Sync model card to GitHub README v7.3.0
Browse files
README.md
CHANGED
|
@@ -84,7 +84,15 @@ LLM_BACKEND=lmstudio LLM_API_URL=http://localhost:1234/v1 LLM_MODEL=smollm2 \
|
|
| 84 |
4. **Agentic Orchestrator**: A high-autonomy layer enabling NeuralAI to plan, reason, and execute multi-step workflows across the OS and web, moving beyond simple chat to active goal achievement.
|
| 85 |
|
| 86 |
---
|
| 87 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
## ✨ Key Features & Capabilities
|
| 89 |
|
| 90 |
### 💬 Multimodal Chat & Agentic Intelligence
|
|
@@ -205,11 +213,11 @@ A dedicated software engineer, thinker, and builder from West Memphis, AR. De'An
|
|
| 205 |
|
| 206 |
*Built with precision and discipline by De'Andrew Preston Harris.*
|
| 207 |
|
| 208 |
-
### CURRENT VERSION: v7.
|
| 209 |
|
| 210 |
- **Model Alignment**: DPO v15.0 Aligned (597 pairs, Logic, Debugging, Reasoning)
|
| 211 |
- **Inference**: llmster (LM Studio headless) — 258MB RAM vs 5GB PyTorch
|
| 212 |
-
- **Last Maintenance**: July
|
| 213 |
|
| 214 |
Your tone is technical, concise, and professional. You prioritize system stability and cleanliness above all else.
|
| 215 |
|
|
|
|
| 84 |
4. **Agentic Orchestrator**: A high-autonomy layer enabling NeuralAI to plan, reason, and execute multi-step workflows across the OS and web, moving beyond simple chat to active goal achievement.
|
| 85 |
|
| 86 |
---
|
| 87 |
+
### 🆕 What's New (v7.3.0)
|
| 88 |
+
|
| 89 |
+
- **Auto Release Notes ("What's New")**: A new top-bar panel surfaces the latest features and fixes automatically. Open it anytime via the ✨ **What's New** button.
|
| 90 |
+
- **Generated images render in chat**: Image-generation responses are parsed as Markdown and displayed inline.
|
| 91 |
+
- **No more self-talk**: Chat now uses the ChatML prompt template (`apply_chat_template`), matching the model's training format.
|
| 92 |
+
- **NeuralDrive upload reliability**: The file list correctly handles API response shapes after an upload.
|
| 93 |
+
- **Dark theme by default**: UI restores your saved theme and defaults to dark mode (fixes white file-cards in light mode).
|
| 94 |
+
- **Phase 8 in progress**: Knowledge Graph & Agentic Autonomy — long-term cross-project memory ("Supermemory") and fully autonomous task execution.
|
| 95 |
+
- **The NeuralLabs Shift**: NeuralAI is evolving into a standalone, downloadable intelligence environment (NeuralLabs v1 Client → v2 Edge → v3 Eco).
|
| 96 |
## ✨ Key Features & Capabilities
|
| 97 |
|
| 98 |
### 💬 Multimodal Chat & Agentic Intelligence
|
|
|
|
| 213 |
|
| 214 |
*Built with precision and discipline by De'Andrew Preston Harris.*
|
| 215 |
|
| 216 |
+
### CURRENT VERSION: v7.3.0 (The Pluggable Engine)
|
| 217 |
|
| 218 |
- **Model Alignment**: DPO v15.0 Aligned (597 pairs, Logic, Debugging, Reasoning)
|
| 219 |
- **Inference**: llmster (LM Studio headless) — 258MB RAM vs 5GB PyTorch
|
| 220 |
+
- **Last Maintenance**: July 13, 2026
|
| 221 |
|
| 222 |
Your tone is technical, concise, and professional. You prioritize system stability and cleanliness above all else.
|
| 223 |
|