π Documentation Enhancement Suggestion
π Documentation Enhancement Suggestion
This observation was generated by Crovia β the AI transparency observation layer.
Crovia does not accuse or judge. It observes publicly available information and suggests improvements.
π Quick Stats
| Metric | Value |
|---|---|
| Source | huggingface |
| Downloads | 1176214 |
| Likes | 1459 |
| Last Updated | 2026-02-09 |
π» Ready-to-Use Code
from transformers import AutoModel, AutoTokenizer
model_id = "google/embeddinggemma-300m"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModel.from_pretrained(model_id)
# Example usage
inputs = tokenizer("Hello, world!", return_tensors="pt")
outputs = model(**inputs)
π Citation
If you use this model, please cite:
@misc {google_embeddinggemma_300m_2026,
author = {google},
title = {google/embeddinggemma-300m},
year = {2026},
url = {https://huggingface.co/google/embeddinggemma-300m},
note = {Accessed via CROVIA transparency registry}
}
βοΈ EU AI Act Compliance Checklist
- Training data disclosed
- License clearly stated
- Intended use documented
- Model limitations documented
- Evaluation metrics provided
- Bias/fairness analysis
π Training Data Transparency
Training Data Status: Documentation not found
No training data section was observed in the public model card.
This is an observation, not an accusation. Many valid reasons exist for this status.
If you'd like to improve documentation, consider adding:
- Dataset names and versions used
- Data collection methodology
- Preprocessing steps applied
- Known limitations
This may help users understand your model better and prepare for upcoming transparency requirements (e.g., EU AI Act).
Enhancement generated by CROVIA Β· Package ID: 928c04ec5665
Generated at: 2026-02-09T19:31:55.837624Z
This suggestion was generated by Crovia β the AI transparency observation layer.
Crovia does not accuse or judge. It observes publicly available information and suggests documentation improvements.
If this suggestion is helpful, consider adding the recommended sections to your model card.
If not applicable, feel free to close this discussion.
Learn more: croviatrust.com Β· What is Crovia?
Hey @CroviaTrust ,
Details about training dataset has been provided already, please refer here: https://huggingface.co/google/embeddinggemma-300m#training-dataset
Thanks!
Hi @srikanta-221 , thank you for pointing this out β you're right that embeddinggemma includes a Training Dataset section.
For context, our automated framework evaluates 19 disclosure elements (the NEC# canon) mapped to 11 regulatory jurisdictions. The existing training dataset description covers the high-level composition well.
Areas where additional structured detail could strengthen compliance coverage include:
Structured dataset references (named sources, proportions, snapshot dates)
Data provenance chain and licensing summaries per source
Environmental impact (training compute/energy estimates)
These are increasingly relevant under the EU AI Act, GDPR Art. 13-14, and similar frameworks.
We'll update our observation to reflect the existing coverage. If a detailed compliance mapping would be useful, we're happy to share one.
Thank you for the engagement β this is exactly the kind of constructive dialogue we hope to enable.