• DanielHesslow's avatar
    Add Gated-SiLU to T5 (#17420) · 607acd4f
    DanielHesslow authored
    
    
    * Add gated-silu to t5 architecture to support UL2
    
    * Fix error message
    
    * formatting
    
    * formatting again
    
    * refactor
    
    * fix classnames in _init_weights
    
    * remove is_gated
    
    * add test
    
    * fix test
    
    * Try without the test?
    
    * Add back the test.
    
    * Improve error message.
    Co-authored-by: default avatarDaniel Hesslow <daniel@lighton.ai>
    607acd4f
test_modeling_t5.py 62.3 KB