Skip to content
DeepForgeby svx
Beginnernlptokenizationtf-idflanguage-models

NLP Starter

27 problems·~10h·4 stages

Text into numbers, from tokenization and n-grams to TF-IDF, edit distance, and attention. You will build the preprocessing pipeline behind search engines and language models with plain Python. Start here if you want to understand what happens before a model sees a sentence.

Your progress

0/27 · 0%

Path readiness 0%Easy 0% · Medium 0% · Hard 0%

Continue path · Tokenization

What you will learn

  • Tokenize, normalize, and build vocabularies from raw text
  • Represent documents with bag-of-words, TF-IDF, and similarity scores
  • Build n-gram language models with smoothing and subword tokenization
  • Retrieve documents and score generated text

1Tokens and Text Normalization

0/8

Turn raw strings into tokens: whitespace and punctuation handling, stopword removal, n-grams, vocabulary indices, one-hot encoding, and sentence splitting.

Stage checkpoint

0/2 to pass

A checkpoint is a short retrieval test drawn from this stage: solve 2 of 3 from memory to mark the stage complete. The stage stays open either way.

Tried this set and did not pass? Record it — the next set swaps in fresh problems from the same stage.
LabRecommended next stepSpam FilterMedium

2Document Representations

0/7

Documents as vectors: bag-of-words, TF-IDF, cosine similarity, offset tokenization, Jaccard and Dice overlap, and PMI collocation scores.

Stage checkpoint

0/2 to pass

A checkpoint is a short retrieval test drawn from this stage: solve 2 of 3 from memory to mark the stage complete. The stage stays open either way.

Tried this set and did not pass? Record it — the next set swaps in fresh problems from the same stage.
LabRecommended next stepSpam FilterMedium

3Language Models

0/6

Predict the next token: BPE merges and encoding, bigram probabilities with Laplace smoothing, log-likelihood, backoff, and WordPiece segmentation.

Stage checkpoint

0/2 to pass

A checkpoint is a short retrieval test drawn from this stage: solve 2 of 3 from memory to mark the stage complete. The stage stays open either way.

Tried this set and did not pass? Record it — the next set swaps in fresh problems from the same stage.
LabRecommended next stepSpam FilterMedium

4Search and Evaluation

0/6

Find and judge text: edit distance, inverted indexes, BM25, log-likelihood ratio collocations, scaled dot-product attention, and ROUGE-2.

Stage checkpoint

0/2 to pass

A checkpoint is a short retrieval test drawn from this stage: solve 2 of 3 from memory to mark the stage complete. The stage stays open either way.

Tried this set and did not pass? Record it — the next set swaps in fresh problems from the same stage.
LabRecommended next stepSpam FilterMedium

Capstone

Project

Close the path with one real artifact from the catalog.

Build a Search Engine

Take the tokenization and TF-IDF documents from this path all the way to ranked search.