Jannik Kossen

Hi @Balandat, thanks for the quick reply. Turns out, you are spot on! Setting `torch.backends.cuda.matmul.allow_tf32 = False` gives the desired behaviour: cpu and gpu results look the same. It might...

Dear Boyuzz, I'm glad you're enjoying our work and that it works well in your setting. Quick comment: We found that – even if accuracy is the desired target quantity...

Hi boyuzz, out of interest – may I ask what project you are using active testing for? You can respond to me in private if you don't feel like sharing...

Hi Lingzhi, thanks for reaching out! > I wonder where exactly are the potential causes. I remember we encountered this same behaviour and could not figure out why memory was...

Dear lihaolin1, thanks for your message and interest in STOVE! We don't have support for moving MNIST built-in. However, it should be very possible to built it in yourself. A)...

@jatinganhotra, we only performed a quick preliminary search for loophole exploitation via `git log` + `--all`. Our findings were: * 20250522_tools_claude-4-sonnet: cheating for pytest-dev__pytest-6202; future leakage for pytest-dev__pytest-5840, pytest-dev__pytest-6202 *...