French Educational Speech Transcriptions, 12.82 Hours of Audio

3,933 transcribed audio segments from the French educational domain, totaling approximately 12.82 hours of audio. The dataset was created by MEscriva using the OpenAI Whisper API and was last updated on December 17, 2025.

Use Cases

• Train automatic speech recognition models based on French educational audio.
• Benchmark speech-to-text accuracy for educational terminology based on high-quality Whisper transcriptions.
• Fine-tune language models for the educational domain using the transcribed text corpus.

Strengths

• Contains 3,933 distinct audio segments for model training.
• Provides approximately 12.82 hours of transcribed speech.
• Transcriptions were performed using the optimized Whisper-1 model to ensure accuracy.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset's specific source, collection method, and potential biases within the educational domain are not detailed.

Provenance

Source

huggingface

Collection Method

Audio transcribed using OpenAI Whisper API (optimized Whisper-1 model).

Freshness

Last updated 2025-12-17 16:56:52; freshness should be verified.