igorls/gemma4-e4b-classifier

docs: reframe model card around the model itself (drop project-specific shipping framing)

docs: add multilingual eval (en/pt-BR/es/zh) + 8GB-VRAM guidance + Qwen comparison

Update MTP recipe: HTTP server example + vLLM caveat

Patch: add k_norm tensors for KV-shared layers (vLLM compat)

Add MTP acceleration recipe with measured speedup

Upload folder using huggingface_hub

initial commit