https://aadonis-ai.github.io/Recent content in Adonis Asonitis on A.AsonitisHugo -- 0.157.0en-usSun, 05 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/neural-audio-codecs/xcodec2/Sun, 05 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/neural-audio-codecs/xcodec2/X-Codec2 setup and usage — single-codebook codec used by Llasa for TTS. Stable version is 1.3.0.https://aadonis-ai.github.io/notebook/neural-audio-codecs/wavtokenizer/Sat, 04 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/neural-audio-codecs/wavtokenizer/WavTokenizer setup — extreme compression (40–75 tok/sec, single codebook). Best for clean TTS; avoid for noisy/degraded speech.https://aadonis-ai.github.io/notebook/build-and-train/hf-trainer/Fri, 03 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/build-and-train/hf-trainer/A complete training script for LLaMA (250M–8B) with a custom tokenizer, label masking, gradient checkpointing, and multi-GPU support via Accelerate.https://aadonis-ai.github.io/notebook/speech-alignment/qwen3-aligner/Fri, 03 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-alignment/qwen3-aligner/Precise word and phoneme-level timestamps using Qwen3-ASR and its paired forced aligner — multilingual, robust to noise.https://aadonis-ai.github.io/notebook/neural-audio-codecs/snac/Fri, 03 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/neural-audio-codecs/snac/SNAC setup and encode/decode wrapper — matches DAC quality at significantly lower token rate via multi-scale residual vector quantization.https://aadonis-ai.github.io/notebook/build-and-train/custom-tokenizer/Thu, 02 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/build-and-train/custom-tokenizer/Building a HuggingFace-compatible tokenizer for neural audio codec tokens — vocabulary design, special tokens, and uploading to the Hub.https://aadonis-ai.github.io/notebook/neural-audio-codecs/dac/Thu, 02 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/neural-audio-codecs/dac/DAC setup and a wrapper class for encoding/decoding audio as flattened token sequences — time-major and codebook-major layouts.https://aadonis-ai.github.io/notebook/voice-conversion/ez-vc/Thu, 02 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/voice-conversion/ez-vc/Easy Zero-shot Any-to-Any Voice Conversion — single encoder architecture, excellent cross-lingual performance, and a clean inference script.https://aadonis-ai.github.io/notebook/speech-alignment/mfa-pyfoal/Thu, 02 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-alignment/mfa-pyfoal/Word and phoneme-level forced alignment with MFA and pyfoal — installation, inference script, and a fix for GitHub rate limit errors.https://aadonis-ai.github.io/notebook/speech-datasets/noise-rirs/Thu, 02 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-datasets/noise-rirs/Downloading ESC-50, the OpenSLR RIR database, DEMAND, and VGG Sound for noise augmentation and speech enhancement training.https://aadonis-ai.github.io/notebook/speech-datasets/clean-speech/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-datasets/clean-speech/Downloading HiFiTTS-2 (high-quality 44kHz) and SparkAudio VoxBox (60k+ hours, multi-language merged corpus).https://aadonis-ai.github.io/notebook/neural-audio-codecs/codec-comparison/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/neural-audio-codecs/codec-comparison/Side-by-side comparison of DAC, SNAC, WavTokenizer, and X-Codec2 — token rates, codebook counts, and when to use each.https://aadonis-ai.github.io/notebook/speech-alignment/espeak-ng/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-alignment/espeak-ng/Installing espeak-ng with and without sudo — including a no-admin local build for HPC clusters.https://aadonis-ai.github.io/notebook/build-and-train/hpc-clusters/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/build-and-train/hpc-clusters/Redirecting pip, HuggingFace, and PyTorch caches to scratch storage to avoid filling your home directory on shared clusters.https://aadonis-ai.github.io/notebook/speech-enhancement/llase-g1/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-enhancement/llase-g1/Setup and inference for LlaSE — a language model-based speech enhancement system that converts degraded audio to high-quality speech.https://aadonis-ai.github.io/notebook/voice-conversion/seed-vc/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/voice-conversion/seed-vc/Zero-shot voice conversion with Seed-VC — diffusion transformer architecture, real-time inference, batch processing, and a comparison with EZ-VC.https://aadonis-ai.github.io/notebook/speech-editing/voicecraft/Wed, 01 Oct 2025 00:00:00 +0000https://aadonis-ai.github.io/notebook/speech-editing/voicecraft/Token-based speech editing with VoiceCraft — setup, MFA forced alignment, and a self-contained editing script for insertions, deletions, and substitutions.https://aadonis-ai.github.io/cv/Mon, 01 Jan 0001 00:00:00 +0000https://aadonis-ai.github.io/cv/Curriculum Vitaehttps://aadonis-ai.github.io/publications/Mon, 01 Jan 0001 00:00:00 +0000https://aadonis-ai.github.io/publications/Research publications and datasets