SebOchs
/

xtremedistil-l6-h256-uncased-squad

Question Answering

Inference Endpoints

Model card Files Files and versions Community

Test model for DL4NLP 2022 HW06

xtremedistil-l6-h256-uncased trained on SQuAD

Hyper parameters

learning rate: 1e-5
weight decay: 0.01
warm up steps: 0
learning rate scheduler: linear
epochs: 1

Metric results on the dev set

F1: 65.48
EM: 51.67

Downloads last month: 108

Inference Providers NEW

Question Answering

This model is not currently available via any of the supported Inference Providers.