PinnedThomas DehaeneinML6teamA Natural Language Processing look at the Belgian governmental agreementA tale of multilingualism, summarization and sentence embeddingsNov 10, 2020Nov 10, 2020
Thomas DehaeneinML6teamBERT is eating your cash: quantization and ONNXRuntime to save💰In 2020, we have trained and open-sourced the first Dutch GPT2 model, in various sizes. Of course we wanted to share this with the world by…May 18, 20222May 18, 20222
Thomas DehaeneinML6teamDutch GPT2: Autoregressive Language Modelling on a budgetFinetuning pretrained English GPT2 models to Dutch with the OSCAR dataset, using Huggingface transformers and fastai.Dec 18, 2020Dec 18, 2020
Thomas DehaeneinTowards Data ScienceGame of Thrones Twitter Sentiment with Keras, Apache Beam, BigQuery and PubSubThe final season of Game of Thrones apparently made a lot of eyebrows frown, so I wanted to dig deeper on how people felt before, during…May 22, 20195May 22, 20195
Thomas DehaeneinTowards Data ScienceAdaptive - and Cyclical Learning Rates using PyTorchThe Learning Rate (LR) is one of the key parameters to tune. Using PyTorch, we’ll check how the common ones hold up against CLR!Mar 20, 20195Mar 20, 20195
Thomas DehaeneinTowards Data ScienceAn NLP View on Holiday Movies — Part II: Text Generation using LSTM’s in KerasLet’s see if we can train an RNN with the input sequences to create a new AI-generated Christmas movie.Dec 31, 20182Dec 31, 20182
Thomas DehaeneinTowards Data ScienceAn NLP View on Holiday Movies — Part I: Topic Modeling using Gensim and SKlearnHolidays are a time for family, friends, snow and as far as my wife is concerned: corny holiday movies.Dec 31, 20182Dec 31, 20182
Thomas DehaeneinBecoming Human: Artificial Intelligence MagazineVariational AutoEncoders for new fruits with Keras and Pytorch.There’s two things you typically love being a Data Scientist at FoodPairing: Machine Learning and food (order up for debate…). So when you…Nov 7, 20187Nov 7, 20187