TIPSv2-B/14
ai model→ stable
TIPSv2
Google's TIPSv2-B/14 is a vision encoder model that performs three dense prediction tasks from a single backbone.
1Total Mentions
+0.60Sentiment (Very Positive)
0.0%Velocity (7d)
First seen: Apr 11, 2026Last active: Apr 11, 2026
Signal Radar
Five-axis snapshot of this entity's footprint
Loading radar…
Mentions × Lab Attention
Weekly mentions (solid) and average article relevance (dotted)
mentionsrelevance
Loading timeline…
Timeline
1- Product LaunchApr 11, 2026
Google releases vision encoder model on Hugging Face for multi-task dense prediction
View source- tasks:
- depth estimation,surface normal prediction,semantic segmentation
- architecture:
- ViT-Base/14 with DPT heads
Recent Articles
No articles found for this entity.
Predictions
No predictions linked to this entity.
AI Discoveries
No AI agent discoveries for this entity.
Sentiment History
Positive sentiment
Negative sentiment
Range: -1 to +1
| Week | Avg Sentiment | Mentions |
|---|---|---|
| 2026-W15 | 0.60 | 1 |