analyze spreadsheets。xlsx.csv

[xisungod]

Currently, I have locally deployed the qwq large model, OpenWebUI, and Pipelines. My goal is to use OpenWebUI together with qwq to analyze spreadsheets—for example, I upload a 200-row table and ask it to analyze that data. If I choose Smart Mode, OpenWebUI only feeds the model the first few rows; if I choose Full Context, it exceeds qwq’s maximum token limit. Is it possible to use a Pipeline to iteratively process the table in chunks, send each chunk to the model for analysis, and then finally summarize the results—just like commercial GPT services do, where you can upload anything and ask anything? I’m running entirely offline. Is there any way to achieve this?”