Hmm just asking, this is for supervise learning? Looks like fasttext is applicable for unsupervised too, but... not mentioned in video... Great video though.
I am having trouble while training the FastText model for 1 GB data as it is taking too long to train on that data, my parameters for Fasttext are: sg=0,size=200, iter=100, window=16 is there any way to make the training quicker. I am using gensim Fasttext.
The process you are explaining around 6min mark. Are you saying that there are 20k input neurons each representing a word and the same 20k output neurons. So that when you are training for Input: Bakshi, Output: is the inputs would be [0, 0, 0, 0, ... 1, ... 0] (where only the neuron representing bakshi is a 1)? And the same for the outputs where the "is" neuron is 1? Thank you
@@tanmayteaches Thank you, I didn't expect you to reply to an over 1 year old video. I was also wondering about the procedure around 12:50, are you saying that whether the network receives "matrix" or "mat" or "at" the network will assume these are the same and have one neuron represent them all. Or are you maybe saying the opposite that instead of having one neuron for "matrix" it is going to add like 10 more neurons for all the variations?
Hello - Neural networks essentially need to analyze a word and predict the contents of the context window. In doing so, it learns a semantic vector representation of words across the vocabulary.
Thanks for asking! I believe this answer should help you out: stackoverflow.com/questions/58337469/how-to-save-fasttext-model-in-vec-format Let me know if that doesn’t work and I can help you out further.
You are great at explaining these concepts. I got a lot out of this video. Thank you!
Well Explained. I learned almost all these concepts from different sources. But you combined them. Nice effort. Keep it up.
Hmm just asking, this is for supervise learning? Looks like fasttext is applicable for unsupervised too, but... not mentioned in video...
Great video though.
Great explaination of word2vec and fasttext. Thanks for your work.
awsm!! u have very clear concepts, thats amazing.
Could u suggest me good websites for python?!
Hey Tanmay , please suggest me which skills I need to become a better comuter science engineer ??
incredible presentation
Superb bro
Love
Form , India (Hindustan)
how can i construct sent2vec from facebook word2vec model?
I am having trouble while training the FastText model for 1 GB data as it is taking too long to train on that data, my parameters for Fasttext are:
sg=0,size=200, iter=100, window=16
is there any way to make the training quicker. I am using gensim Fasttext.
The process you are explaining around 6min mark. Are you saying that there are 20k input neurons each representing a word and the same 20k output neurons. So that when you are training for Input: Bakshi, Output: is the inputs would be [0, 0, 0, 0, ... 1, ... 0] (where only the neuron representing bakshi is a 1)? And the same for the outputs where the "is" neuron is 1? Thank you
Yes, that is correct. The inputs & outputs are one-hot encoded. This may seem like a lot, and it is, but the method does work.
@@tanmayteaches Thank you, I didn't expect you to reply to an over 1 year old video. I was also wondering about the procedure around 12:50, are you saying that whether the network receives "matrix" or "mat" or "at" the network will assume these are the same and have one neuron represent them all. Or are you maybe saying the opposite that instead of having one neuron for "matrix" it is going to add like 10 more neurons for all the variations?
Thanks sir...... Can u plz tell us about yourself plz..... And how we all can work with you ..... Thanks again
Is fasttext using hot encoding too?
thankyou so much Tanmay ...u really inspire me and motivate me
✌👍
riya
Fantastic video 👏🏻
Awesome
Please explain BERT
Please can I have contact with you Tanmay please
I want your help
What is the work of neural networks on Context window?
Hello - Neural networks essentially need to analyze a word and predict the contents of the context window. In doing so, it learns a semantic vector representation of words across the vocabulary.
Maybe not related but how old are you? :) you look like a teenager!
Where is the dataset?
you are really very good ....never seen boy like u ....👍
Absolutely true 😘😘❤💕💖💞💝😇
oh yeah👍✌
I am like your little subscriber
How did you go from bin file to vec file?
Thanks for asking! I believe this answer should help you out: stackoverflow.com/questions/58337469/how-to-save-fasttext-model-in-vec-format
Let me know if that doesn’t work and I can help you out further.
@@tanmayteaches Thank you very much for the reply. Your channel is awesome.
Hi
Hi Tanmay what's up
Kool.