debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check#18692
Merged
danbev merged 2 commits intoJan 11, 2026
Conversation
ggerganov
approved these changes
Jan 8, 2026
Member
|
See my comments in #18464. If my understanding is correct there, then we likely don't need to query the pooling here at all and this can be simplified. |
Member
Author
Thanks, I'll take a look at this now 👍 |
This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE.
This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
danbev
force-pushed
the
llama-debug-pooling-enabled-check
branch
from
aab1d98 to
f31c38d
Compare
ggerganov
approved these changes
Jan 11, 2026
gary149
pushed a commit
to gary149/llama-agent
that referenced
this pull request
Jan 13, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
dillon-blake
referenced
this pull request
in dillon-blake/llama.cpp-BL
Jan 15, 2026
* debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
Seunghhon
pushed a commit
to Seunghhon/llama.cpp
that referenced
this pull request
Apr 26, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
ljubomirj
pushed a commit
to ljubomirj/llama.cpp
that referenced
this pull request
May 6, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
my-other-github-account
pushed a commit
to my-other-github-account/llama.cpp
that referenced
this pull request
May 15, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
my-other-github-account
pushed a commit
to my-other-github-account/llama.cpp
that referenced
this pull request
May 15, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
phibya
pushed a commit
to ziee-ai/llama.cpp
that referenced
this pull request
May 29, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
fewtarius
pushed a commit
to fewtarius/CachyLLama
that referenced
this pull request
May 30, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
fukuro-kun
pushed a commit
to fukuro-kun/fukuro-llama-cpp-turboquant
that referenced
this pull request
Jul 5, 2026
…-org#18692) * debug : include LLAMA_POOLING_TYPE_UNSPECIFIED in pooling check This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE. * debug : normalize both pooled and token embeddings This commit updates debug.cpp to normalize embeddings for both pooled and non-pooled outputs. For pooled embeddings, normalization is applied to the single vector, and for non-pooled embeddings, normalization is applied to each token embedding vector individually. The motivation for this is to enable non-pooled embeddings to be normalized which was not possible previously.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This commit updates the pooling check in the debug example to also include LLAMA_POOLING_TYPE_UNSPECIFIED and not just LLAMA_POOLING_TYPE_NONE.