-
```
What steps will reproduce the problem?
1: Open Power VR options
2: *Edit - Delete - Add lines as needed*
3: *Edit - Delete - Add lines as needed*
What is the expected output? What do you see i…
-
```
What steps will reproduce the problem?
1: Open Power VR options
2: *Edit - Delete - Add lines as needed*
3: *Edit - Delete - Add lines as needed*
What is the expected output? What do you see i…
-
当我使用命令 model = YOLOv10('yolov10n.yaml')选定模型后,训练yolov10n.yaml时,进行完一个epoch周期训练再进行验证时会报错CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling `cublasSgemv(handle, op, m, n, &alpha, a, lda, x, incx, &beta,…
-
将GLM-10B-chinese模型切分为MP_SIZE=8, 然后finetune seq2seq任务时,在训练阶段一切正常,在eval阶段报错IndexError。
GLM-10B-chinese模型切分为MP_SIZE=8后,每个rank embedding为维度为:50048/8=6256。
但是eval阶段时好像将模型当作MP_SIZE=1的情况进行评估模型。请问这块怎么配置或者怎么…
-
It would be nice to have automatic CSS generation included in this loader.
This is a gulp extension I wrote that I'm wonder if your interested on collaborating on integrating. This also generates a …
-
## ❓ Questions and Help
### Before asking:
1. search the issues.
2. search the docs.
#### What is your question?
I change the `ddp-backend` from `no_c10d` to `fully_sharded` as it is said…
-
```
Google translate AJAX API seems to have a character limit. 2000 I think. That
means that pages like this fail
http://www.flickr.com/photos/shenyun2010/4626974300/in/photostream/ (use top of
t…
-
I would greatly appreciate your help with this error. Here is [tutorial] (https://huggingface.co/blog/fine-tune-whisper) that i followed. Thanks in advance.
''from transformers import Seq2SeqTrai…
-
Hello,
I managed without problem to run `torchinfo.summary` on `YOLOv5n` that I loaded from the torch hub.
```pycon
>>> torchinfo.summary(torch.hub.load('ultralytics/yolov5', 'yolov5n'), in…
-
This issue will document the list of ideas for recipes.