-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
What caused this error? KeyError: 'loss' #22
Comments
It seems you might have specified a batch size that is larger than the total data size, so that the loss is not computed by call back function by the end of the epoch, because no epoch can run. |
Thank you for your reply. |
@panzhengsen Yes it's the |
yes, I printed |
I see, can you read in your training and validation data (for example, by As an example, the input to DeepSEA train/val data are two numpy arrays: |
ok. |
I believe there is no problem with my data set construction.
The error message is as follows:
`
['traindata', 'trainxdata']
['validdata', 'validxdata']
BUILDING
controller = GeneralController
modeler = EnasCnnModelBuilder
knowledge = zero
reward = LossAucReward
manager = EnasManager
time budget set to: 24:00:00
env_type = EnasTrainEnv
Trial -1: Start training model with sample_arc...
Traceback (most recent call last):
File "amber_deepsea.py", line 124, in
amb.run()
File "/home/zhouss/AMBER/amber/wrapper.py", line 104, in run
self.env.train()
File "/home/zhouss/AMBER/amber/architect/trainEnv.py", line 418, in train
self.manager.get_rewards(trial=-i, model_arc=None, nsteps=warmup_nsteps)
File "/home/zhouss/AMBER/amber/architect/manager.py", line 544, in get_rewards
verbose=self.verbose,
File "/home/zhouss/AMBER/amber/modeler/child.py", line 743, in fit
prog = ", ".join(["%s=%.4f" % (name, hist.history[name][-1]) for name in to_print])
File "/home/zhouss/AMBER/amber/modeler/child.py", line 743, in
prog = ", ".join(["%s=%.4f" % (name, hist.history[name][-1]) for name in to_print])
KeyError: 'loss'
`
hist.histroy keys: dict_keys([])
I printed hist.histroy keys() to check the keywords and found it was empty
The text was updated successfully, but these errors were encountered: