Skip to main content

Synthese vocale neuronale monospeaker francais — modeles high (Conformer) et light (FastPitch) au choix (ONNX)

Project description

lectura-monospeaker

Moteur de synthese vocale neuronale monospeaker pour le francais — deux modeles au choix : high (Matcha-Conformer) et light (FastPitch) + HiFi-GAN (ONNX).

Installation

# Version minimale (API distante, zero deps)
pip install lectura-monospeaker

# Version locale (inference ONNX)
pip install lectura-monospeaker[onnx]

Pour le pipeline complet texte → audio, utilisez pip install lectura-tts-mono (inclut le G2P).

Utilisation

Depuis des phonemes IPA

from lectura_monospeaker import creer_engine

# Modele high (Conformer, meilleure qualite) — par defaut
engine = creer_engine(mode="local")
result = engine.synthesize_phonemes(
    "bɔ̃ʒuʁ",
    phrase_type=0,
    pitch_range=1.3,
)

# Modele light (FastPitch, plus rapide/leger)
engine_light = creer_engine(mode="local", model="light")
result = engine_light.synthesize_phonemes("bɔ̃ʒuʁ")

Raccourci synthetiser()

from lectura_monospeaker import synthetiser

# High (defaut)
audio = synthetiser("Bonjour le monde.")

# Light
audio = synthetiser("Bonjour le monde.", model="light")

Via l'API distante

from lectura_monospeaker import creer_engine

engine = creer_engine(mode="api", api_url="https://api.lectura.world")
result = engine.synthesize("Bonjour")

Modeles

Modele Architecture Taille (INT8) Qualite Vitesse
high (defaut) Matcha-Conformer + HiFi-GAN ~29 Mo Meilleure ~30x temps-reel
light FastPitch + HiFi-GAN ~28 Mo Bonne ~50x temps-reel

Controles prosodiques

Parametre Defaut Description
duration_scale 1.0 Vitesse globale
pitch_shift 0.0 Decalage F0 (demi-tons)
pitch_range 1.3 Variation F0 (1.0 = neutre)
energy_scale 1.0 Intensite
pause_scale 1.0 Duree des pauses
phrase_type 0 0=decl, 1=inter, 2=excl, 3=susp

Architecture

  • high : Matcha-Conformer — phonemes → mel-spectrogramme via flow matching ODE
  • light : FastPitch — phonemes → mel-spectrogramme via FFT decoder
  • HiFi-GAN V1 : mel → audio 22050 Hz (partage)
  • Runtime : ONNX (pas de dependance PyTorch)

Emplacements des modeles

Recherche dans l'ordre :

  1. Parametre models_dir explicite
  2. $LECTURA_MODELS_DIR/tts_mono/
  3. ~/.lectura/models/tts_mono/
  4. Modeles embarques dans le package (version privee)

Chaque emplacement peut contenir des sous-repertoires conformer/ et fastpitch/ (nouveau layout) ou les fichiers directement (ancien layout, retrocompatible).

Licence

Licence proprietaire Lectura. Voir LICENCE.txt.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

lectura_monospeaker-4.0.0.tar.gz (29.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

lectura_monospeaker-4.0.0-py3-none-any.whl (26.4 kB view details)

Uploaded Python 3

File details

Details for the file lectura_monospeaker-4.0.0.tar.gz.

File metadata

  • Download URL: lectura_monospeaker-4.0.0.tar.gz
  • Upload date:
  • Size: 29.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.11.9

File hashes

Hashes for lectura_monospeaker-4.0.0.tar.gz
Algorithm Hash digest
SHA256 d3b58f5d52517c3c9dc31521c1f1cbde6ea1d891687ff122258390de3dda1dcd
MD5 d6f29f790f6c23e587cd19bbe0d2a05d
BLAKE2b-256 bf41824d591e1e900ad86a17b94a6e54a632343293e2fa4fcd85c071a65132b1

See more details on using hashes here.

File details

Details for the file lectura_monospeaker-4.0.0-py3-none-any.whl.

File metadata

File hashes

Hashes for lectura_monospeaker-4.0.0-py3-none-any.whl
Algorithm Hash digest
SHA256 5aac80b32138e45659ee5317b10ea998b40dc76847ca642a011060fc37d296ef
MD5 9fd58b7cbefc9c2235a62cd5f6110f92
BLAKE2b-256 611384f8db392636a87c74a0fe898a0be0744beca6a7ccae188ae04010b36ab8

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page