tag:blogger.com,1999:blog-7583720.post5392093426561347512..comments2024-03-05T03:17:02.289-08:00Comments on Salmon Run: Deep Learning Models for Question Answering with KerasSujit Palhttp://www.blogger.com/profile/06835223352394332155noreply@blogger.comBlogger17125tag:blogger.com,1999:blog-7583720.post-33635010340171604932020-04-29T12:49:42.300-07:002020-04-29T12:49:42.300-07:00Glad I could help!
Glad I could help!<br />Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-40216858685329423412019-02-23T06:06:39.358-08:002019-02-23T06:06:39.358-08:00I am trying to make such model. But i am a beggine...I am trying to make such model. But i am a begginer. But after reading your post i got some knowledge to step forwardAnonymoushttps://www.blogger.com/profile/03887034004591070862noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-76100976434546572442017-05-13T17:20:40.827-07:002017-05-13T17:20:40.827-07:00Thanks Prakhar.Thanks Prakhar.Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-67230362642231413402017-05-11T10:44:56.269-07:002017-05-11T10:44:56.269-07:00Very well written. Very well written. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7583720.post-23764080937408160212017-03-26T12:01:00.795-07:002017-03-26T12:01:00.795-07:00Hi Rafat, difficult to say without more context. M...Hi Rafat, difficult to say without more context. Maybe dump out the layer.input_shape and output_shape for each layer and see if that helps? Or do model.summary()?<br />Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-82539579796172334962017-03-23T16:10:47.328-07:002017-03-23T16:10:47.328-07:00Dear Sujit pal,
Your blog really helping me a lo...Dear Sujit pal, <br />Your blog really helping me a lot learning the LSTM model with other models. Thanks a lot for that. I have one question for you I trying to use cosine similarity as defined in the original paper. for that I have written the codes. But I am facing problem in matching the shape of the model output and the Y output.<br />Error when checking model target: expected dense_5 to Rafathttps://www.blogger.com/profile/09645734277820906056noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-12572438771918728652017-03-15T03:30:47.099-07:002017-03-15T03:30:47.099-07:00Thanks for the reply. It solved the problem. Now c...Thanks for the reply. It solved the problem. Now can you please tell me a way to find the sentence similarity with the weights generated. I tried your approach for prediction but it is not giving me the correct result. I want the get a similarity score between 1 or 0. Or scores of 1 and 0 will also work. Thanks in advanceRafathttps://www.blogger.com/profile/09645734277820906056noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-11193285577266254042017-03-02T12:06:29.408-08:002017-03-02T12:06:29.408-08:00Thanks Rafat, and max_seqlen is the maximum number...Thanks Rafat, and max_seqlen is the maximum number of words in the question (and answer, which is much shorter) for all questions in the given dataset. Let me know if you have more questions and I will try to answer.<br />Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-47878342396157613892017-03-01T20:35:57.854-08:002017-03-01T20:35:57.854-08:00Dear Sujit pal,
I really am an admirer of your bl...Dear Sujit pal, <br />I really am an admirer of your blog, I am working to find the question question similarity, So I was following this particular blog using keras. is seq_maxlen the length of the maximum seq length of the the questions or the length of the data set.<br /><br />I am having problems specifying the parameters of the LSTM and attention model, Can you please help me a little bit Rafathttps://www.blogger.com/profile/09645734277820906056noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-61140326523389962192016-12-01T11:15:12.470-08:002016-12-01T11:15:12.470-08:00Hi Abewabu, thanks for the kind words. OpenNLP has...Hi Abewabu, thanks for the kind words. OpenNLP has a MaxEnt classifier (http://maxent.sourceforge.net/howto.html). I am not sure about this, but I am guessing you will extract concept features and match the concepts found in the question and the answer. The best answer will be the one with the highest probability.<br />Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-84549181844807637582016-12-01T08:40:41.056-08:002016-12-01T08:40:41.056-08:00Thanks Malaikannan, glad you liked it. Was good me...Thanks Malaikannan, glad you liked it. Was good meeting at the Demystifying AI and DL event.<br />Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-28515958624555360162016-11-23T21:43:11.206-08:002016-11-23T21:43:11.206-08:00Pretty cool blog post on Q&A. Actively working...Pretty cool blog post on Q&A. Actively working on this field, i understand the complexity involved in solving this problem. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-7583720.post-30633485428138352152016-11-13T04:27:04.621-08:002016-11-13T04:27:04.621-08:00Dear Sujit Pal,
I have reading your blog and I a...Dear Sujit Pal, <br /><br />I have reading your blog and I am one your folowers. I always enjoy reading your posts. I always appreciate you. Realy I am proud of you. <br /><br />Next, I want to ask is please help me on MaxEnt(Maximum Entropy) java implementation. I am working on evaluating open questions answer using ontology (domain knowledge) and MaxEnt as evaluation techniques. I am challengedAnonymoushttps://www.blogger.com/profile/11055112451383940885noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-33281420832754119382016-10-25T08:19:08.707-07:002016-10-25T08:19:08.707-07:00Thanks for the kind words, Elias.Thanks for the kind words, Elias.Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-81178246152731623502016-10-25T04:53:55.122-07:002016-10-25T04:53:55.122-07:00I've reading your blog a couple of years now a...I've reading your blog a couple of years now and I wanted to thank you for all the time and quality that you try to put into each post. I really enjoy reading them. Can't wait for the next post ! Elie A.https://www.blogger.com/profile/05255845203944572247noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-48419629608276288872016-10-11T08:04:50.440-07:002016-10-11T08:04:50.440-07:00Thank you for the kind words.
Thank you for the kind words.<br />Sujit Palhttps://www.blogger.com/profile/06835223352394332155noreply@blogger.comtag:blogger.com,1999:blog-7583720.post-3548735253487237402016-10-11T06:27:14.275-07:002016-10-11T06:27:14.275-07:00Interesting post! Thank you for sharing your work....Interesting post! Thank you for sharing your work.Anonymousnoreply@blogger.com