NeedsMoar

The only real problem with NTFS alternate streams is that very few people know about them. Hash data isn't large enough that I'd worry about storing it there. I think...

Kornia requires flash attention 2 but doesn't list it in its requirements.txt because they're apparently idiots so there's no way of even knowing what version it wants because the flash_attn_2...

> Hi - few weeks ago I opened an issue on CPU bottleneck, finally found out the root cause. It wasn't the CPU bottleneck really - it was the CPU...

Also, 1-2s for SDXL unets from a cloud provider? Is that on some compute instance or do you have a 40 / 56 / 100Gb connection to the internet? :-)...

To give people an idea, the default build of flash attention by itself on a 32/64 core threadripper pro 5975WX with 512GB of ram on older versions of the makefile...

That's not a hacker, that's a moron (along with anybody who blindly downloads a suspiciously password protected zip file from mediafire and runs anything in it).

By rewriting half of it, adding JNI functions, and compiling with the NDK as a shared library.

> I tried defining NOMINMAX before including windows.h that did not fix the problem. Here are the error messages I'm getting. > > 1>D:\Craig\Visual Studio 2022\Projects\CraigsSystem\src\portfolioreturnobjectivefunction.cpp(6,9): warning C4603: 'NOMINMAX': macro...

The error checking compiler version is almost certainly because ninja and isn't installed and just aren't finding the compiler right at that stage. Its setup.py has an explicit hook /...

Torch's minimum compiler version is 19.0.24215 btw. The highest version I have is the same 14.41.34120 you have, which is compiler version 19.41.34120, the lowest is 19.29.30154 from march of...