Hugging Face's logo

AvitoTech
/
SigLIP2-giant-e5small-v2-gating

Feature Extraction
Transformers
Safetensors
avito_gated_fusion
siglip
siglip2
vision
text
clip
multimodal
image-text-embeddings
pet-recognition
custom_code
Model card Files
xet
 Community
2

Fix AutoModel loading, add image processor and Apache-2.0 license

#2
by Slenser0 - opened
base: refs/heads/main
←
from: refs/pr/2
Discussion Files changed
+61696
-116
Slenser0

Hi, and thanks for releasing these models.

This follows up on the report on the SigLIP2-Base repo.

  • The architecture was not loadable. config.json advertised this composite checkpoint as a plain SiglipModel, so AutoModel built a differently shaped model and initialised it randomly. The repo now ships configuration_avito_gated.py and modeling_avito_gated.py -- a faithful port of the FaceRecognizer class from the model card -- plus a config.json that describes both towers. Tensor names are unchanged, so existing load_state_dict code keeps working.
    Also: the download snippet in the card points at AvitoTech/SigLIP2-giant-e5small-v2-gating-for-animal-identification, which 404s. It now points here.
  • Image processor. Added preprocessor_config.json, copied unchanged from google/siglip2-giant-opt-patch16-384, so the repo is self-contained and the card no longer has to send users to another repository for preprocessing.
  • License. Added license: apache-2.0 and a LICENSE file (see below).

Verification

The reference is the wrapper this checkpoint was trained with: it is built from
google/siglip2-giant-opt-patch16-384 and loaded from the published model.safetensors with
load_state_dict(strict=True) (0 missing / 0 unexpected keys). Its embeddings are then
compared against AutoModel.from_pretrained on the files in this PR.

check result
missing / unexpected / mismatched keys 0 / 0 / 0
max abs difference vs. that reference 0.0
embedding dimensionality 512
image processor output vs. google/siglip2-giant-opt-patch16-384 identical (max abs diff 0.0)

No weight values change anywhere in this PR.

About the license

The repository currently has no license field and no LICENSE file, which is what prompted
the original question. This PR proposes Apache-2.0, matching the other public AvitoTech
models on the Hub and the Apache-2.0 base model(s) this is derived from -- but that call is
yours. If you would rather use different terms, say so and I will amend the PR; if you would
rather add the license yourself, feel free to drop the LICENSE file and the frontmatter
line from this PR and take just the loading fix.
Fix AutoModel loading, add image processor and Apache-2.0 licensec98d3b86
AvitoTech1 changed pull request status to merged

Sign up or log in to comment