Rebecca J

Every time I've had this issue it was related to running out of VRAM -- maybe using a different GPU or shrinking your dataset could help? In my experience, the...

Hello there. While playing on a shared network, I saw some issues that appear to be related to network corruption of the save game, in that it happens intermittently when...

``` Traceback (most recent call last): File "generate.py", line 88, in main() File "generate.py", line 71, in main generate_samples_interactive( File "/workspace/gpt-neox/megatron/text_generation_utils.py", line 745, in generate_samples_interactive for ( File "/workspace/gpt-neox/megatron/text_generation_utils.py", line...

With that line added on my end... ``` Context prompt >>> this cat DEBUG tensor([0, 1], device='cuda:3') tensor([0, 0, 0], device='cuda:3') DEBUG tensor([0, 1], device='cuda:2') tensor([ 58, 46434, 0], device='cuda:2')...

See #604 for the greater prototyping effort underway to ensure that all processes have the correct context_length and context_tokens.

This is resolved

Initial torchtyping integration is occurring in #600

You might want to take a look at #16 if it hasn't been addressed.

I've been training NeRFs with this code recently, and have made a few changes to the COLMAP extraction method, which appears to help significantly with training at the expense of...

slash-under/instant-ngp@efdd42a851039b689b84b84191e0358dbd35f07d It's about 1-1.25 GB of RAM per thread at peak usage during initial extraction for about 250 3840x2160 frames, but greatly improves NeRF model quality with large datasets in...