Update model card with proper metadata and documentation

README.md

@@ -0,0 +1,75 @@
+---
+language:
+  - en
+license: apache-2.0
+library_name: gguf
+tags:
+  - gguf
+  - vibevoice
+  - realtime
+  - tts
+  - speech
+  - tiny
+  - affectively
+  - edgework
+  - aether
+  - distributed-inference
+  - edge-deployment
+base_model: fixie-ai/ultravox-v0_5-llama-3_2-1b
+base_model_relation: quantized
+pipeline_tag: text-to-speech
+---
+
+# VibeVoice Realtime 0.5B (GGUF, Q4_K_M)
+
+> **Production-ready** GGUF quantization of [fixie-ai/ultravox-v0_5-llama-3_2-1b](https://huggingface.co/fixie-ai/ultravox-v0_5-llama-3_2-1b) for distributed speech synthesis — powered by the [Aether](https://github.com/affectively-ai/aether) edge inference runtime.
+
+## Highlights
+
+- **0.5B parameters** — Ultra-low-latency voice model. Designed for real-time speech synthesis and voice interaction.
+- **~0.3 GB** Q4_K_M quantized — optimized for distributed edge inference
+- **LLaMA architecture** — proven, stable, well-tested
+- **Aether runtime compatible** — layer-sharded across distributed nodes via [Edgework.ai](https://edgework.ai)
+
+## Model Details
+
+| Property | Value |
+|----------|-------|
+| Base model | [fixie-ai/ultravox-v0_5-llama-3_2-1b](https://huggingface.co/fixie-ai/ultravox-v0_5-llama-3_2-1b) |
+| Parameters | 0.5B |
+| Architecture | LLaMA |
+| Quantization | Q4_K_M |
+| Format | GGUF |
+| Size | ~0.3 GB |
+| License | apache-2.0 |
+
+## Usage
+
+### With llama.cpp
+
+```bash
+./llama-cli -m vibevoice-realtime-0.5b-q4_k_m.gguf -p "Your prompt here" -n 256
+```
+
+### With Aether (Distributed Inference)
+
+This model is deployed across the [Aether](https://github.com/affectively-ai/aether) distributed inference network. Weights are layer-sharded and distributed across multiple edge nodes for parallel inference.
+
+## Deployment Architecture
+
+This model runs on the **Aether distributed inference runtime** — our custom engine that shards model layers across multiple nodes for parallel execution:
+
+1. **Coordinator** receives requests and manages token generation
+2. **Layer nodes** each hold a subset of model layers
+3. **Hidden states flow** between nodes via gRPC
+4. **Zero cold start** via warm pool scheduling
+
+Deployed via [Edgework.ai](https://edgework.ai) — bringing fast, cheap, and private inference as close to the user as possible.
+
+## About
+
+Published by [AFFECTIVELY](https://huggingface.co/affectively-ai) · Managed by [@buley](https://huggingface.co/buley)
+
+We quantize and publish **production-ready models** for distributed edge inference via the [Aether](https://github.com/affectively-ai/aether) runtime. Every release is tested for correctness and stability before publication.
+
+- [All models](https://huggingface.co/affectively-ai) · [GitHub](https://github.com/affectively-ai) · [Edgework.ai](https://edgework.ai)