Profile cover photo
Profile photo
Trung Tran
28 followers
28 followers
About
Posts

Been playing around with Recurrent Neural Networks recently, I attempted to create a text generator using the famous Harry Potter as training data. And below is the result I achieved, a generated sample with J. K. Rowling's unique style:

"Because Gripe meant, of course, the first thing that have certainly mentioned your aunt and uncle while wearing those too many of you tonight.... A pair of free Stunners called force the pipes in Britain. If the Howler doesn't believe it! This is brilliant!" said Fleur to Ron and Hermione, impatient again. Hermione snapped and then, with a slight sneer ful inches as she reached the stone gargoyle. "They're only socks, who was doing this year?"

Wanna make your own text generator, check out my newest post at here:
https://chunml.github.io/ChunML.github.io/project/Creating-Text-Generator-Using-Recurrent-Neural-Network/

Hope you enjoy the post :)
Add a comment...

Post has attachment
Been playing around with Caffe for a while, it's definitely one of the coolest Deep Learning frameworks we have nowadays.
"Being able to train your Network without writing a line of code" maybe the main point which helps Caffe stand out from the crowd but that also makes Caffe look like a damn black box.
So this blog post is for anyone who is struggling to train Caffe's Network using your own dataset. So glad if this can help save you quite some time!
Add a comment...

Post has attachment
Add a comment...

Post has attachment
Faced a problem which caused the result images not to show up, and seems like so many guys are dealing with this too. So hope this post can help :)
Add a comment...

Post has attachment
Here's my experience in installing Caffe on Ubuntu 16.04 with CUDA and cuDNN to make use of the power of GPU. I also did some speed comparison between CPU and GPU to have a better view on how GPU can significantly outperform CPU in training Deep Neural Network. Enjoy!
Add a comment...

Post has attachment
My 9th tutorial on Machine Learning: Regularization. Regularization is a must-know technique if you want to deal with Overfitting problem. It's definitely an efficient tool you have just added to your toolbox. Hope it helps!
Add a comment...

Post has attachment
8th tutorial on Machine Learning is on! This time I talked about a cool but easy-to-understand algorithm called Decision Tree. So enjoy the tutorial and I'll be glad if it can help!
Add a comment...

Post has attachment
https://www.coursera.org/account/accomplishments/certificate/RJ9VZPFBQXTD

Finally complete Machine Learning course. Great thanks to Andrew Ng for providing such well constructed course!
Add a comment...

Post has attachment
Check out my 7th post on Machine Learning. This time I talked about Cross Validation, a technique to evaluate the Model's performance.
Add a comment...

Post has attachment
Done 9th week!
Photo
Add a comment...
Wait while more posts are being loaded