lge-arc-advancedai / auptimizer Goto Github PK

Describe the bug
When USER is not defined in a system, the following error appears:

INFO - Executing command: /opt/conda/bin/python -m aup.setupdb.sqlite .aup/env.ini --user root --cpu 4 --log info
Traceback (most recent call last):
  File "/opt/conda/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/opt/conda/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/opt/conda/lib/python3.6/site-packages/Auptimizer-1.2-py3.6.egg/aup/setupdb/sqlite.py", line 152, in <module>
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 829, in __call__
    return self.main(*args, **kwargs)
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 782, in main
    rv = self.invoke(ctx)
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 1066, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 610, in invoke
    return callback(*args, **kwargs)
  File "/opt/conda/lib/python3.6/site-packages/Auptimizer-1.2-py3.6.egg/aup/setupdb/sqlite.py", line 144, in main
  File "/opt/conda/lib/python3.6/site-packages/Auptimizer-1.2-py3.6.egg/aup/utils.py", line 89, in get_default_username
  File "/opt/conda/lib/python3.6/os.py", line 669, in __getitem__
    raise KeyError(key) from None
KeyError: 'USER'
CRITICAL - Failed in setup commands

To Reproduce
Steps to reproduce the behavior:

export USER
python -m aup.setup

Expected behavior
Assign a default user name.

I noticed that it is easy to 'auptimize' over hyper-parameters which are integer or float types, but is it possible to optimize over categorical hyper-parameters, for example loss type, weight decay type, etc?

Hi i am facing issue with save model if auto conversion is used.
Step1:
In training script i have added the
aup_save_model(save_model, model)

and save_model call back is implemented as
def save_model(model):
import os
print("save_model...")
os.makedirs('save_model')
model.save('./save_model/mnist.h5')

Step2: convert the script in hpo format.
python -m aup.convert train_script aup_config function
Step3: Training with hpo converted file
python -m aup experiment.json
Call back written for saving the model (save_model) is not getting hit.
I doubt the aup.convert is not able to handle the save_model changes

Result: model is not getting saved

I get the following error when running Auptimizer on AWS
catch unexpected error size mismatch in put! 0 != 110

I believe there is an error at https://github.com/LGE-ARC-AdvancedAI/auptimizer/blob/master/src/aup/Proposer/hpbandster/core/dispatcher.py#L232 where del ... leaves a None in the worker_pool without ever reducing the size of the collection. I think filter or some other reduction might be needed when cleaning up.

Is your feature request related to a problem? Please describe.
Does auptimizer work with any python3.x versions? The requirements.txt file confuses me referencing both 2.7 and 3.x.

Describe the solution you'd like
I'd like to see py3 support.