A Deep Dive into NLP with PyTorch
summary
Hands-on workshop that shows exactly what gets built and how: PyTorch notebooks on GitHub and Colab GPU runtimes, a bag‑of‑words classifier (vocab ~3,000, batch 512), embedding experiments loading pretrained GloVe via magnitude, character‑level generation and small GRU/LSTM sentiment models, and a sequence‑to‑sequence translator with seq2seq with attention and a transformer overview. Models train for just ~10 epochs for demos but still produce plausible outputs; code uses Adam, CountVectorizer and standard PyTorch modules.
AI-written summary · about
chapters (11)
- 0:00NLP with PyTorch intro
- 2:19PyTorch basics overview
- 9:53Feed-forward neural network
- 24:15NLP text preprocessing
- 35:29Bag-of-words classifier
- 44:29Word embeddings training
- 50:59Recurrent neural networks for sequences
- 1:01:16GRU and LSTM in PyTorch
- 1:11:28Sequence-to-sequence models with attention
- 1:26:17Transformers architecture
- 1:35:44Training for translation
related talks
01h 32mIntroduction to PyTorch
01h 30mIntroduction to Deep Learning & Natural Language Processing
01h 26mDeep Neural Networks with PyTorch
01h 59mDeep Learning and Modern NLP
01h 27mHands-On Computer Vision with PyTorch
17mF8 Refresh | Deep Learning with PyTorch
01h 32mNeural Networks for Natural Language Processing
41mGPU Training Made Easy-Pre-train & fine-tune NLP models