-
### Description
We propose overlapping the all gather op and the matmul op by executing partial out product of the matmul whenever a chunk is recieved in during the ring all gather. A ring all gather…
-
Ensure that the URL and email fields follow the correct format.
For depth fields:
depth_from (if specified) must be less than depth_to.
For sample dates:
If a parent sample is specified, val…
-
## 环境
硬件环境:modelarts 910B 八卡
cann7.0.0beta1
软件环境:
transformers 4.41.2
ms-swift 2.2.0.dev0 /home/ma-user/work/gengyichen/swift
## 第一种版本
torch …
-
同标题
-
**To Reproduce**
Steps to reproduce the behaviour:
1. Add a comment to a block of text
2. Go to a different page
3. In the Reviewing Pane, select the comment
4. Word does not scroll to the d…
-
This code:
```julia
using ClimaOcean
using Oceananigans
# 2 degree grid
grid = LatitudeLongitudeGrid(size = (180, 75, 1),
halo = (7, 7, 7),
…
-
### Reminder
- [X] I have read the README and searched the existing issues.
### System Info
CANN | cann_8.0.rc1
-- | --
PyTorch | pytorch_2.1.0
PyTorch_npu | 2.1.0.post3-20240413
…
-
### Is there an existing issue for this?
- [X] I have searched the existing issues
### Is your feature request related to a problem? Please describe.
随着越来越多的国内厂商采购国产GPU,如百度、科大讯飞等采购了华为910B,国产化适配越来越重…
bjgyl updated
3 months ago
-
你好,我在npu的4卡上部署qwen2-72b-instruct模型的时候,出现现存溢出。同样的模型、同样的4卡使用mindie可以推理。使用命令如下
```
ASCEND_RT_VISIBLE_DEVICES=0,1,2,3 NPROC_PER_NODE=4 swift deploy --model_type qwen2-72b-instruct --model_id_or_path .…
-
Hello!
I'm having a few issues going on in a mobile browser.
When I open the application in a mobile browser iOS or Android during the scroll, some widgets disappear that are in the tree belo…