medium10 pts · first answer countsmodel-servingtoolingarchitecture
Pre-processing on the client is slow
Clients send raw text, then a Python service tokenises, calls Triton, and post-processes, adding many network hops. Which Triton feature can run all steps server-side?
Discussion 0
Join free to post in the discussion.
Related questions
- When batch inference fitsMLOps Engineer
- Traffic before the model has loadedMLOps Engineer
- Why export a model to ONNXMLOps Engineer
- Protocol for internal inference callsMLOps Engineer
- Latency and throughputMLOps Engineer
- What TTFT measuresMLOps Engineer
- Gap between streamed tokensMLOps Engineer
- Two phases of LLM generationMLOps Engineer