71-vueT10

/

chess-guesser-distilbert

like 1

I can allocate an RTX A6000 to the task, and I already have clean formatted PGN games in Parquet format. My current task is building an "annotator" that can speak and reason about a chess position, and I can see this model being useful in my experiments.

I've never fine-tuned an LLM before, only trained chess neural nets to play, so this is of special interest to me as a learning project.