| license: mit | |
| tags: | |
| - moe | |
| - eve-swarm | |
| - specialist | |
| base_model: | |
| - anthonym21/Eve-2-MoE-IT-272M | |
| datasets: | |
| - bitext/Bitext-customer-support-llm-chatbot-training-dataset | |
| <div align="center"> | |
| <img src="eve-2-swarm.jpg" alt="Eve-2 Swarm" width="600"> | |
| </div> | |
| # Eve-2-MoE-NanoRouter-272M | |
| Intent classification and routing. | |
| ## Training | |
| - Hardware: NVIDIA H200 SXM | |
| - Method: Full Fine-Tuning (FFT) | |
| - Samples: 25,000 | |
| - Time: 4.4 min | |