Voice typing models (v0.1.0)

This release includes models for Joplin's pre-release voice typing feature (for Android). Each model includes a README with additional information about the model source and/or build.

The whisper-small models are the largest (and slowest) of the attached models, but likely to have the highest quality. whisper-base is smaller than whisper-small and whisper-tiny is the smallest.

The whisper-tiny-fr-v2 model was fine-tuned more recently, more than doubling the number of training iterations and may provide better accuracy than whisper-tiny-fr.

The whisper-base--efficient-encoder.zip model should be faster than whisper-base.zip, but requires that a pull request be merged.

Note: whisper-base-fr-efficient-encoder-q8_0 and whisper-base-fr-efficient-encoder are both based on whisper-base-fr-v2.