search

huawei-noah / vega

AutoML tools chain
http://www.noahlab.com.hk/opensource/vega/
Other
842 stars 175 forks source link

vega: command not found #114

Closed vanessasidrim closed 3 years ago

vanessasidrim commented 3 years ago

I performed the installation following the instructions in the install.md file, but when I try to run the vega command, I get "vega: command not found"

zhangjiajin commented 3 years ago

@vanessasidrim

When Vega is installed, commands such as vega are installed in the ~/.local/bin directory and the path is added to $PATH. However, $PATH of the current user terminal does not contain this path. Therefore, login again or make the setting take effect.

vanessasidrim commented 3 years ago

The vega command worked, however, when trying to run SPNAS on several gpus using the vega.pipeline command, a No module named error is returned. Could you help me, please.

zhangjiajin commented 3 years ago

@vanessasidrim

Please provide more detailed information, such as logs.

vanessasidrim commented 3 years ago

I try execute CUDA_VISIBLE_DEVICES=3,4,5 vega.pipeline ./nas/sp_nas/spnas.yml and get as return bash: vega.pipeline: command not found

zhangjiajin commented 3 years ago

@vanessasidrim 

CUDA_VISIBLE_DEVICES=3,4,5 vega ./nas/sp_nas/spnas.yml
vanessasidrim commented 3 years ago

I tried this but the following error is returned to me:

Notes: msgpack: Variation is ok, as long as everything is above 0.6 python: Variation is sometimes ok, sometimes not. It depends on your workloads warnings.warn (version_module.VersionMismatchWarning (msg [0] ["warning"])) distributed.protocol.pickle - INFO - Failed to deserialize b '\ x80 \ x04 \ x95 \ xd7 \ x01 \ x00 \ x00 \ x00 \ x00 \ x00 \ x00 \ x8c \ x1evega.core.scheduler.worker_env \ x94 \ x8c \ tWorkerEnv \ x94 \ x93 \ x94) \ x81 \ x94} \ x94 (\ x8c \ x11workers_each_node \ x94K \ x01 \ x8c \ x0cdevice_quota \ x94K \ x03 \ x8c \ x10master_host_name \ x94N \ x8c \ nmaster_pid \ x94 \ xmaster \ x0bdevice_list \ x94] \ x94 \ x8c \ r_backend_type \ x94 \ x8c \ x07PYTORCH \ x94 \ x8c \ x0fdevice_category \ x94 \ x8c \ x03GPU \ x94 \ x8c \ ttemp_path \ x94/ x8c/7/8/8/9/8/9/9/9/9/8x8x8/9x8x8/8/8/8/8/8/8/8/8/8/8/8/8x8/8/8/8/8/8x8/8/8/8x8/8/8x8x8/ both temp \ x94 \ x8c \ x14worker_null_file \ x94 \ x8c4 / vega / examples / tasks / 0528.093414.763 / temp / .vega_null \ x94 \ x8c \ x18worker_device_folder \ x94 \ x8c6 / vega / examples / tasks / 0528.0934. vega_device \ x94 \ x8c \ r_cuda_devices \ x94] \ x94 (\ x8c \ x013 \ x94 \ x8c \ x014 \ x94 \ x8c \ x015 \ x94e \ x8c \ x14_ori_rank_table_file \ x94Nub. ' Traceback (most recent call last): File "/root/anaconda3/lib/python3.8/site-packages/distributed/protocol/pickle.py", line 67, in loads return pickle.loads (x) ModuleNotFoundError: No module named 'vega' distributed.utils - ERROR - No module named 'vega'

zhangjiajin commented 3 years ago

@vanessasidrim

It is possible that the Vega was not installed properly.

Run the following commands to install and pay attention to the information displayed during the installation:

pip3 install --user -U noah-vega
vanessasidrim commented 3 years ago

The problem had occurred when installing the dark, I need to install the library in the environment