microsoft / deep-language-networks

We view Large Language Models as stochastic language layers in a network, where the learnable parameters are the natural language prompts at each layer. We stack two such layers, feeding the output of one layer to the next. We call the stacked architecture a Deep Language Network - DLN
MIT License
87 stars 12 forks source link

Fix dev test seed #47

Closed matheper closed 5 months ago

matheper commented 5 months ago

Fix seed when resizing and batching dev and test sets. Loading the dataset was already using a fixed seed.

MarcCote commented 5 months ago

That's great. Thanks @matheper