| | --- |
| | title: Indro-ai |
| | emoji: 📈 |
| | colorFrom: purple |
| | colorTo: red |
| | sdk: static |
| | pinned: true |
| | license: apache-2.0 |
| | --- |
| | |
| | # Indro-ai Research |
| |
|
| | **Pushing the Boundaries of Reasoning in Small Language Models.** |
| |
|
| | ## 🎯 Our Mission |
| | At Indro-ai, we are dedicated to building Sovereign Small Language Models (SLMs) that don't just predict the next word but actually **reason**. Our focus is on high-quality curated data and efficient architecture. |
| |
|
| | ## 🚀 Active Project: Indro-Veda (500M) |
| | Indro-Veda is our flagship 500M parameter model trained on a balanced diet of: |
| | - **Mathematics:** For logical structure. |
| | - **Code:** For algorithmic reasoning. |
| | - **Educational Data:** For high-quality knowledge. |
| |
|
| | ## 📊 Dataset Stats |
| | - **Total Tokens:** 3 Billion+ (Curated) |
| | - **Framework:** PyTorch/XLA (Optimized for TPU/GPU) |
| |
|
| | --- |
| | "Gyanam Paramam Balam" (Knowledge is Supreme Power) |
