Shorts Studio caption models
Speech-to-text models for Indian languages, converted to whisper.cpp's ggml format (8-bit, q8_0) for offline captions in Shorts Studio. Nothing here was retrained: each file is a format conversion of the original model below.
| File | Language | Original model | Licence |
|---|---|---|---|
| ggml-hindi-small-q8_0.bin | Hindi | vasista22/whisper-hindi-small | Apache-2.0 |
| ggml-tamil-small-q8_0.bin | Tamil | vasista22/whisper-tamil-small | Apache-2.0 |
| ggml-telugu-small-q8_0.bin | Telugu | vasista22/whisper-telugu-small | Apache-2.0 |
| ggml-kannada-small-q8_0.bin | Kannada | vasista22/whisper-kannada-small | Apache-2.0 |
| ggml-gujarati-small-q8_0.bin | Gujarati | vasista22/whisper-gujarati-small | Apache-2.0 |
| ggml-malayalam-small-q8_0.bin | Malayalam | vrclc/Whisper-small-Malayalam | Apache-2.0 |
| ggml-bengali-medium-q8_0.bin | Bengali | bengaliAI/tugstugi_bengaliai-asr_whisper-medium | Apache-2.0 |
The vasista22 models are by the SPRING Lab, IIT Madras. All originals are fine-tuned from OpenAI Whisper (MIT). Licensed under the Apache License 2.0: https://www.apache.org/licenses/LICENSE-2.0
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support