bloominstituteoftechnology / ds-unit-4-sprint-3-deep-learning Goto Github PK

I meant to post this issue after the sprint to provide example requirement.txt and environment.yml files but slipped my mind, until I received the update on my issue earlier in the unit.

Others and I consistently ran into issues installing with these files out of the box as provided throughout the entire unit 4.

For conda env, use of the --from-history flag will only list user installed conda packages instead of every single package in the env. conda env export --from-history > environment.yml

For .py files, the package pigar will generate a requirements.txt based off of import statements instead of every single package. It also references the files and line the import statement was made in as a comment. pip install pigar then just type and run pigar in terminal.

I have not yet found a good method for dealing with .ipynb as compared to above. Other than converting ipynb to py then use pigar and remove the py files.

There are minor ceavats like psychopg2 typically isnt directly imported due to use of flask-sqlalchemy, thus pigar wont catch it, for example. But it solves the more daunting issue of the dependency conflict nightmare.

These combined should generate more OS agnostic dependency files with less conflicts and errors for students, assuming they even use these files to create the env.

Learning Objectives
At the end of the lecture you should be to:

Part 1: Describe the componenets of an autoencoder
Part 2: Train an autoencoder
Part 3: Apply an autoenocder to a basic information retrieval problem

No 4.3.3 learning objects on main or day page of Training Kit
https://learn.lambdaschool.com/ds/sprint/rec5inpL5t6W30CFH

Currently using MNIST dataset for lecture. Would greatly benefit from using a real-world dataset such as CooperHewitt - or some kind of other highly visual dataset.

In 'LS_DS_432_Convolution_Neural_Networks_Assignment' the 'Starter code has:

from tensorflow.keras.layers import Dense, GlobalAveragePooling2D()

Should drop the parens() at the end of the line to be:

from tensorflow.keras.layers import Dense, GlobalAveragePooling2D

Several typos in this paragraph...making interpretation unclear (e.g. 'expert' is a verb...or not?)

Some typos emphasized:
"Using the Keras functional API, we will need to additional additional full connected layers to our model. We we removed the top layers, we removed all preivous fully connected layers. In other words, we kept only the feature processing portions of our network. You can expert with additional layers beyond what's listed here. The GlobalAveragePooling2D layer functions as a really fancy flatten function by taking the average of each of the last convolutional layer outputs (which is two dimensional still)."

• Include informaiton on annotation tools such as:

• Prodigy
• Microsoft One
  https://github.com/jsbroks/awesome-dataset-tools/