Instructor v1.15.4
Instructor v1.15.4 is a patch release addressing response-model normalization, JSON stream extraction bug fixes, image autodetect handling, and CLI fine-tuning file ID resolution.
原始内容为英文；当前页面提供中文导航与来源说明，具体事实请以原文为准。
为什么值得关注
Instructor 1.15.4 delivers several runtime and CLI fixes: response-model normalization resolves AttributeError crashes when using list or primitive models; streamed JSON extraction preserves backticks within string values rather than parsing them as markdown fences; Image.autodetect() supports raw JPEG, PNG, GIF, and WebP bytes; list BaseModel detection correctly routes through iterable handling; and CLI fine-tuning now correctly submits uploaded validation file IDs instead of local paths to OpenAI. Additionally, docs and default model strings were refreshed.
可执行摘要
Instructor v1.15.4 fixes runtime crashes with primitive/list response models, stops backtick misinterpretation in streamed JSON, enhances Image.autodetect() byte handling, and corrects CLI fine-tuning validation file parameter passing.
- Agent 实用度
- 82/100
- 可信度
- 96%
- 机器格式
- JSON + Markdown
开发者应核对什么
- Upgrade instructor to version 1.15.4 via PyPI (pip install -U instructor).
- Verify pipelines relying on list BaseModel or primitive scalar response models to ensure normalization works as expected.
- Ensure any workflows passing raw image bytes to Image.autodetect() conform to supported formats (JPEG, PNG, GIF, WebP).