Module: Natual Language Processing.pdf (presentation). slide # 64
More coming up on CloudxLab
? Word2vec - Vector Representations of Words - Covered in RNN
? Deep Learning - LSTM - Long Short-Term Memory - Covered in RNN
? GloVe - Global Vectors for Word Representation
? spaCY - Industrial-Strength Natural Language Processing in Python
? Hands-on using Stanford CoreNLP
? List of APIs available for chatbots etc
Where can I find ppts and notebooks for the remaining topics (last 4)?
Thanks Rajtilak. Is there any planned date by which the topics will be available? What if the course duration is completed before the new topics are added? The access will be provided to learners for the new topics?
Unfortunately, as of now there are no planned dates. However, if any part of this course is updated you will find new content within your playlist. The course materials will be available to you for a lifetime.
use_idf can be set to True or False. Could you please explain how and where "use_idf" parameter will impact the model? I would also like to know the different scenarios and the value to be used for use_idf (if possible).
use_idf=True enables inverse-document-frequency reweighting. The inverse document frequency is a measure of whether a term is common or rare in a given document corpus. It is obtained by dividing the total number of documents by the number of documents containing the term in the corpus.
If height and width dimensions are same then--
In CNN we calculate the next dimension(if padding is valid) by the formula=> (n-2p-(f-1))/stride ; here n=height=width, f=filter size
I want to know if the dimensions are different say height not equals to width, and we apply no padding then what will be the formula for calculating the next dimensions on applying filter
I just want to know how we can use the other state of the art CNN architectures like resnet,VGG etc, in tensorflow by importing them like we can do in the Keras.
If not can we use keras for importing the architectufes and then further use it in tensorflow.
I believe we have covered this in Deep Learning lecture when we covered pre-trained layers. Would request you to go back and check. Also, just so that you know, we mostly do not use vanilla Keras. We mostly use Keras with a Tensorflow (or some other library like Theano) backend. Only in rare cases do we use vanilla Keras.
but it was mainly for the model which was made in tensorflow only but it did not give any example of the state of the art architectures such as Resnet,VGG etc.
So., can you please mention some of the links so that i can go through it thoroughly.
Yes sir it's been uploaded, but they don't match with the slides on the video. I think there is a next sequel on word2vec (word embedding) and seq2seq( machine translation).
Please login to comment
40 Comments
Hi,
slides for NLP and video content doesn't match. If there is new video for NLP kindly share link.
Upvote ShareI have added the NLP chapter separately now.
Upvote ShareMy jupyter lab is resetting more frequently when running the project of tensorflow
Upvote ShareHi,
Please check our FUP:
https://cloudxlab.com/faq/6/what-are-the-limits-on-the-usage-of-lab-or-what-is-the-fair-usage-policy-fup
Also, check out the below discussion:
https://discuss.cloudxlab.com/t/explained-code-taking-too-long-to-execute/5304
Thanks.
Upvote ShareWhile running tensorflow in kaggle.com i am getting error tensorflow has no module placeholder
Upvote ShareHi,
Please check if Kaggle has Tensorflow 1 installed.
Thanks.
Upvote ShareHi,
Module: Natual Language Processing.pdf (presentation). slide # 64
More coming up on CloudxLab
? Word2vec - Vector Representations of Words - Covered in RNN
? Deep Learning - LSTM - Long Short-Term Memory - Covered in RNN
? GloVe - Global Vectors for Word Representation
? spaCY - Industrial-Strength Natural Language Processing in Python
? Hands-on using Stanford CoreNLP
? List of APIs available for chatbots etc
Where can I find ppts and notebooks for the remaining topics (last 4)?
Regards,
punit
Upvote ShareHi,
We do not have any slides as of now, however, we are in the process of revamping the entire courseware where we plan to include them.
Thanks.
Upvote ShareThanks Rajtilak. Is there any planned date by which the topics will be available? What if the course duration is completed before the new topics are added? The access will be provided to learners for the new topics?
Regards,
Punit
Upvote ShareHi,
Unfortunately, as of now there are no planned dates. However, if any part of this course is updated you will find new content within your playlist. The course materials will be available to you for a lifetime.
Thanks.
Upvote ShareHi,
Notebook - sklearn_text_analyser.ipynb
Section - From occurrences to frequencies
from sklearn.feature_extraction.text import TfidfTransformer
tf_transformer = TfidfTransformer(use_idf=False).fit(X_train_counts)
X_train_tf = tf_transformer.transform(X_train_counts)
X_train_tf.shape
use_idf can be set to True or False. Could you please explain how and where "use_idf" parameter will impact the model? I would also like to know the different scenarios and the value to be used for use_idf (if possible).
Regards,
Punit
1 Upvote ShareHi Punit,
Good question!
Please find an explanation from the below link:
https://stackoverflow.com/questions/34847210/what-exactly-does-use-idf-do-when-creating-a-tfidftransformer-in-sklearn
Thanks.
1 Upvote ShareThanks Rajtilak. It would be helpful if you can answer following query, which is not available not the link.
I would also like to know the different scenarios and the value to be used for use_idf.
Regards,
Punit
Upvote ShareHi,
use_idf=True enables inverse-document-frequency reweighting. The inverse document frequency is a measure of whether a term is common or rare in a given document corpus. It is obtained by dividing the total number of documents by the number of documents containing the term in the corpus.
Thanks.
Upvote SharePlease tell suggest me some more DATA AUGMENTATION TECHNIQUES TO INCREASE number of 3D IMAGES
Upvote Shareapart from POSITION, SIZE and COLOUR.
Hi,
Please go through the below discussion:
https://machinelearningmastery.com/how-to-configure-image-data-augmentation-when-training-deep-learning-neural-networks/
Thanks.
Upvote ShareHi,
If height and width dimensions are same then--
In CNN we calculate the next dimension(if padding is valid) by the formula=> (n-2p-(f-1))/stride ; here n=height=width, f=filter size
I want to know if the dimensions are different say height not equals to width, and we apply no padding then what will be the formula for calculating the next dimensions on applying filter
Upvote ShareHi,
Here is the formula for the same:
https://www.quora.com/How-can-I-calculate-the-size-of-output-of-convolutional-layer/answer/Rockson-Agyeman
Thanks.
Upvote ShareAccording to me NLP is more easier when done using Keras as compared to tensorflow isn't it ?
Upvote ShareHi,
Yes, Keras is earsier than TF1.0. But then, this is my personal opinion.
Thanks.
Upvote ShareHi,
The NLP vedio is missing please add it too ASAP.
Only Autoencoder is covered in this not NLP.
THANKS.
Upvote ShareThis comment has been removed.
Hi,
NLP is covered in this video itself.
Thanks.
Upvote ShareI don't think it is fully covered because the slides includes tfidf and many more things but they aren't told in the vedio.
Please have a look.
Upvote ShareHi,
You are correct. Not all concept have been covered in the video, but we have covered them in the slides.
Thanks.
Upvote ShareYes, so what to do now??
Upvote ShareThis comment has been removed.
Hi,
Please study the slides as we did in Python, Numpy and some of the other topics, along with the codes from the notebook.
Thanks.
Upvote ShareHi,
I just want to know how we can use the other state of the art CNN architectures like resnet,VGG etc, in tensorflow by importing them like we can do in the Keras.
If not can we use keras for importing the architectufes and then further use it in tensorflow.
Please clarify my doubt.
THANKS.
Upvote ShareThis comment has been removed.
Hi,
I believe we have covered this in Deep Learning lecture when we covered pre-trained layers. Would request you to go back and check. Also, just so that you know, we mostly do not use vanilla Keras. We mostly use Keras with a Tensorflow (or some other library like Theano) backend. Only in rare cases do we use vanilla Keras.
Thanks.
Upvote ShareYes i just took a look at those slides,
but it was mainly for the model which was made in tensorflow only but it did not give any example of the state of the art architectures such as Resnet,VGG etc.
So., can you please mention some of the links so that i can go through it thoroughly.
THANKYOU..
Upvote ShareHi,
I have a query that i want to work on a dataset and it is 6GB in size can i upload it into my cloudxlab account and work on it?????
Upvote ShareHI, Anubhav.
Kindly refer to this usage and limits of your lab : https://cloudxlab.com/faq/6/what-are-the-limits-on-the-usage-of-lab-or-what-is-the-fair-usage-policy-fup#:~:text=Questions%20and%20Answers-,What%20are%20the%20limits%20on%20the%20usage%20of%20lab%20or,for%20educational%20and%20PoC%20purposes.&text=Here%20are%20the%20limits%20as,the%20replication%20factor%20of%203.
All the best!
Upvote ShareHi,
I am not able to access my jupyter notebook it is giving 500 error everytime since last some days.
Please eradicate this issue.
Upvote ShareHi Anubhav,
Your disk space quota has exceeded 3GB limit. Please follow instructions given here
Upvote ShareThe slides on NLP are not uploaded. Where can I get them?
Upvote ShareHi,
I have added the slides to NLP, could you please check and confirm.
Thanks.
Upvote ShareYes sir it's been uploaded, but they don't match with the slides on the video. I think there is a next sequel on word2vec (word embedding) and seq2seq( machine translation).
Thank you.
Upvote ShareHi,
There are constant updates. Have you checked the latest notebooks from our GitHub repository? Let me know if you cannot find it there.
Thanks.
Upvote Share