Single model: manifest and card now describe kokoro_batched.onnx only

Single model: manifest and card now describe kokoro_batched.onnx only

Remove kokoro.onnx: superseded by kokoro_batched.onnx, which is faster even at batch 1

Add kokoro_batched.onnx: variable-length batched graph (masked AdaIN + packed LSTMs), supersedes kokoro.onnx

Add kokoro_batched.onnx: variable-length batched graph (masked AdaIN + packed LSTMs), supersedes kokoro.onnx

Add kokoro_batched.onnx: variable-length batched graph (masked AdaIN + packed LSTMs), supersedes kokoro.onnx

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

sync from vernacula repo

initial commit