garythung / trashnet Goto Github PK

I followed the instructions and encountered a little problem with genicam
This is error:

Traceback (most recent call last):
  File "trash_classifier.py", line 25, in <module>
    camera = pylon.InstantCamera(pylon.TlFactory.GetInstance().CreateFirstDevice())
  File "/home/trungluc12706/.local/lib/python3.6/site-packages/pypylon/pylon.py", line 1614, in CreateFirstDevice
    return _pylon.TlFactory_CreateFirstDevice(self, *args)
_genicam.RuntimeException: No device is available or no device contains the provided device info properties. : RuntimeException thrown (file 'TlFactory.cpp', line 692)

I'm newbie, please help me.

Dear Sir/Madam

I am using windows and i saw that torch works only for linux. I saw that the libraries that you require are the following:

luarocks install torch
luarocks install nn
luarocks install optim
luarocks install image
luarocks install gnuplot

if I install the libraries in windows using pip install nn and so on (except torch that works in linux).
Can i perform the same steps as you did?

Hi!
Great work with the model and the idea.

I'm working on something very very similar and have a few questions.
I can understand due to the limited availability of the data, the CNN models don't perform as expected. If you can explain the dataset retrieval process, it would be very helpful.
How did you get the dataset? The README suggests, all of it was manually built. If yes then did you try some data mining or scraping other than the manual process?

PS. The images look to be captured in very specific conditions like a white cardboard background etc.
How do you see this translate into a real-world scenario where a similar model is deployed using a camera??

HI, @garythung

I run th train.lua, but met this error:

17:47:34 weightDecay = 0.025
17:47:34 learningRate = 1.25e-05
17:47:34 Starting training for 100 epochs
/tmp/luarocks_cunn-scm-1-4893/cunn/lib/THCUNN/ClassNLLCriterion.cu:57: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [7,0,0] Assertion t >= 0 && t < n_classes failed.
/tmp/luarocks_cunn-scm-1-4893/cunn/lib/THCUNN/ClassNLLCriterion.cu:57: void cunn_ClassNLLCriterion_updateOutput_kernel(Dtype *, Dtype *, Dtype *, long *, Dtype *, int, int, int, int, long) [with Dtype = float, Acctype = float]: block: [0,0,0], thread: [27,0,0] Assertion t >= 0 && t < n_classes failed.
THCudaCheck FAIL file=/tmp/luarocks_cutorch-scm-1-9564/cutorch/lib/THC/generic/THCStorage.c line=32 error=59 : device-side assert triggered
/home/ctilab/torch/install/bin/luajit: cuda runtime error (59) : device-side assert triggered at /tmp/luarocks_cutorch-scm-1-9564/cutorch/lib/THC/generic/THCStorage.c:32
stack traceback:

What's wrong with me?

My CUDA & Luarocks Package install is allright, may be..

Hi, can I get the updated link of the google drive for the dataset? Thank you.

I tried this

git clone [email protected]:datasets/<dataset ID>

but it doesnt download the actual images. Just readme and other files. How can I obtain the actual files?