请问 dygraph 用的paddle版本是什么呐,必须使用编译版本的么?
已关闭
请问 dygraph 用的paddle版本是什么呐,必须使用编译版本的么?
Created by: ARQlalala
hi,问个简单的问题,用develop和paddle 1.8.4 时候LinearLrWarmup这些 paddle.optimizer.lr_scheduler还有ParallelEnv都没有呐,必须要使用编译版本的paddle么
Created by: ARQlalala
嗯嗯 环境准备好了, 对于单机多卡来说,运行指令如下: python3 -m paddle.distributed.launch
--selected_gpus="0,1,2,3"
tools/train.py
-c ./configs/ResNet/ResNet50.yaml
-o print_interval=10 ~然后错误如下: Please make sure the version is good with your code. INFO 2020-09-17 14:11:16,592 utils.py:269] terminate all the procs ERROR 2020-09-17 14:11:16,593 utils.py:439] ABORT!!! Out of all 4 trainers, the trainer process with rank=[0, 1, 2, 3] was aborted. Please check its log. INFO 2020-09-17 14:11:19,593 utils.py:269] terminate all the procs
请问这个是需要修改什么参数呐
Created by: littletomatodonkey
嗯嗯 环境准备好了, 对于单机多卡来说,运行指令如下: python3 -m paddle.distributed.launch --selected_gpus="0,1,2,3" tools/train.py -c ./configs/ResNet/ResNet50.yaml -o print_interval=10 ~
然后错误如下: Please make sure the version is good with your code. INFO 2020-09-17 14:11:16,592 utils.py:269] terminate all the procs ERROR 2020-09-17 14:11:16,593 utils.py:439] ABORT!!! Out of all 4 trainers, the trainer process with rank=[0, 1, 2, 3] was aborted. Please check its log. INFO 2020-09-17 14:11:19,593 utils.py:269] terminate all the procs
请问这个是需要修改什么参数呐
你好,这个是版本check出现问题了,因为现在发布的2.0.0b0版本,所以check会失败,你可以拉下最新的clas代码,或者修改下代码的这里
Created by: ARQlalala
嗯嗯好的,请教下,在paddle-gpu 2.0-beta版本中使用paddle.randperm函数,报错
Error Message Summary:
UnimplementedError: Place CUDAPlace(0) is not supported. Please check that your paddle compiles with WITH_GPU or WITH_XPU option or check that your train process hold the correct gpu_id if you use Executor. (at /paddle/paddle/fluid/platform/device_context.cc:67) [operator < randperm > error]
这个是数据的操作要转成cuda模式么,在paddle中需要怎么转呢,谢谢~
Created by: littletomatodonkey
嗯嗯好的,请教下,在paddle-gpu 2.0-beta版本中使用paddle.randperm函数,报错
Error Message Summary:
UnimplementedError: Place CUDAPlace(0) is not supported. Please check that your paddle compiles with WITH_GPU or WITH_XPU option or check that your train process hold the correct gpu_id if you use Executor. (at /paddle/paddle/fluid/platform/device_context.cc:67) [operator < randperm > error]
这个是数据的操作要转成cuda模式么,在paddle中需要怎么转呢,谢谢~
你是怎么使用的呢?应该不用转成cuda模式的,看着是你的Paddle安装的不是gpu版本的,我下面这段代码运行没啥问题
import paddle paddle.disable_static() out1 = paddle.randperm(5)