my-speech-datasets
This dataset is used to train Mozilla's DeepSpeech model, containing public‑domain speech data, currently focusing on Spanish speech, with plans to add more languages. It includes 120 hours of clean Spanish speech and 100 hours of clean speech from a single speaker, in LJSpeech format.
Dataset description and usage context
Dataset Overview
Data Collection Method
The dataset was created by automatically aligning transcripts with Windows speech‑recognition output and validating with Mozilla DeepSpeech models. Multiple language models were used during validation; the initial model was trained on voxforge Spanish data, followed by the model with the highest confidence scores on Windows speech‑recognition results.
Supported Languages
- Spanish
- 120 hours of clean speech data, address: 120h of clean speech
- 100 hours of clean speech from a single speaker, LJSpeech format, address: 100h of clean speech from a single speaker
License
The dataset is released under a public‑domain license.
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.