1개 답변
- 최신
- 최다 투표
- 가장 많은 댓글
0
You can do local testing by running the container in serve mode as a docker. Then using Curl/Postman to send an HTTP request and inspecting the response.
The request can be CSV/JSON or binary (a parquet file in your case).
If you're able to run the Pytorch model in serve mode locally, then this local testing provides a lot of coverage before running in Batch Transform itself.
답변함 5년 전
관련 콘텐츠
- AWS 공식업데이트됨 일 년 전