Compact multilingual Whisper, a step up from tiny
Whisper Base is OpenAI's 74M-parameter multilingual Whisper checkpoint. It uses the standard
Whisper encoder-decoder architecture for automatic speech recognition and speech translation,
trained with large-scale weak supervision on 680k hours of labelled speech. Base offers a
meaningful accuracy gain over tiny while staying small and fast enough for low-resource
devices. This OpenASR repo repackages the original openai/whisper-base weights as .oasr
packs that run natively in the OpenASR runtime with no Python at inference time. For most users
the q8_0 build is the recommended default; q4_k is for tighter memory budgets and fp16 is for
verification or maximum fidelity.
.oasr packs run with no Python at inference, engineered for CPU and Apple SiliconThese are CLI / local-server examples. The desktop app runs this model without typing a command — see the desktop install path above.
$ openasr pull whisper-base:q8 ↓ whisper-base.oasr 102.8 MB ✓ verified sha256 $ openasr transcribe meeting.wav --backend native --model-pack ~/.openasr/models/whisper-base/q8_0/whisper-base-q8_0.oasr ✓ local transcript · 0 bytes sent
$ openasr serve --backend native --model-pack ~/.openasr/models/whisper-base/q8_0/whisper-base-q8_0.oasr --addr 127.0.0.1:8080 ▶ http://127.0.0.1:8080 · model=whisper-base · 0 bytes will leave this host
from openai import OpenAI client = OpenAI(base_url="http://127.0.0.1:8080/v1", api_key="local") audio = open("meeting.wav", "rb") text = client.audio.transcriptions.create(model="whisper-base", file=audio)