-
One issue I have consistently as a media server user is convincing my content creation teams that the long export times of HAP codec is worthwhile. I've done some basic testing of different HAP encode…
-
I've been thinking about the procedural generation of terrain, and I've begun to write some algorithms for it. I've run into some issues. Fair warning, this is a long post. Lots of questions.
First…
-
Hey @johndpope, I noticed that you used NoiseInjectionBlock with StyledConv.
In [supplementary](https://openaccess.thecvf.com/content/CVPR2024/supplemental/Gao_Implicit_Motion_Function_CVPR_2024_su…
-
## Objective
Develop an intelligent sampling algorithm that can extract representative content from entire documents or collections of documents, ensuring balanced representation in validation prompt…
-
I benchmarked the downstream performance of the ABQ W4A16 (per-channel, symmetric) quantized version of the [MiniCPM-1B-sft-llama-format](https://www.modelscope.cn/models/OpenBMB/MiniCPM-1B-sft-llama-…
-
## Goal-State/What/Result
Have separate Tokio runtimes for query/AI computation (inference, embeddings, etc) and acceleration refreshes. The goal is that all queries should stay fast no matter what e…
-
Inspiration is taken from how git packs achieve high compression rates with good random access performance. The 2 key components are (1) a clustering strategy to store similar objects together and (2)…
-
This would be useful for FFT's or integration over a subset of the volume as discussed in #1101 . Currently algorithms that rely on such computations are forced to compute on the full volume, which im…
-
-
Store metadata with each model and checkpoint, such as the training source code and parameters so I know which models correspond to which problems, parameters, runs, algorithm revisions, etc.
Consi…