t5_1.1_large / README.md
Divyasreepat's picture
Upload folder using huggingface_hub
e7c0f5d verified
|
raw
history blame
735 Bytes
metadata
library_name: keras-hub

This is a T5 model uploaded using the KerasHub library and can be used with JAX, TensorFlow, and PyTorch backends. Model config:

  • name: t5_backbone
  • trainable: True
  • vocabulary_size: 32128
  • hidden_dim: 1024
  • intermediate_dim: 2816
  • num_layers: 24
  • num_heads: 16
  • activation: gelu
  • key_value_dim: 64
  • dropout: 0.1
  • use_gated_activation: True
  • layer_norm_epsilon: 1e-06
  • tie_embedding_weights: False

This model card has been generated automatically and should be completed by the model author. See Model Cards documentation for more information.