nlp & cv lab

by Manas Pandya

fifteen lab tasks in nlp and computer vision. models are trained in pytorch, exported to onnx and run right here in the browser, so nothing gets uploaded anywhere.

code on github

• 01 / multitodo

  image caption generation

  describe an image in one sentence.

• 02 / nlptodo

  seq2seq translation

  translate text between languages.

• 03 / nlptodo

  text summarization

  shorten long text with a transformer.

• 04 / nlptodo

  question answering

  answer questions from a passage.

• 05 / nlptodo

  grammar error correction

  find and fix grammar mistakes.

• 06 / nlplive

  text emotion classification

  28-way multi-label emotions, running in your browser.

  try the demo →

• 07 / cvlive

  bilateral image denoising

  classic and learned bilateral filters, live on the gpu.

  try the demo →

• 08 / cvtodo

  eye detection & facial landmarks

  locate eyes and face keypoints.

• 09 / cvtodo

  human activity recognition

  classify actions from video frames.

• 10 / cvtodo

  vehicle detection & counting

  detect and count vehicles in traffic video.

• 11 / multitodo

  visual question answering

  ask questions about an image.

• 12 / cvtodo

  table detection & extraction

  find tables in documents and pull out the data.

• 13 / multitodo

  signboard text recognition & translation

  read sign text and translate it.

• 14 / multitodo

  product recognition & description

  recognise a product and write a description.

• 15 / multitodo

  multimodal sentiment analysis

  sentiment from text and image together.