-
> All the variables should be inserted at the top of the entry block, so this alloca, store, alloca pattern would ought to look like alloca, alloca, store, ...
_Originally posted by @schweitzpgi in…
-
### 🐛 Describe the bug
- Exact code worked fine on 2.2.1 (includes 2.3.0, 2.3.1)
- Deployment is with a docker container: nothing besides pytorch has changed (same CUDA version, etc).
- `nvidia-s…
-
I can't find info on this error anywhere! I am running Docker Desktop with WSL. My docker compose file looks like this:
```
version: '3'
services:
app:
container_name: "sd"
bui…
-
Hello! Please add video memory usage! `nvidia-smi` provides info. For coding with Pytorch and Tensorflow using CUDA we need this info!
Thanks very much.
-
**Describe the bug**
As described in the title, after finishing all of the installs and building nemo and megatron-lm from source, assuming that the model has been trained with megatron-lm.
**St…
-
### Description of the bug | 错误描述
(data) (base) randy@3080Ti:/mnt/projects/settings$ magic-pdf -p small_ocr.pdf
2024-09-04 00:07:47.984 | INFO | magic_pdf.libs.pdf_check:detect_invalid_chars:5…
-
I've followed the instructions in [Nvidia container toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) and [Docker as a non-root user](https://do…
-
**Describe the bug**
Circular import error with PyTorch nightly. If I uninstall deepspeed it works fine.
```
Traceback (most recent call last):
File "/test/oss.py", line 322, in
mp.spawn…
-
## Description
I'm trying to deserialize model on device A and serialize it on device B, but failed。Except for the graphics driver, the environment of A and B is the same. Does the graphics…
-
Hello, I am running Fedora 40 Kinoite (Bazzite) with an NVIDIA GeForce RTX 4070 Ti (proprietary drivers, version 560.35.03). My aim was to install Jellyfin via a podman systemd container file and use …