Falco
(Falco)
1 oktober 2024 om 16:16
32
Yes.
Yes, each model produces different vector representations.
It’s basically one call per topic, so very easy to calculate.
If most of your topics are long, they will be truncated to 8k tokens, otherwise they will use your topic length.
Yes.
Overgrow:
I assume that for both related topics and AI-powered search, all posts need to be vectorized only once, so I can calculate the total number of words in posts table and derive the number of tokens needed. The same process would apply to the daily addition of posts. I’m neglecting the search phrases for now.
Both work at the topic level, so one per topic.
2 likes
May I know how to properly add the gems without forking the plugin with the suggested PR?
I’m trying the scale-to-zero feature on HuggingFace and I just need to use the rake task for backfill embeddings.
jlcoo
(Jiang Long)
7 juli 2025 om 08:33
40
why return 418 error code when I using discourse ai embeddings full search in DiscourseAi::Embeddings::EmbeddingsController#search as JSON? Could you help me?
Falco
(Falco)
Heeft dit topic gesplitst
14 oktober 2025 om 19:55
41
Falco
(Falco)
Heeft dit topic gesplitst
15 oktober 2025 om 15:36
42
Can Deepseek be used as provider? Which model?
How to estimate cost of creation as I have quite large DB (14.3 GB).
Falco
(Falco)
25 september 2026 om 13:23
44
Any model exposing the OpenAI compatible API can be used.
Number of public topics times the embedding model dimensions times 16 for the data in bytes, plus number of topics times dimensions divided by 8 for the index.
2 likes