-
### Quick Summary
---
Latest build 1.3.0 stated it added servers for Arcsys games, though when attempting to make a lobby in BlazBlue: Calamity Trigger, the game crashes (does not close, is froze…
-
### Description
Hello. I play dota 5000 games. Of these, I was the best player 500 times. My rank reaches 2000 mmr and then for some reason it drops rapidly. The losing streak can be 20 games in a ro…
-
### Reminder
- [X] I have read the README and searched the existing issues.
### System Info
- `llamafactory` version: 0.9.1.dev0
- Platform: Linux-3.10.0-957.el7.x86_64-x86_64-with-glibc2.35
- Py…
-
### Description
1. INTERMITTENT DISCONNECTION ON RANKED!!!!! GAME!!!!!!
2. NO HERO OR CREEP DISPLAY UPON RECONNECTING!
### Example Match ID (and possibly Timestamp)
_No response_
### Screenshot…
-
## ❓ Questions and Help
Hi, everyone. When i use `ddp`, i have encounter some question..
And I want to running on `single node 4 gpus` on gcp.
If i set `num_work=0`, it will be work but training is…
-
## Issue description
I am based on the POMO code [POMO](https://github.com/yd-kwon/POMO) to change it to a single machine multi-GPU graphics card running mode
Let me explain the specific code lo…
-
## Motivation
The current PyTorch Operator focuses on a one process per pod architecture, which may not fully utilize the generic design of PyTorch and can underperform. When users adopt `torchrun`…
-
somewhere, display the rank order of each compound - maybe directly under the selection box, in very small italics font.
 and Bare Metal (BM) and got different training results.
## Experiment settings
- Two V100 GPU mac…