Multimodal Retrieval Augmented Generation (RAG) using the Gemini API in Agent Platform Reviews

Somesh S. · Reviewed 3 месяцев ago

Vivek N. · Reviewed 3 месяцев ago

Divya R. · Reviewed 3 месяцев ago

Amritha V. · Reviewed 3 месяцев ago

Rohith V. · Reviewed 3 месяцев ago

Sarun K. · Reviewed 3 месяцев ago

Anuram V. · Reviewed 3 месяцев ago

Jenny Jolly M. · Reviewed 3 месяцев ago

Pranav B. · Reviewed 3 месяцев ago

Gowthaman N. · Reviewed 3 месяцев ago

Anisetti R. · Reviewed 3 месяцев ago

Sau Yee W. · Reviewed 3 месяцев ago

Kazuto K. · Reviewed 3 месяцев ago

BHANU TEJA KUMAR T. · Reviewed 3 месяцев ago

Kazuto K. · Reviewed 3 месяцев ago

MARCELO D. · Reviewed 3 месяцев ago

Pedro L. · Reviewed 3 месяцев ago

fail Vertex AI!!!! Task 4. Build metadata of documents containing text and images ResourceExhausted: 429 Resource exhausted. Please try again later. Please refer to https://cloud.google.com/vertex-ai/generative-ai/docs/error-code-429 for more details.

Teodoro C. · Reviewed 3 месяцев ago

it was impossible to run the lab within an hour, keep getting 429 error!!!!!

Sogand A. · Reviewed 3 месяцев ago

We are constantly receiving 429 errors. At this point, it is impossible to tell whether any progress is being made at all.

Fumiko K. · Reviewed 3 месяцев ago

The lab could not be completed due to repeated 429 quota/rate limit errors during the core metadata extraction step using Gemini multimodal processing. The issue occurred specifically in the `get_document_metadata()` pipeline while processing PDF pages and generating image descriptions/embeddings. Even after: * reducing the number of pages, * enabling `add_sleep_after_page=True`, * increasing sleep intervals significantly, * simplifying prompts, * and retrying multiple times, the process still failed repeatedly with 429 errors. Because `text_metadata_df` and `image_metadata_df` could not be generated successfully, the following RAG retrieval and similarity search steps could not be completed. This appears to be an infrastructure/shared quota issue rather than a user-side implementation problem. Since this metadata extraction step is the core part of the lab, the lab became effectively impossible to complete within the allocated session time and credits.

Jiwoon J. · Reviewed 3 месяцев ago

Samruddhi S. D. · Reviewed 3 месяцев ago

Very frequent resource is exhausted.

Nitin P. · Reviewed 3 месяцев ago

Kazuto K. · Reviewed 3 месяцев ago

Kazuto K. · Reviewed 3 месяцев ago