Pinned Repositories
distil-whisper
Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.
ParallelWaveGAN
Unofficial Parallel WaveGAN (+ MelGAN & Multi-band MelGAN & HiFi-GAN & StyleMelGAN) with Pytorch
attention_sinks
stablelm 3b 4e1t
axolotl
for testing
llm_attention_sinks
Fork for attn sinks and beyond
relax_attention_sinks
for attention sink impl
StyleTTS2
StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models
onnxruntime
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
mlc-llm
Universal LLM Deployment Engine with ML Compilation
StyleTTS2
StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models
kmn1024's Repositories
kmn1024/attention_sinks
stablelm 3b 4e1t
kmn1024/axolotl
for testing
kmn1024/StyleTTS2
StyleTTS 2: Towards Human-Level Text-to-Speech through Style Diffusion and Adversarial Training with Large Speech Language Models
kmn1024/llm_attention_sinks
Fork for attn sinks and beyond
kmn1024/relax_attention_sinks
for attention sink impl