Ceshine Lee[Notes] (Ir)Reproducible Machine Learning: A Case StudyI just read this (draft) paper named “(Ir)Reproducible Machine Learning: A Case Study” ( blog post; paper). It reviewed 15 papers that…Jan 17, 2022Jan 17, 2022
Ceshine LeeHow to Create a Documentation Website for Your Python PackageUse Sphinx to (semi-)automatically generate documentation from docstringsDec 23, 20211Dec 23, 20211
Ceshine LeeinVeritable[Notes] Gradient Checkpointing with BERTA brief analysis of huggingface’s implementationJun 18, 2021Jun 18, 2021
Ceshine LeeinVeritable[Paper] Adafactor: Adaptive Learning Rates with Sublinear Memory CostEssential for fine-tuning T5 v1.1 and mT5 modelsApr 18, 2021Apr 18, 2021
Ceshine LeeinVeritable[Paper] Rethinking Cooperative Rationalization: Introspective Extraction and Complement ControlBuilding competitive self-explaining NLP modelsMar 24, 2021Mar 24, 2021
Ceshine LeeinVeritableReducing the SentencePiece Vocabulary Size of Pretrained NLP ModelsUseful for fine-tuning on a subset of available languagesFeb 22, 2021Feb 22, 2021
Ceshine LeeinVeritable[PyTorch Lightning] Log Training Losses when Accumulating GradientsThe global step is not what you think it isJan 23, 2021Jan 23, 2021
Ceshine LeeinVeritableGenerating Synthetic Tabular Data Using GANA case study: detecting credit fraudDec 30, 2020Dec 30, 2020
Ceshine LeeinVeritable[Paper] Are We Really Making Much Progress?A Worrying Analysis of Recent Neural Recommendation ApproachesDec 9, 2020Dec 9, 2020
Ceshine LeeinVeritableAutomatic Testing Your SQLite Database with Great ExpectationsA great tool for eliminating pipeline debtsNov 21, 2020Nov 21, 2020
![[Notes] (Ir)Reproducible Machine Learning: A Case Study](https://miro.medium.com/v2/resize:fill:160:106/0*I74gvIAJBmSM7JSW.jpg)
![[Notes] (Ir)Reproducible Machine Learning: A Case Study](https://miro.medium.com/v2/resize:fill:320:214/0*I74gvIAJBmSM7JSW.jpg)
![[Notes] Gradient Checkpointing with BERT](https://miro.medium.com/v2/resize:fill:160:106/0*JEiyRk0EIhEOsB5u.jpg)
![[Notes] Gradient Checkpointing with BERT](https://miro.medium.com/v2/resize:fill:320:214/0*JEiyRk0EIhEOsB5u.jpg)
![[Paper] Adafactor: Adaptive Learning Rates with Sublinear Memory Cost](https://miro.medium.com/v2/resize:fill:160:106/0*cYtYMg4UDRu6NKuc.jpg)
![[Paper] Adafactor: Adaptive Learning Rates with Sublinear Memory Cost](https://miro.medium.com/v2/resize:fill:320:214/0*cYtYMg4UDRu6NKuc.jpg)
![[Paper] Rethinking Cooperative Rationalization: Introspective Extraction and Complement Control](https://miro.medium.com/v2/resize:fill:160:106/0*9iK2rLh3qGeaSm57.jpg)
![[Paper] Rethinking Cooperative Rationalization: Introspective Extraction and Complement Control](https://miro.medium.com/v2/resize:fill:320:214/0*9iK2rLh3qGeaSm57.jpg)
![[PyTorch Lightning] Log Training Losses when Accumulating Gradients](https://miro.medium.com/v2/resize:fill:160:106/0*yimj7gw-9iCdEzdP.jpg)
![[PyTorch Lightning] Log Training Losses when Accumulating Gradients](https://miro.medium.com/v2/resize:fill:320:214/0*yimj7gw-9iCdEzdP.jpg)
![[Paper] Are We Really Making Much Progress?](https://miro.medium.com/v2/resize:fill:160:106/0*dX05hYzIDOrWWb76.jpg)
![[Paper] Are We Really Making Much Progress?](https://miro.medium.com/v2/resize:fill:320:214/0*dX05hYzIDOrWWb76.jpg)
