Update README.md

README.md

@@ -1,19 +1,9 @@
----
-title: README
-emoji: "🛰️"
-colorFrom: blue
-colorTo: green
-sdk: static
-pinned: false
----
-
-![Ratio1 banner](./assets/banner.svg)
-
 # Ratio1
 
 **Decentralized AI infrastructure for builders and partner teams.**
 
-Ratio1 helps teams build, deploy, and operate intelligent products without giving up control of their models, data paths, runtime, or economics.
+Ratio1 helps teams build, deploy, and operate intelligent products without giving
+up control of their models, data paths, runtime, or economics.
 
 [Website](https://ratio1.ai) • [GitHub](https://github.com/Ratio1) • [Documentation](https://github.com/Ratio1/docs)
 
@@ -21,7 +11,10 @@ Ratio1 helps teams build, deploy, and operate intelligent products without givin
 
 ## Positioning
 
-Ratio1 is built for teams that want AI infrastructure they can operate, extend, and trust. We focus on practical delivery: developer tooling, distributed execution, controllable data paths, and network economics that fit production workloads instead of locking them into a single hosted stack.
+Ratio1 is built for teams that want AI infrastructure they can operate, extend,
+and trust. We focus on practical delivery: developer tooling, distributed
+execution, controllable data paths, and network economics that fit production
+workloads instead of locking them into a single hosted stack.
 
 ## What We Do
 
@@ -64,3 +57,24 @@ This Hugging Face organization is the public front door for Ratio1 assets:
 - dataset cards for publishable data surfaces
 - demos and static Spaces for product and partner narratives
 - documentation links back to the `ratio1-docs` source
+
+### Public Models
+
+| Model | Task | Artifact | Status | Notes |
+| --- | --- | --- | --- | --- |
+| [`ratio1/edgeguard-cypher-qwen3-4b-v0.4-gguf`](https://huggingface.co/ratio1/edgeguard-cypher-qwen3-4b-v0.4-gguf) | EdgeGuard text-to-Cypher for cybersecurity graph investigation | GGUF Q4_K_M | Public release | Use behind an EdgeGuard read-only/schema validator. This model is intended to assist guarded graph-query workflows, not to execute arbitrary database operations. |
+| [`ratio1/redmesh-cybersecqwen-4b-lora-v0.1`](https://huggingface.co/ratio1/redmesh-cybersecqwen-4b-lora-v0.1) | Defensive CVE/CWE and vulnerability triage | PEFT LoRA adapter | Experimental public release | Adapter-only release for research and reproducibility. Not production-promoted. |
+
+### Public Datasets
+
+| Dataset | Task | Format | License | Notes |
+| --- | --- | --- | --- | --- |
+| [`ratio1/Drone_Detection_Survaillance`](https://huggingface.co/datasets/ratio1/Drone_Detection_Survaillance) | Drone object detection and surveillance imagery | Image and XML annotation files | CC-BY-NC-4.0 | Public computer-vision dataset with a declared train split. Dataset Viewer preview and parquet statistics may not be available for this file layout. |
+
+### Use Notes
+
+- Read each model or dataset card before use.
+- Treat experimental models as research artifacts unless the card explicitly says they are production-promoted.
+- Keep EdgeGuard query-generation models behind schema validation, read-only execution controls, and bounded result inspection.
+- For LoRA adapters, load the adapter with the base model named in the model card.
+- Check license terms before using datasets or model outputs in commercial or redistributed work.