CoolFace
Datasetpublic

dlyog/af_heart_arm_tts_dataset

af_heart_arm_tts_dataset The distillation corpus used to train dlyog/af_heart_arm_tts — a single-voice TTS model built for Arm inference on NVIDIA DGX Spark. 3.0000 hours · 4,104 clips · 24 kHz mono 16-bit PCM, synthesized by Kokoro-82M speaking af_heart. This is knowledge distillation: the teacher generated every clip, so the student can approach it but never exceed it. The pipeline that produced this — and that reproduces it for any other voice — is at… See the full description on the dataset page: https://huggingface.co/datasets/dlyog/af_heart_arm_tts_dataset.

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
0likes199downloads
5 commits on main
6cf1e782mo ago

Delete wavs/README.md

tarunchy
dac57032mo ago

Upload 4104 files

tarunchy
1fceb6c2mo ago

Create wavs/README.md

tarunchy
f5d30c52mo ago

Upload 2 files

tarunchy
83f8e6c2mo ago

initial commit

tarunchy