babysor / babysor/MockingBird

为什么main的新版本,我开启训练后只能使用CPU,没法使用GPU,训练速度低到 0.08setp/s

Open
#902 3 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
36.9k
Forks
5.2k
PR merge metrics
No merged PRs in 30d

Description

为什么main的新版本,我开启训练后只能使用CPU,没法使用GPU,训练速度低到 0.08setp/s

{| Epoch: 52/3334 (3/3) | Loss: 1.005 | 0.081 steps/s | Step: 0k | }
{| Epoch: 53/3334 (3/3) | Loss: 0.9955 | 0.081 steps/s | Step: 0k | }
{| Epoch: 54/3334 (3/3) | Loss: 0.9905 | 0.081 steps/s | Step: 0k | }
{| Epoch: 55/3334 (3/3) | Loss: 0.9797 | 0.081 steps/s | Step: 0k | }

任务管理器中,GPU显示为3%,CPU为60%

是还需要修改什么参数吗

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reproducing the reported main-version training run and inspect its GPU/CPU selection and training parameters; compare the reported Task Manager utilization and 0.081 steps/s. Done means identifying whether a parameter or environment issue prevents GPU use and documenting the required setting.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
machine-learning, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.