关于“在 Vertex AI 中使用 Gemini API 执行多模态检索增强生成 (RAG)”的评价

Gabriel Henrique P. · 已于 about 1 year前审核

C.M. J. · 已于 about 1 year前审核

Cedric C. · 已于 about 1 year前审核

Navin S. · 已于 about 1 year前审核

Navin S. · 已于 about 1 year前审核

Vasundhara S. · 已于 about 1 year前审核

Alberto A. · 已于 about 1 year前审核

Good

Coin O. · 已于 about 1 year前审核

Eduardo G. · 已于 about 1 year前审核

Akkamgari S. · 已于 about 1 year前审核

Chethan N. · 已于 about 1 year前审核

While passing image data to gemini 1.5 pro model as stated in the notebook, it says "Unfortunately, I cannot process or "see" images and am unable to provide the analysis requested. I can only process the textual information provided. ". And proceeds to provide answers based on image descriptions only, or suggesting how to get answers. This somewhat contradicts the multimodal capability lab is trying to show. It looks like there is some issue with notebook code, which is converting images to string (just an object repr in python) before sending it to Gemini. So it just sends text instead of actual images.

Nij R. · 已于 about 1 year前审核

Pedro Henrique M. · 已于 about 1 year前审核

nice

Icha Fadhlilah P. · 已于 about 1 year前审核

Airi Fabianne O. · 已于 about 1 year前审核

Sumam P. · 已于 about 1 year前审核

Jeff G. · 已于 about 1 year前审核

at first there was problem with 'import fitz'. after kernel restart the problem went away

Sloppy C. · 已于 about 1 year前审核

Charles D. · 已于 about 1 year前审核

Lúcio L. · 已于 about 1 year前审核

Kayê O. · 已于 about 1 year前审核

ok

Ahmed A. · 已于 about 1 year前审核

Alan Fernando de O. · 已于 about 1 year前审核

Daniel S. · 已于 about 1 year前审核

Reginaldo B. · 已于 about 1 year前审核