kipgparker/soft-prompt-tuning

Issues

• Is padding redundant with the later version of transformers?

  #11 opened a year ago by TinfoilHat0
  0

• How to generate text?

  #7 opened 3 years ago by luke-thorburn
  10

• Prompt tuning significantly slows down code & uses more memory than without?

  #13 opened a year ago by patricks-lab
  0

• question on training loss

  #9 opened 3 years ago by zluw1117
  1

• how could I apply prompt tuning to GPT-3?

  #12 opened 2 years ago by smith-co
  0

• Has anyone implemented initializing prompts for class labels?

  #10 opened 2 years ago by dawei0716
  0

• Input clarification from example

  #8 opened 3 years ago by JosephGatto
  0

• Some question about the "LM Adaptation"

  #1 opened 4 years ago by qcwthu
  1

• What is initialization from vocab for?

  #2 opened 3 years ago by EmElleE
  1

• did it doenst need backpropagation process？

  #4 opened 3 years ago by qyccc
  4

• Some question about "multi-task mixing training"

  #6 opened 3 years ago by czwlines
  0

• the length of n_tokens

  #5 opened 3 years ago by wofeichangaiwoai
  0