-
-
### System Info
Accelerate version: 1.1.0
transformers version: 4.46.2
DeepSpeed version: 0.14.4
Platform: Linux 5.15.0-101-generic #111-Ubuntu SMP x86_64 GNU/Linux
Python version: 3.10.14
PyT…
-
### 详细叙述
如题，这个提交导致我自己生产的planet文件没法在var分区上持久化。在/etc/config/zero下可持久化，每次重启服务或者重启路由，都可一直在，放var后都是默认zerotier自己生成的planet给覆盖了。求revert
![image](https://github.com/coolsnowwolf/lede/assets/20067401/17be5a78…
-
Latest build works great - but unfortunately I can't get the zero to reboot on a 'reboot now' command.
I made an armbian build based on the zero.wip config - that's rebooting just fine, but doesn't re…
-
Obtaining file:///home/house365ai/xxm/vllm2
Installing build dependencies ... done
Checking if build backend supports build_editable ... done
Getting requirements to build editable ... error
…
-
I am getting "float division by zero" error whenever I try to quantize mixtral related models with autogptq,
and here is my code.
```
from transformers import AutoTokenizer, TextGenerationPipeli…
-
I am having an issue where I don't get the error responsible for the ContainerError unless I turn off tty. Is the expected behavior?
With tty enabled:
```
ContainerError: Command '['train']' in …
-
**Describe the bug**
When I set `bf16.enabled` to `true` and `weight_quantization.quantized_initialization`, the model would repeatedly generate the same token.
**To Reproduce**
Run the follo…
Atry updated
5 months ago
-
### What happened + What you expected to happen
By running the given example of AlphaZero, line 51: mask = obs["action_mask"] in ray/rllib/algorithms/alpha_zero/ranked_rewards.py" reports there is …
-
### Prerequisites
- [X] I'm using the latest version
- [X] I've read the relevant [documentation](https://knip.dev)
- [X] I've searched for [existing issues](https://github.com/webpro-nl/knip/iss…