๐Ÿ“– Step 9: AI/LLM#310 / 350

Inference

Inference

๐Ÿ“–One-line summary

The process where a trained model generates output for new input.

๐Ÿ’กEasy explanation

The process of a trained model actually answering questions. If training is 'studying,' inference is 'taking the test.' It happens every time you call an API.

โœจExample

ํ•™์Šต = ๊ณต๋ถ€ ยท ์ถ”๋ก  = ์‹œํ—˜

๐Ÿ“–

Training

์ˆ˜๊ฐœ์›” ยท ๊ณ ๋น„์šฉ

GPU ์ˆ˜์ฒœ ๊ฐœ

โœ๏ธ

Inference

๋ฐ€๋ฆฌ์ดˆ ๋‹จ์œ„

API ํ˜ธ์ถœ๋งˆ๋‹ค ์‹คํ–‰