We start a company-name recognition task with a small scale and low quality training data, then using skills to enhanced model training speed and predicting performance with least artificial participation. The methods we use involve lite pre-training models such as Albert-small or Electra-small with financial corpus, knowledge of distillation an…
tf-keras code of Face Ear Landmark Detection System (with Multi-Task Learning).
A face recognition system with low computational cost
Implemen Distilling the knowledge in a neural network with TensorFlow
Easily generate synthetic data for classification tasks using LLMs
基于亚马逊一页纸写作原则 & MBB咨询方法论，把任何文本转换为一页纸形式（HTML&PDF）；An agent skill that compresses anything into one page worth reading.
This repo is all the code that I used to participant in Prostate Cancer Grade Assessment contest on Kaggle
To associate your repository with the distilling-the-knowledge topic, visit your repo's landing page and select "manage topics."