Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Nomic Embed integration via llama.cpp BERT implementation #2086

Merged
merged 41 commits into from
Mar 13, 2024
Merged

Nomic Embed integration via llama.cpp BERT implementation #2086

merged 41 commits into from
Mar 13, 2024

Commits on Feb 28, 2024

  1. llmodel_c: simplify casts and callbacks 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Feb 28, 2024
    Configuration menu
    Copy the full SHA
    3077b97 View commit details
    Browse the repository at this point in the history

Commits on Feb 29, 2024

  1. modellist: fix a memory leak 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Feb 29, 2024
    Configuration menu
    Copy the full SHA
    3d8159b View commit details
    Browse the repository at this point in the history

  2. modellist: remove modelDirPath(), which was never used 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Feb 29, 2024
    Configuration menu
    Copy the full SHA
    536d666 View commit details
    Browse the repository at this point in the history

Commits on Mar 8, 2024

  1. llamamodel: initial LLamaModel::embed implementation 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    520b3af View commit details
    Browse the repository at this point in the history

  2. s/prompts/texts/ 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    7b37051 View commit details
    Browse the repository at this point in the history

  3. python: adapt to the new LLModel::embed API 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    4f11471 View commit details
    Browse the repository at this point in the history

  4. llamamodel: whitelist BERT and Nomic BERT 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    54e8fbe View commit details
    Browse the repository at this point in the history

  5. python: Nomic Embed is actually working now 

    Fix a few bugs in the implementation so that it actually works.

Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    e38140e View commit details
    Browse the repository at this point in the history

  6. llamamodel: blacklist old MiniLM quant 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    735127a View commit details
    Browse the repository at this point in the history

  7. llamamodel: only allow LLamaModel::embed on embedding models 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    9134410 View commit details
    Browse the repository at this point in the history

  8. chat: initial integration of new embedding code with chat UI 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    5671c0e View commit details
    Browse the repository at this point in the history

  9. llamamodel: use embedding code from "fix embeddings" PR 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    4751cd3 View commit details
    Browse the repository at this point in the history

  10. matryoshka 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    99f1e93 View commit details
    Browse the repository at this point in the history

  11. fix n_batch for embedding models 

    The intended value is the trained context length.

Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    b998791 View commit details
    Browse the repository at this point in the history

  12. update to latest changes from PR 5796 (merged) 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    979b947 View commit details
    Browse the repository at this point in the history

  13. better alignment with Atlas API 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    9b51271 View commit details
    Browse the repository at this point in the history

  14. closer adherence to Atlas API 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    8e5645c View commit details
    Browse the repository at this point in the history

  15. it compiles! 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    d00a9e0 View commit details
    Browse the repository at this point in the history

  16. python fixup 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    6c10840 View commit details
    Browse the repository at this point in the history

  17. chat: don't do anything with pre-GGUF .bin files 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    5c10a2d View commit details
    Browse the repository at this point in the history

  18. chat: restrict local embed to all-MiniLM-L6-v2-gguf2.f16.gguf for now 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    56768ff View commit details
    Browse the repository at this point in the history

  19. models3.json: add new MiniLM quant 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    910b472 View commit details
    Browse the repository at this point in the history

  20. models3.json: add Nomic Embed v1 and v1.5 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    17fe0b3 View commit details
    Browse the repository at this point in the history

  21. Merge branch 'main' into new-bert

    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    ab8bfd9 View commit details
    Browse the repository at this point in the history

  22. llmodel_c: fix typo 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    5f7103e View commit details
    Browse the repository at this point in the history

  23. Merge branch 'main' into new-bert

    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    bb1fc75 View commit details
    Browse the repository at this point in the history

  24. fix botched merge 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    6ee8c76 View commit details
    Browse the repository at this point in the history

  25. chat: fix filename bug caused by merge 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    eabc9cd View commit details
    Browse the repository at this point in the history

  26. chat: update isEmbeddingModel logic for models.json 

    We need to be able to set embeddingModel in models.json, otherwise we
cannot reliably know whether the model is an embedding model because we
cannot open it to check its architecture.

Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    122ae59 View commit details
    Browse the repository at this point in the history

  27. modellist: also whitelist nomic-embed-text-v1.txt 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 8, 2024
    Configuration menu
    Copy the full SHA
    c39fa52 View commit details
    Browse the repository at this point in the history

Commits on Mar 10, 2024

  1. python: use slice to index embedding_ptr 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 10, 2024
    Configuration menu
    Copy the full SHA
    33a0cd6 View commit details
    Browse the repository at this point in the history

Commits on Mar 12, 2024

  1. python: change dimensionality exception to a warning 

    This is a warning in the Nomic client, so it should be a warning here.

Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    db10b7d View commit details
    Browse the repository at this point in the history

  2. Merge branch 'main' into new-bert

    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    cb214df View commit details
    Browse the repository at this point in the history

  3. style: do not use braces with single-line ifs 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    46d64d6 View commit details
    Browse the repository at this point in the history

  4. embllm: make sendAtlasRequest private 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    39fb564 View commit details
    Browse the repository at this point in the history

  5. modellist: clarify arument to EmbeddingModels 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    e3fde3b View commit details
    Browse the repository at this point in the history

  6. llamamodel: handle null d_ptr->model in auto-prefix embed() 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    223090b View commit details
    Browse the repository at this point in the history

  7. llamamodel: fix missing const 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    5c1ddb0 View commit details
    Browse the repository at this point in the history

  8. llamamodel: make magic constant less magic 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 12, 2024
    Configuration menu
    Copy the full SHA
    9b93ab1 View commit details
    Browse the repository at this point in the history

Commits on Mar 13, 2024

  1. models3.json: assume these changes will be released in v2.7.4 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 13, 2024
    Configuration menu
    Copy the full SHA
    967616e View commit details
    Browse the repository at this point in the history

  2. modellist: never list or use models with disableGUI set 

    Signed-off-by: Jared Van Bortel <jared@nomic.ai>
    @cebtenzzre
    cebtenzzre committed Mar 13, 2024
    Configuration menu
    Copy the full SHA
    ad86bf5 View commit details
    Browse the repository at this point in the history