Fine-tuning for writing code from prompts
#2
by
Johnguitarist
- opened
Instead of using this model for code completion in a causal LM context, how might I use it to write code from prompts?
Example input: 'function to scale numerical features in a dataframe'
would return a python function to perform this.
I've been trying to fine-tune this for a while now but it seems CausalLM isn't the mode I was expecting,
What imports and steps would I need to achieve results like I described above?
Many thanks!
bomp