This model was derived from the bert-base-uncased checkpoint by replacing the GELU with ReLU activation function and continued pre-training to adapt it to the change of the activation function.

Downloads last month
16
Safetensors
Model size
110M params
Tensor type
I64
·
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Datasets used to train mpiorczynski/relu-bert-base-uncased