Originally created by @Gyro0o on GitHub (Feb 24, 2024).
Can you add the option to use ollama embeddings with customized models (such as nomic-embed-text...)
Alongside the native implementation.
Originally created by @Gyro0o on GitHub (Feb 24, 2024).
Can you add the option to use ollama embeddings with customized models (such as nomic-embed-text...)
Alongside the native implementation.
@antonsapt4 commented on GitHub (Feb 25, 2024):
Hi Guys, is https://huggingface.co/nomic-ai/nomic-embed-text-v1.5 is not currently supported?
https://github.com/open-webui/open-webui/pull/772 been following till it merged, just want to know if this related and now we can use it to change the embed via dockerfile with nomic file
@tjbck @Gyro0o
Hello, I have tried to change the default embedding model for "mxbai-embed-large-v1". Docker build went fine, but when I try to ingest any document I get error:
"Embedding dimension 1024 does not match collection dimensionality 384"
I guess the new embedding model needs different size of vector db, but how do I configure that? Thanks.
edit: seems to be working after resetting vector storage.. sorry about that
@vaclcer commented on GitHub (Mar 26, 2024):
Hello, I have tried to change the default embedding model for "mxbai-embed-large-v1". Docker build went fine, but when I try to ingest any document I get error:
"Embedding dimension 1024 does not match collection dimensionality 384"
I guess the new embedding model needs different size of vector db, but how do I configure that? Thanks.
edit: seems to be working after resetting vector storage.. sorry about that
@vaclcer is it possible for you to create a guide or instructions on how to it. Would love to try out the new embedding model.
@JJassonn69 commented on GitHub (Apr 10, 2024):
@vaclcer is it possible for you to create a guide or instructions on how to it. Would love to try out the new embedding model.
@sammcj commented on GitHub (Apr 16, 2024):
Just tried it out, works as described - Thanks! 🌮
<img width="918" alt="image" src="https://github.com/open-webui/open-webui/assets/862951/2af43f17-87ad-4af2-8045-0a8c076c3ab1">
<img width="1537" alt="image" src="https://github.com/open-webui/open-webui/assets/862951/163e2e22-b74c-411b-83eb-92748dbb4251">
😅 And I just went through the trouble of switching to another built-in model, deleting my documents and vector DB the other day. This opens so many new possibilities!
@justinh-rahb commented on GitHub (Apr 16, 2024):
😅 And I just went through the trouble of switching to another built-in model, deleting my documents and vector DB the other day. This opens so many new possibilities!
Delete the knowledge database and re-create it solved my problem. Seems like because I changed the embedding model for this knowledge database, which is initialized with the embedding size of the previous embedding model (even it's empty now)
@makecent commented on GitHub (Mar 10, 2025):
Delete the knowledge database and re-create it solved my problem. Seems like because I changed the embedding model for this knowledge database, which is initialized with the embedding size of the previous embedding model (even it's empty now)
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @Gyro0o on GitHub (Feb 24, 2024).
Can you add the option to use ollama embeddings with customized models (such as nomic-embed-text...)
Alongside the native implementation.
@antonsapt4 commented on GitHub (Feb 25, 2024):
Hi Guys, is https://huggingface.co/nomic-ai/nomic-embed-text-v1.5 is not currently supported?
https://github.com/open-webui/open-webui/pull/772 been following till it merged, just want to know if this related and now we can use it to change the embed via dockerfile with nomic file
@tjbck @Gyro0o
@oliverbob commented on GitHub (Feb 25, 2024):
Nomic embed text is the best in the world.
Better than OpenAI's . Following....
@MikeyBeez commented on GitHub (Mar 7, 2024):
Has this been merged yet?
@vaclcer commented on GitHub (Mar 26, 2024):
Hello, I have tried to change the default embedding model for "mxbai-embed-large-v1". Docker build went fine, but when I try to ingest any document I get error:
"Embedding dimension 1024 does not match collection dimensionality 384"
I guess the new embedding model needs different size of vector db, but how do I configure that? Thanks.
edit: seems to be working after resetting vector storage.. sorry about that
@JJassonn69 commented on GitHub (Apr 10, 2024):
@vaclcer is it possible for you to create a guide or instructions on how to it. Would love to try out the new embedding model.
@vaclcer commented on GitHub (Apr 10, 2024):
I have a feeling this will be sorted out in upcoming v0.1.118. Lets see.
@tjbck commented on GitHub (Apr 14, 2024):
Added to our dev branch, let us know if you encounter any issues!
@sammcj commented on GitHub (Apr 16, 2024):
Love your work @tjbck :)
@sammcj commented on GitHub (Apr 16, 2024):
Just tried it out, works as described - Thanks! 🌮
@justinh-rahb commented on GitHub (Apr 16, 2024):
😅 And I just went through the trouble of switching to another built-in model, deleting my documents and vector DB the other day. This opens so many new possibilities!
@makecent commented on GitHub (Mar 10, 2025):
Delete the knowledge database and re-create it solved my problem. Seems like because I changed the embedding model for this knowledge database, which is initialized with the embedding size of the previous embedding model (even it's empty now)