You can deploy this model and start earning money today!

arpit763/
canopylabs-orpheus-3b-0-1-ft-hindi-finetune

Name: canopylabs-orpheus-3b-0-1-ft-hindi-finetune
Availability: InStock

safetensors

Views

No views yet

language:

hi tags:
text-to-speech
tts
hindi
audio
speech-synthesis
fine-tuned license: apache-2.0 pipeline_tag: text-to-speech datasets:
SPRINGLab/IndicTTS-Hindi base_model: canopylabs/orpheus-3b-0.1-ft
metrics:
mos inference: false # set true if you have an inference API config

🎙️ Hindi TTS — Fine-tuned Model

A fine-tuned Text-to-Speech model for Hindi (हिन्दी), capable of generating natural, fluent speech from Hindi text including complex conjunct consonants and nasalization.

🔊 Audio Sample

Input text: "श्रीमती विश्वविद्यालय की प्रयोगशाला में क्षत्रिय छात्रों को त्रिकोणमिति की संज्ञाएँ समझा रही थीं।"

---

📌 Model Details

Property	Details
Model Type	Text-to-Speech (Acoustic Model)
Language	Hindi (`hi`)
Fine-tuned by	`arpit763`
Framework	`Hugging Face Transformers`
License	Apache 2.0

🧠 Model Description

This model was fine-tuned on Hindi speech data starting from <base-model-id>. It is designed to synthesize clear, natural-sounding Hindi speech, with improved handling of:

Conjunct consonants (e.g. क्ष, त्र, श्र, ज्ञ)
Nasalization and anusvara (ं, ँ)
Long compound words common in formal/literary Hindi
Correct pitch and prosody for declarative and interrogative sentences

📂 Repository Structure

language: hi tags:

text-to-speech
tts
hindi license: mit

My Fine-tuned Hindi TTS Model for educational purpose

Description

Fine-tuned on Hindi speech data for high-quality TTS.

Usage

# Add your inference code here

arpit763/canopylabs-orpheus-3b-0-1-ft-hindi-finetune