QuickDraw | Implementation of Quickdraw - an online game | Machine Learning library

Your conversion is correct, the bug is in the rendering code. It must be is_down = data[i][2] == 0 instead of is_down = data[i-1][2] == 0 in draw_stroke3.

This error didn't show up with the other paths as in all but two cases the new path starts at the end of the previous path. In the other case where you really move to a new start point the additional line coincided with a line already drawn.

UPDATE AND CORRECTION:
I noticed that I mis-interpreted the meaning of the pen-up bit: in fact it shows that the pen is to be lifted after drawing the current stroke, not for the current stroke as I though at first. Therefore your rendering code appears to be OK and the bug is in the stroke3 file generation.
I guess you can do it much simpler by recording the end points for each operation along with the op code (1 = move, 0 = draw) for the current operation. After conversion to a numpy array we can easily convert these absolute positions the relative displacements by do the difference of the first two columns and then shift the third column with the op codes backwards by one position:

Looks like you're having an interpreter version mismatch.

Remove your environment, then in your project folder try:

First of all, if I may say so:

Since I'm a complete beginner in machine learning and I have no knowledge about >how neural network work yet. I wanted to try with scikit-learn's algorithms.

This is not a good way to approach ML in general, I strongly suggest you start studying the basics at least, otherwise you won't be able to tell what's going on at all (it's not something you can figure out by trying it).

Back to your problem, applying Naive Bayes methods to raw images it's not a good strategy: the problem is that each pixel of your image is a feature and with images you can get a very high number of dimensions easily (also assuming each pixel is independant of its neighbors it's not what you want). NB is commonly used with documents and looking at this example on wikipedia might help you understand a bit more the algorithm.

In short, NB boils down to computing joint conditional probabilities, which boils down to counting co-occurences of features (words in wikipedia's example) being co-occurences of pixels in your case, which in turn boils down to computing a huge matrix of occurences that you need to formulate your NB model.

Now, if your matrix is made of all the words in a set of documents, this can get pretty expensive in both time and space (O(n^2)/2), with n being the number of features; instead, imagine the matrix being composed of ALL the pixels in your training set, as you're doing in your example... this explodes really fast.

That's why cutting your dataset to 1000 images allows your PC to not run out of memory. Hope it helps.

The problem was in the parse_tfexample function. In it, there is a dictionary with an element keyed "drawing" which is a sparse tensor. So I just converted it to dense using tf.sparse.to_dense(). Here is the code of the parse_tfexample:

The problem lies with how the data split is done. Notice that there are 5 classes and you do 0.2 validation split. By default there's no shuffling and in your code you feed the data in a sequential order. What that means:

• Training data consists entirely of 4 classes: 'cat.npy', 'dog.npy', 'apple.npy', 'banana.npy'. That's the 0.8 training split.
• Test data is 'flower.npy'. That's your 0.2 validation split. The model was never trained on this so it gets terrible accuracy.

Such results are only possible thanks to the fact that the validation_split=0.2, so you get close to perfect class separation.

Solution

Maybe is your dataset have empty line, you can try add this to check the error string

python train_model.py \ --training_data=rnn_tutorial_data/training.tfrecord-?????-of-????? \ --eval_data=rnn_tutorial_data/eval.tfrecord-?????-of-????? \ --classes_file=rnn_tutorial_data/training.tfrecord.classes

AFAIK using the above command works as well, it will simply read all the files in the folder where you have download the data files one by one.

create_tfrecord_for_prediction is certainly not my own creation, this code was mostly picked from another file from tensorflow guys create_dataset.py

Below I have pasted the almost all of the new code i added including the my modifications to the main() function

First of all, you didn't copy the code right. There are some mistakes.

1. in CharInfo change

   [FieldOffset(2)] public CharUnion Char;

   to

   [FieldOffset(0)] public CharUnion Char;

2. Change $CONOUT to this CONOUT$

3. First set the Attribute and then the AsciiChar

And of course, if you want the correct foreground-/background color representation then you have to delete the 2 << 4. I don't even know why you put it there.