---
license: apache-2.0
language:
- hi
- en
tags:
- whisper
- whisper.cpp
- ggml
- hinglish
- speech-recognition
---

# AutoCaptions models

GGML conversions of open speech models, used by the AutoCaptions Premiere Pro plugin
(offline Hinglish / Hindi / English captions with [whisper.cpp](https://github.com/ggml-org/whisper.cpp)).
The plugin downloads these files on first run; they are not modified beyond format conversion.

| File | Original model | Licence |
|---|---|---|
| `ggml-apex.bin` | [Oriserve/Whisper-Hindi2Hinglish-Apex](https://huggingface.co/Oriserve/Whisper-Hindi2Hinglish-Apex), converted to GGML fp16 | Apache-2.0 |
| `ggml-large-v3-turbo.bin` | [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) (whisper.cpp GGML) | MIT |
| `ggml-silero-v6.2.0.bin` | [Silero VAD](https://github.com/snakers4/silero-vad) v6.2 (whisper.cpp GGML) | MIT |

All credit for the models goes to their authors. Licence texts:
[Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0) · [MIT (Whisper)](https://github.com/openai/whisper/blob/main/LICENSE) ·
[MIT (Silero VAD)](https://github.com/snakers4/silero-vad/blob/master/LICENSE)

## GPU pack (`gpu-<hash>/`)

Downloaded by the plugin only on PCs with an NVIDIA graphics card, into its own private folder.

| File | Source | Licence |
|---|---|---|
| `ggml-cuda.dll` | [ggml](https://github.com/ggml-org/ggml) CUDA backend built for this plugin version | MIT |
| `cudart64_12.dll`, `cublas64_12.dll`, `cublasLt64_12.dll` | NVIDIA CUDA 12 runtime + cuBLAS, unmodified | [CUDA EULA](https://docs.nvidia.com/cuda/eula/), redistributable under Attachment A as part of the AutoCaptions plugin |

The NVIDIA files are distributed only as components of AutoCaptions and are not licensed for any other use.