---

base_model: unslothai/whisper-large-v3

license: apache-2.0

library_name: whisper.cpp

pipeline_tag: automatic-speech-recognition

tags:

• whisper
• whisper.cpp
• ggml
• f16
• unsloth-studio
• unsloth

---

Whisper Large v3 F16 for Unsloth Studio

Run maximum-quality, fully local speech-to-text dictation in Unsloth Studio. Whisper Large v3 is the most accurate option in the default Whisper lineup when transcription quality matters more than download size or speed.

Run in Unsloth Studio

1. Install or update Unsloth Studio.
2. Open Settings > Voice.
3. Open the local dictation model picker and select Whisper Large v3.
4. Let Studio download and cache the model.
5. Use the microphone button in the chat composer to dictate locally.

The model runs on your device through whisper.cpp. Your recorded audio does not need to be sent to a hosted transcription service.

Model file

• whisper-large-v3.bin: native F16 model for whisper.cpp
• Download size: approximately 3.1 GB
• Best for: highest transcription quality on capable hardware

whisper.cpp uses a custom GGML binary format for Whisper. The model file is therefore named .bin, not .gguf, even though this repository follows the common -GGUF repository naming convention.

No low-bit quantization was applied. Matrix weights are stored as F16, while tensors that whisper.cpp requires in F32 remain F32.

Manual whisper.cpp usage

whisper-cli -m whisper-large-v3.bin -f audio.wav

Integrity

• Source model.safetensors SHA-256: a8e94b85976e5864ba3e9525c7e6c83b2a1eca42d4b797a0c7c24d778e40fd95
• Converted model SHA-256: e4fb6f478cfec2e2bbd4b271654338b08492550bbf221361a29d890bc22fbe01
• Conversion tool: ggml-org/whisper.cpp commit 080bbbe85230f624f0b52127f1ae1218247989f9

The converted model was loaded by whisper.cpp and passed an end-to-end transcription test.