bugfix: Update modeling_t5.T5Stack.forward() for Gradient Checkpointing

bugfix: Add attention mask for generation

Adding `safetensors` variant of this model (#1)

Create README.md

Upload tokenizer

Upload model

initial commit