Access Denied Tensorflow Error? about pygta5

[J-XianSheng]

set
MODEL_NAME = 'anything'

LOAD_MODEL = False

[mantoone]

Great!

There's a bug with the motion script. You need to add screen as a parameter to the function:
So change:
def motion_detection(t_minus, t_now, t_plus):
To:
def motion_detection(t_minus, t_now, t_plus, screen):

And in the test_model script add screen to the function call:
https://github.com/Sentdex/pygta5/blob/master/3.%20test_model.py#L148

[mantoone]

Did you modify the collect script and set the number of outputs (keys) to 17? If so you also need to change it in the training script. If you didn't intend to do that can you show the code of your collection script?

…

On Jul 5, 2017 2:16 PM, "Daniel Mossaband" ***@***.***> wrote: Hey @mantoone <https://github.com/mantoone> . I'm also having the same error. I changed MODEL_NAME = 'anything' LOAD_MODEL = False And I only have up to training_data-200.npy So I set FILE_I_END = 200 And now I get training/training_data-37.npy training_data-37.npy 500 --------------------------------- Run id: anything Log directory: log/ --------------------------------- Training samples: 450 Validation samples: 50 -- Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)' training/training_data-69.npy training_data-69.npy 500 --------------------------------- Run id: anything Log directory: log/ --------------------------------- Training samples: 450 Validation samples: 50 -- Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)' training/training_data-190.npy training_data-190.npy 500 --------------------------------- Run id: anything Log directory: log/ --------------------------------- Training samples: 450 Validation samples: 50 -- Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)' training/training_data-112.npy training_data-112.npy 500 --------------------------------- Run id: anything Log directory: log/ --------------------------------- Training samples: 450 Validation samples: 50 -- Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)' training/training_data-157.npy training_data-157.npy 500 --------------------------------- Run id: anything Log directory: log/ --------------------------------- Training samples: 450 Validation samples: 50 -- Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)' training/training_data-58.npy training_data-58.npy 500 --------------------------------- Run id: anything Log directory: log/ --------------------------------- Training samples: 450 Validation samples: 50 -- Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)' training/training_data-104.npy training_data-104.npy 500 Over and over until I stop it — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#50 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFrn1RHdFGZeaUobJKRTlAHuJ1kBnbj9ks5sK3BsgaJpZM4N3tsc> .

[Caleb-J773]

Okay that fixed it kinda but now there's a new error. It can't actually find any npy files when they are there.

[Errno 2] No such file or directory: 'J:/phase10-random-padded/training_data-1778.npy' [Errno 2] No such file or directory: 'J:/phase10-random-padded/training_data-1276.npy' I tried changing up the directory it was listing to my pc but it still can't find any training files.

[mantoone]

Where did you save the files? It works fine for me, but I did need to change the paths in the scripts.

[Caleb-J773]

I made a new folder and saved it to my desktop
C:\Users\User1\Desktop\drivers

[mantoone]

If you want to save the recorded data there change the line https://github.com/Sentdex/pygta5/blob/master/1.%20collect_data.py#L101
file_name = 'X:/pygta5/phase7-larger-color/training_data-{}.npy'.format(starting_value)
to
file_name = 'C:/Users/User1/Desktop/drivers/training_data-{}.npy'.format(starting_value)

[Caleb-J773]

Okay but it's also doing the same errors to the other files, like training data and testing data. but im sure testing data is just acting like that cause well training data file isn't working, i'll try your step

[Caleb-J773]

Alright this isn't working Training data is just refusing to work, There must be something wrong with the current code.
[Errno 2] No such file or directory: 'C:/Users/User1/Desktop/drivers/training_data-1127.npy'
[Errno 2] No such file or directory: 'C:/Users/User1/Desktop/drivers/training_data-172.npy'
[Errno 2] No such file or directory: 'C:/Users/User1/Desktop/drivers/training_data-1575.npy'
[Errno 2] No such file or directory: 'C:/Users/User1/Desktop/drivers/training_data-249.npy'
[Errno 2] No such file or directory: 'C:/Users/User1/Desktop/drivers/training_data-874.npy'
[Errno 2] No such file or directory: 'C:/Users/User1/Desktop/drivers/training_data-119.npy'
training_data-2.npy 500

(It skips -1 )

It's not even training anymore.

[mantoone]

Did you check that the folder contains those files? I needed to adjust the FILE_I_END variable. Take a look at https://github.com/Sentdex/pygta5/blob/master/Versions/v0.03/2.%20train_model.py#L11

If you have data until training_data-100.npy, change it to FILE_I_END = 100

[Caleb-J773]

The folder has training data 1 and 2 , that's it
just for the start

[mantoone]

Try FILE_I_END = 2

[Caleb-J773]

Traceback (most recent call last): File "C:\Users\User1\Desktop\drivers\test_model.py", line 245, in <module> main() File "C:\Users\User1\Desktop\drivers\test_model.py", line 148, in main delta_count_last = motion_detection(t_minus, t_now, t_plus) File "C:\Users\User1\Desktop\drivers\motion.py", line 46, in motion_detection dst = cv2.addWeighted(screen,1.0, delta_view,0.6,0) NameError: name 'screen' is not defined
Test_model is erroring out now. I have everything installed and just verified, Python has all the requirements satisfied

Also doing what you told me above allowed me to train finally, but running into this problem ^^

[Caleb-J773]

Okay now it's giving this error
Traceback (most recent call last):
File "test_model.py", line 245, in
main()
File "test_model.py", line 190, in main
motion_log.append(delta_count)
NameError: name 'delta_count' is not defined

[mantoone]

I suppose delta_count should be delta_count_last on the line:
https://github.com/Sentdex/pygta5/blob/master/3.%20test_model.py#L190

[Caleb-J773]

Okay well that somehow worked but it's still not training. I waited 1 hour for it to train and it "trained" yet the file never updated. and just started to repeat.

[mantoone]

What was the output from the "2. train_model.py" script? Did you add the same MODEL_NAME to the test_model script?

[Caleb-J773]

I added the same model name, There's no error other then MODEL SAVED. then it just repeats the same training data over and over

[mantoone]

The model is saved every 10th epoch. It should finish automatically when it's done. By default after 30 epochs (30 passes over the data).

[DanMossa]

Hey @mantoone . I'm also having the same error. I changed

MODEL_NAME = 'anything'

LOAD_MODEL = False

And I only have up to
training_data-200.npy

So I set FILE_I_END = 200

And now I get

training/training_data-37.npy
training_data-37.npy 500
---------------------------------
Run id: anything
Log directory: log/
---------------------------------
Training samples: 450
Validation samples: 50
--
Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)'
training/training_data-69.npy
training_data-69.npy 500
---------------------------------
Run id: anything
Log directory: log/
---------------------------------
Training samples: 450
Validation samples: 50
--
Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)'
training/training_data-190.npy
training_data-190.npy 500
---------------------------------
Run id: anything
Log directory: log/
---------------------------------
Training samples: 450
Validation samples: 50
--
Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)'
training/training_data-112.npy
training_data-112.npy 500
---------------------------------
Run id: anything
Log directory: log/
---------------------------------
Training samples: 450
Validation samples: 50
--
Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)'
training/training_data-157.npy
training_data-157.npy 500
---------------------------------
Run id: anything
Log directory: log/
---------------------------------
Training samples: 450
Validation samples: 50
--
Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)'
training/training_data-58.npy
training_data-58.npy 500
---------------------------------
Run id: anything
Log directory: log/
---------------------------------
Training samples: 450
Validation samples: 50
--
Cannot feed value of shape (64, 17) for Tensor 'targets/Y:0', which has shape '(?, 9)'
training/training_data-104.npy
training_data-104.npy 500

Over and over until I stop it

[DanMossa]

@mantooneoops I did change it.

Thanks mate

[mantoone]

You're welcome!

[frossaren]

@mantoone It would be nice if you did a pull request on your fix for screen not defined in motion. I think many people have the same problem

[mantoone]

@frossaren There it is now: #64

[frossaren]

Great! Den 12 aug. 2017 18:18 skrev mantoone <[email protected]>: @frossaren<https://github.com/frossaren> There it is now: #64<#64> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#50 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ATxZZ-TGTx19VqCI66rUvlDjISTcohHXks5sXdA-gaJpZM4N3tsc>.