Developing a Linguistic Realizer
Blog Article Linguistic Realizer
- Definition of a Linguistic Realizer
Step 1: Find the Tokens with Spacy
- Use PAWS / random corpus as base
- Use Spacy to identify all relevant tokens (nouns, pronouns)
Step 2: Create Masked Dataset with Masked Sentences / Target
And Divide into Test, Train, Eval - Set
Step 3: Training!
- Feed Masked Dataset to T5 Model
Step 4: Inference with a Target Set
- Define Target BoW to accept as generated results &
- Find probabilities for generated results for accepting / rejecting.