Изгради го својот вокабулар за кодирање во Пајтон и истражување на вештачка интелигенција! Научи да зборуваш за структурата на кодот, библиотеките и новите технолошки достигнувања, додека ги подобруваш своите програмски вештини. Идеално за средно напредни учесници кои сакаат да дискутираат за технолошки теми пофлуентно.
Stories
Lena Optimizes Python Algorithm
English
Lena loves writing code in Python. She has been working on an AI project for a large company, but her algorithm hasn't been optimized yet. "Have you debugged the program yet?" her team asked last week. She said she had processed the datasets efficiently, but the model still needed work. Despite being complex, the project was not impossible. They used an outdated library, so she had to implement a new one. "May we train the models quickly?" she asked. The scientists confirmed they could. Every day, she writes code to analyze data and improve features. He used to debug manually, but now they automate the process. Your project? It might be next!
Neural Network Improves Over Time
English
The scientists were excited about their new neural network. They had built it last month to automate repetitive tasks, but the initial results weren’t accurate. "Have you preprocessed the data correctly?" she asked. He admitted the dataset wasn’t analyzed properly before training. Despite their difficulties, they decided to refactor the old code. "We must fine-tune the parameters," he said. They integrated new APIs and deployed the model on the server. Even though errors crashed the program at first, it now handles them gracefully. The results are better this week. "Were the tasks automated yet?" she wondered. Next, they will visualize the data to improve accuracy.
Model Training Challenges
English
The team was excited—they had just trained a new classifier. The data had been split into training and test sets, and the features were encoded manually. Despite their efforts, the model was overfit. "Have you evaluated the performance yet?" asked the lead engineer. She confirmed the results were incomplete. "We must normalize the data," she said. They scaled the variables, but the model still underfit. "May we try regularization?" someone suggested. By morning, the model had been reevaluated. The normalization improved accuracy, but the performance was not high enough. "Your dataset might be too small," he warned. "Has the model been loaded successfully?" They visualized the results and found a mistake—some features had not been scaled properly. "Did you check the encoding?" she asked. The team used to encode data automatically, but now they did it manually. Despite the challenges, they were determined to improve the model. "We must train it again," she said. The classifier has been working since yesterday, but the results have not improved yet. "Have you tried adjusting the hyperparameters?" he asked. The team knew they had to be patient—good models take time.
Sample flashcards 20
да пишува код→to write code
Таа пишува код секој ден.→She writes code every day.
Дали тој пишува код за своите проекти?→Does he write code for his projects?
Тие не пишуваат код во Пајтон.→They do not write code in Python.
дебагира програма→to debug a program
Дали сте го дебагирале програмот уште?→Have you debugged the program yet?
Програмот сè уште не е дебагиран.→The program hasn't been debugged yet.
да користи библиотека→to use a library
да користи Пајтон библиотека→to use a Python library
Може ли да ја користиме оваа Пајтон библиотека за нашиот проект?→May we use this Python library for our project?
Вашиот тим не смее да ја користи таа застарена библиотека.→Your team may not use that outdated library.
да тренира модел→to train a model
Научниците тренираат AI модели со големи податочни множества.→Scientists train AI models with large datasets.
Дали моделот е веќе обучен?→Has the model been trained yet?
да оптимизира алгоритам→to optimize an algorithm
да го оптимизира алгоритамот→to optimize the algorithm
Тој рече дека го оптимизирале алгоритамот.→He said they had optimized the algorithm.
Иако е сложен, алгоритамот беше брзо оптимизиран.→Despite being complex, the algorithm was optimized quickly.
да имплементира функција→to implement a feature
Порано ги имплементиравме функциите рачно.→We used to implement features manually.
A sample of the collection — the full set lives in the app.
Learn this topic in Taalhammer
Spaced repetition, pronunciation and audio — in the app.
