adkar-recite-wav2vec2-ar-ctc

• Base model: jonatasgrosman/wav2vec2-large-xlsr-53-arabic (Apache-2.0)
• Input: [1, T] float32 mono waveform @ 16 kHz, T a multiple of 320
• Output: [1, T/320, 52] character logits (vocab in vocab_ar.json)
• Usage: CTC forced alignment against a known target text (no open-vocab decoding) — see the app's reciteCTC grader.
• Runtime: react-native-executorch generic ExecutorchModule.