[WIP] added README for MNIST #68
Conversation
|
Thanks for opening your first pull request in this repository! Someone will review it when they have a chance. In the mean time, please be sure that you've handled the following things, to make the review process quicker and easier:
Thank you again for your contributions! 👍 |
There was a problem hiding this comment.
Hey @codeboy5, awesome work. Some handful comments. Let me know what you think. I will gives this a more in depth review tomorrow morning. Thanks a lot 💯
| MNIST dataset contains images of handwritten digits. It is one of the most common datasets used for image classifcation. It has 60,000 28x28 grayscale images under the training set and 10,000 28x28 grayscale images under the test set. Each pixel has a value between 0 and 255. | ||
|
|
||
|  | ||
| *Sample Images From The MNIST Database* |
There was a problem hiding this comment.
Could we have this in a separate line ?
Also The images might look better with centre alignment. Is there any way to do that?
There was a problem hiding this comment.
Yeah, I found a way online, would try it and see.
Kaggle/README.md
Outdated
|
|
||
| ##### 1. DigitRecognizer | ||
| For More Details related to the Model, check out the **DigitRecognizer** Directory. | ||
| After running the model for 50 cycles, I was able to obtain the best validation set accuracy of **97.2143%**. |
There was a problem hiding this comment.
Instead of using I here can we use one here?
There was a problem hiding this comment.
the model would be more apt I believe
| ### Processing The Dataset | ||
|
|
||
| ##### 1. Loading The Dataset From The File To A Useable Format | ||
| The Load Class is used to load a matrix from the file.The type of file is guessed automatically from the extension. For a list of supported file types please refer to [Load Class](https://github.com/mlpack/mlpack/blob/master/src/mlpack/core/data/load.hpp). Please note that since mlpack requires matrices to be in column major format and usually data is stored in a row major format, the matrix by default is transposed. In case the data is originally in column major format, consider setting the parameter to False. Refer to documentation of Load Class for further information. |
There was a problem hiding this comment.
Do this we could add a code snippet here that shows this function loading mnist.
There was a problem hiding this comment.
Sure that would be beneficial, would make that change.
I have added a README for the dataset and how to process it to be used in the models.
@kartikdutt18 Kindly go through this and let me know what can I improve on.
As discussed in #65