spinf-serve releases open-source batch scoring inference server under Apache 2.0
What happened
spinf-serve released version 0.1 of its inference server under the Apache 2.0 license, reported on the Hacker News front page on October 31, 2026. The server targets batch throughput: it answers multiple questions from a single read of the content and returns probabilities for each option, for batch scoring and decisions. It also extracts text, numbers and amounts from the same read. It runs on a single 24 GB NVIDIA GPU and offers an OpenAI-compatible API.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageShow HN: Spinf-serve – inference server for bulk scoring and decisions
spinf-serve 发布 0.1 版推理服务器，以 Apache 2.0 开源，单次读取内容即可回答多个问题并返回各选项的概率。它支持单张 24 GB NVIDIA GPU 部署和 OpenAI 兼容 API，面向批量吞吐，并支持从同次读取中提取文本、数字和金额。
Heat over time
Not enough continuous observations to draw a trend yet.