infiniflow
[Question]: Difference between Local LLM and Online LLM
Describe your problem
Ragflow calls the large model deployed locally with Ollama, and it responds to questions very quickly. However, when calling the large model API, it seems to be lost somewhere, resulting in a very slow response. Moreover, it appears to "dumb down" the large model, and the answers are quite silly. ——— This issue has been resolved. When calling the external large model API, Ragflow hides the thinking process, which is why it seems slow. The results are actually consistent with what is expected, and there is no "dumbing down" of the model.
We intend to create an international community, so we encourage using English for communication.
Different large models have different capabilities, and local deployed large models and online large models also vary in terms of the accuracy of answering questions and latency.
——— This issue has been resolved. When calling the external large model API, Ragflow hides the thinking process, which is why it seems slow. The results are actually consistent with what is expected, and there is no "dumbing down" of the model.