Add custom modeling code

Add custom modeling code

Upload final HF-compatible base model with architecture code

Fix KV cache handling in Attention layer

Add custom modeling code

Upload final HF-compatible 362M base model

Upload final HF-compatible 362M base model (pytorch_model.bin format)

Create README.md

Upload config.json with huggingface_hub

Upload pytorch_model.bin with huggingface_hub

initial commit