wundr-ai/echoread-gemma4-e4b-v3.2.1-litertlm-gpu

Promote Run C R1 (GPU composites) to model.litertlm — device A/B: 21 tok/s, 370ms prefill, 1.5GB; R2 (2-bit PTQ) rejected for repetition loops

Extend card with Run C R1/R2 bundles

Add Run C R2 bundle (composites + patch_1099 + 2-bit embed/PLE/lm_head channelwise mobile-parity recipe)

Add Run C R1 bundle (composites + patch_1099, dynamic_wi4b32_afp32)

Upload README.md with huggingface_hub

Upload model.litertlm with huggingface_hub

initial commit