Ask HN: Compression Ratio of LLMs?

4 points by FileSorter ↗ HN
Does anyone know what the lossy compression ratio of a LLM is?

((Final .safetensors [GB]) / (Total Training Data [GB])) * 100 = ?

3 comments

[ 6.3 ms ] story [ 21.4 ms ] thread
I believe this wouldn't be meaningful, since any size LLM can be trained on any amount of data.

You could measure how well it memorizes via prediction accuracy on the training set, but this wouldn't indicate whether it generalizes well.

OpenAI’s GPT-3 model (175B) has an archive size of about 350 GB, with training data estimated in the hundreds of terabytes, resulting in a highly compressed ratio.