Synchromesh: Reliable Code Generation from Pre-trained Language Models.
Gabriel PoesiaAlex PolozovVu LeAshish TiwariGustavo SoaresChristopher MeekSumit GulwaniPublished in: ICLR (2022)
Keyphrases
- language model
- code generation
- pre trained
- language modeling
- application development
- software development
- n gram
- probabilistic model
- training examples
- training data
- language modelling
- speech recognition
- information retrieval
- modeling language
- retrieval model
- query expansion
- statistical language models
- model driven
- software reuse
- formal specification
- rapid prototyping
- control signals
- relevance model
- smoothing methods
- language models for information retrieval
- design tools
- appearance variations
- object oriented
- training set
- data sets