girafe-ai/ml-course

Open Machine Learning course

ml-course's Issues

Typo in generative models presentation

https://github.com/ml-mipt/ml-mipt/blob/advanced/week15_generative/week15_Generative_models.pdf

First slide should say "advanced", not "advansed".

Thanks for such awesome material!

ml-learning

Wrong link in 23f_yandex_ml_trainings/homeworks/assignment01_knn/knn_assignment_0_01.ipynb

Instead, this worked for me:

'''
If you are using Google Colab, uncomment the next line to download `k_nearest_neighbor.py`. 
You can open and change it in Colab using the "Files" sidebar on the left.
'''
!wget https://raw.githubusercontent.com/girafe-ai/ml-course/23f_yandex_ml_trainings/homeworks/assignment01_knn/k_nearest_neighbor.py

may you help to find code for Прикладное машинное обучение. Семинар 5. BERT for text classification

it is written for youtube video you have code for this videos on https://github.com/girafe-ai/ml-course
but there is not
may you help to find code ?
https://www.youtube.com/watch?v=F3jsMAI5EF4&t=6s
https://www.youtube.com/watch?v=1DygevyV2eA&t=656s

Link to slides on backprop sem

In table Расписание Курса
Link to Marchenkoff presentation on his sem on backprop is broken. The pres I consider is shown by him in the video of the backprop sem, there are a lot of formulas on the NN from the 5th homework.

Add 12 and 13 material of Lectures

Branch have not material of 12 and 13 lecture

Wrong link for ML handbook in README.md

Current link gives error 404.

Correct link is https://education.yandex.ru/handbook/ml6

Incorrect variable name in hw02_laplace

ml-course/homeworks/hw02_laplace/distribution.py

Line 46 in 16d098a

return np.exp(self.logpdf(value))

I think, return np.exp(self.logpdf(values)) should be here (because we have values in function args)

Confusing useless loop in naive bayes

ml-course/week0_01_naive_bayes/week0_01_01_naive_bayes__completed.ipynb

Line 744 in 8d2b127

" for idx in range(X.shape[1]):\n",

We get correct results after the first iteration. During all other iterations logsumexp returns vector of zeros, because probabilities are already normalized

fix formula in week0_01_01_naive_bayes.ipynb Probabilistic model

P(\mathbf{x}i) = \sum{k=1}^K P(y_i = C_k) P(\mathbf{x}_i|y_i=C_k) should be P(\mathbf{x}i) = \sum{k=1}^K P(\mathbf{x}_i|y_i=C_k), as was stated in video lecture.

Legacy imports in Lab02_NMT

from torchtext.datasets import TranslationDataset, Multi30k
from torchtext.data import Field, BucketIterator

Этот код стал легаси два года назад, возможно пришло время от него избавиться. Сверх того, первые два импорта нигде не используются

Branch 23s_advanced has README from 22f_basic

Exam program doesn't exist in branch 23f_basic

Word embeddigns presentation

В презентации явно не те иллюстрации, что нужны для CBOW и Skip-gram
Презентация:
https://github.com/girafe-ai/ml-course/blob/23s_advanced/week01_word_embeddings/MADE__NLP_01_Word_embeddings.pdf
Оригинальная статья:
https://arxiv.org/pdf/1411.2738.pdf

what video is for week0_02_Linear_regression_and_SGD.ipynb

looks really great
may you share what video is for
https://github.com/girafe-ai/ml-mipt/blob/master/week0_02_linear_reg/week0_02_Linear_regression_and_SGD.ipynb