litert-community/Qwen3-TTS-12Hz-0.6B-Base

Card: definition line (LiteRT, 2026-09)

Card: base_model_relation: quantized (list under the base model's Quantizations)

Add measured Raspberry Pi 5 (CPU) section

Add measured Galaxy S26 NPU/GPU section

Card: add the measured Pixel 8a rows (GPU vs CPU) that the Android sweep produced

Add fast split codec (codec_partA/partB): ~2.5x codec on device, end-to-end RTF 6.7->2.06

Add fast folded int8 MTP (mtp_folded_int8.tflite): ~5x MTP on device, RTF 6.7->3.3

Model card: Android app section, Pixel 8a numbers, vocab/merges rows

Add merges.txt for the Android sample tokenizer

Add vocab.json/merges.txt for the Android sample tokenizer

Add Qwen3-TTS-12Hz-0.6B-Base LiteRT graphs, host tables, tokenizer, demo voice

initial commit