Hugging Face's logo

Google
Enterprise
company
Verified
Activity Feed

AI & ML interests

Google ❤️ Open Source AI

Recent Activity
sonali-kumari11  new activity about 5 hours ago
google/gemma-4-12B-it:👋 Training support for transformers/megatron backends
osanseviero  new activity about 5 hours ago
google/gemma-4-12B-it:Wrong context size in config.json?
osanseviero  new activity about 5 hours ago
google/gemma-4-12B-it:fix max_position_embeddings 128Ki --> 256Ki
View all activity

Papers

Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories

PARCEL: Pool-Anchored Resampling with Conditioned Elastic Queries for Efficient Vision-Language Understanding

View all Papers

google 's collections 50

MedGemma Concept Apps
Collection of concept apps built built with MedGemma models to inspire the community.
VideoPrism
VideoPrism is a foundational video encoder that enables state-of-the-art performance on a large variety of video understanding tasks.
HAI-DEF Concept Apps
Collection of concept apps built around HAI-DEF open models/libraries to inspire the community. Learn more at http://goo.gle/hai-def`
Gemma 3 QAT
Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory
Health AI Developer Foundations (HAI-DEF)
Groups models released for use in health AI by Google. Read more about HAI-DEF at http://goo.gle/hai-def
BERT release
Regroups the original BERT models released by the Google team. Except for the models marked otherwise, the checkpoints support English.
T5 release
The original T5 transformer release was done in two steps, the original T5 checkpoints and the improved T5v1
SigLIP
Contrastive (sigmoid) image-text models from https://arxiv.org/abs/2303.15343
MedGemma Release
Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.
SigLIP2
TimesFM Release
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.
MedGemma Concept Apps
Collection of concept apps built built with MedGemma models to inspire the community.
VideoPrism
VideoPrism is a foundational video encoder that enables state-of-the-art performance on a large variety of video understanding tasks.
HAI-DEF Concept Apps
Collection of concept apps built around HAI-DEF open models/libraries to inspire the community. Learn more at http://goo.gle/hai-def`
MedGemma Release
Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.
Gemma 3 QAT
Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory
SigLIP2
Health AI Developer Foundations (HAI-DEF)
Groups models released for use in health AI by Google. Read more about HAI-DEF at http://goo.gle/hai-def
BERT release
Regroups the original BERT models released by the Google team. Except for the models marked otherwise, the checkpoints support English.
T5 release
The original T5 transformer release was done in two steps, the original T5 checkpoints and the improved T5v1
SigLIP
Contrastive (sigmoid) image-text models from https://arxiv.org/abs/2303.15343
TimesFM Release
TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.