Falcon-Perception Tech Report: 0.6B Model Beats SAM 3 in Segmentation
vanstriendaniel · x · 2026-08-12
The developer shared the paper and technical details for the Falcon-Perception model. The technical report reveals that despite having only 0.6B parameters, it outperforms SAM 3 in open-vocabulary segmentation, achieving a 68.0 Macro F1 score on the SA-Co dataset compared to SAM 3's 62.3. A complete recipe for using the model to build detection datasets is expected to be released later this week.
Related event: Falcon-Perception 0.6B Beats SAM 3 in Segmentation(2 posts)→
More from Models
- With $1.2T Valuation and Upcoming IPO, Users Frustrated by Claude Opus Being 'Lazy' — kimmonismus · 2026-08-12
- MOSS-VL Quantized Models Run Locally on 24GB VRAM for Multimodal Tasks — huggingface · 2026-08-12
- ACE-Step 1.5 Japanese Denpa Fine-Tuned Weights Released, Full Fine-Tune Not LoRA — MuziqueComfyUI · 2026-08-12
- Report: Ilya Sutskever's SSI Pivots to Test-Time Training for New Reasoning Engine — iruletheworldmo · 2026-08-12
- Liquid Crow Released: Squeezing Physical Cognition into a 450M-Parameter Micro-World Model — helloiamleonie · 2026-08-12
- Meta vs NVIDIA 30B Agent Models: Local Execution vs Cloud Routing — eyishazyer · 2026-08-12