#Using different embedding models in VectorDBs

1 messages · Page 1 of 1 (latest)

viscid plank
#

Can I search a 1536 dimension text-embedding-small vector
in a vector database created with 1536 dimension text-embedding-large vectors?

hallow wingBOT
#

Hi @viscid plank. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

The information provided in the extracts does not directly answer your question about the interchangeability of vectors generated by different embedding models in a vector database.

However, it's important to note that the text-embedding-3-small and text-embedding-3-large models generate embeddings of different characteristics, even though they may have the same dimensionality (1536 for the small model and 3072 for the large model, but can be reduced to 1536 using the dimensions parameter). These models might capture different levels of semantic and syntactic information from the text, which could affect the performance of similarity search in the vector database.

If you want to use vectors generated by a different model in a vector database, it would be advisable to ensure that the model is suitable for your specific use case and that the vectors are compatible with your database schema and search algorithm.

For more detailed information, you may want to refer to the OpenAI Embeddings Guide (https://platform.openai.com/docs/guides/embeddings) and the documentation of your specific vector database.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

#

@viscid plank kapa.ai is still learning and improving, please let us know how it did by reacting below

viscid plank
#

@hallow wing what's the gpt 3.5 turbo price right now

hallow wingBOT
#

Let me check the knowledge base for you. One moment! :books:

#

Based on the information provided in the extracts, the pricing for the GPT-3.5 Turbo models is as follows:

  • GPT-3.5-turbo: $1.50 for input usage and $2.00 for output usage
  • GPT-3.5-turbo-16k: $3.00 for input usage and $4.00 for output usage

Please note that these prices are in terms of megatokens. For the most accurate and up-to-date pricing, you should refer to the official OpenAI pricing page.

Source: OpenAI Community Post (https://community.openai.com/t/554656)

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon: