gemma-3n-e2b-it config.json

text_config: hidden 2048, 30 layers, 8 q heads, 2 kv heads, head_dim 256, sliding_window 512, layer_types with 6 full_attention entries in 30 layers, final_logit_softcapping 30.0, max_position_embeddings 32768. Separate vision_config (gemma3n_vision) and audio_config. | SOURCE AUDIT 2026-10-03: kind was 'spec-sheet' but the URL is a raw Hugging Face config.json served as text/plain, not a vendor spec sheet; set to 'repo'. The schema field it fills (layer/window/head counts) is machine-readable model metadata and should be read as such.