-
Hello,
Thanks for your great work on nvdiffrec!
This feature improved more quality(good PSNR then nvdiffrec), but therunning failed on optimization phase.
**So, could you please give me …
-
**Describe the bug**
I want to pretrain a BERT model on 8 A100 40G GPUS. The problem which I have is that I run out of CPU memory (not GPU memory). I cannot understand why. I am trying to load a 75G …
-
### Problem Description
On Llama3 70B Proxy Model, the training stalls & gpucore dumps. The gpucore dumps are 41GByte per GPU thus i am unable to send it. Probably easier for yall to reprod this er…
-
### Description
I started a work to increase the performance in `WebGPURenderer` and it will work for both `Backends`, it's about a better management of `Bindings` and `Cache`.
### Solution
-…
-
### Informations
- **Qiskit Aer version**:
Requirement already satisfied: qiskit-aer-gpu in /global/homes/g/gzquse/.conda/envs/qiskit-summer/lib/python3.11/site-packages (0.14.2)
Requiremen…
-
Hello,
Your paper looks very promising. I can't wait the release of model and code 😍
Could you provide more information about inference speed? And max resolution of generated video?
-
Hi, I'm having an issue with tensorrt conversion of model which uses 3D convolutions and processes 5D input.
Code to reproduce error (I cut the model to minimal example):
```
import tensorflow as…
-
These two functions, for solving banded matrices and performing symmetric indefinite factorizations, show up a lot in optimization and would be great to have in JAX.
The scipy routines are basicall…
-
Using an explicit SYCL queue instance for Kokkos::SYCL targeting a GPU results in a SYCL (icpx) error:
```
terminate called after throwing an instance of 'sycl::_V1::runtime_error'
what(): Nati…
-
**Motivation**
* Faster than raw operations by 200-1000x (Lazy optimization, Rust, Fastest Algorithms for every simple operation)
* Big amount of supported features
* Less memory intensive
* Can…