Views
No views yet
dac_16khz_encoder_fp16.tflite (43 MB) — audio[1,1,16000] → latent[1,1024,50], GPU.dac_16khz_deconly_zs_fp16.tflite (105 MB) — latent[1,1024,50] → audio, GPU.dac_rvq.bin (1.2 MB) — RVQ weights (12 codebooks) for the CPU quantizer (float32 LE).audio -> encoder.tflite (GPU) -> z -> RVQ.encode (CPU) -> codes[12,50]
-> RVQ.decode (CPU) -> z_q -> decoder.tflite (GPU) -> audioLITERT_CL, 1 partition,
no CPU fallback); warm RTF ~0.82; reconstruction corr 1.0 vs PyTorch DAC.ConvTranspose1d are rewritten to a GPU-clean zero-stuff form (the real DAC's
odd stride-5 transposed conv fails converter legalization, and TRANSPOSE_CONV is rejected by Mali).
The RVQ uses EMBEDDING_LOOKUP + int64 indices (Mali-rejected) so it runs on CPU. So the float conv
graph stays fully on the GPU.