to_stream_mapping` is constructed....

I also tried assigning each elem pair that communicates with each other to a unique stream index to avoid any sort of conflicts, and that seems to hang as well....