Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,47 @@
+---
+language: en
+license: mit
+library_name: transformers
+tags:
+- chess
+- mamba
+- moe
+---
+
+# ♟️ KiyEngine V3 (Mamba-MoE)
+
+Chess engine powered by Mamba-MoE architecture. Final training loss: **5.46**
+
+## Model Architecture
+
+- **Type**: Mamba with Mixture of Experts (MoE)
+- **Parameters**:
+  - d_model: 384
+  - n_layers: 4
+  - n_experts: 8
+  - top_k: 2
+  - d_state: 16
+  - d_conv: 4
+  - expansion_factor: 2
+  - vocab_size: 768
+
+## Usage
+
+```python
+from transformers import AutoConfig, AutoModel
+
+# Load the model
+config = AutoConfig.from_pretrained("Kiy-K/KiyEngine-V3-Mamba-MoE", trust_remote_code=True)
+model = AutoModel.from_pretrained("Kiy-K/KiyEngine-V3-Mamba-MoE", trust_remote_code=True)
+```
+
+## Training Details
+
+Model trained on chess data with final loss of 5.46.
+
+## Files
+
+- `model.safetensors`: Model weights
+- `config.json`: Model configuration
+- `configuration_kiyengine.py`: Configuration class definition
+- `modeling_kiyengine.py`: Model implementation (if available)