GPU utilities and helpers for Strolid services.
cd local-repos/strolidlib
poetry installfrom strolidlib import is_cuda_available, move_to_gpu_maybe
from strolidlib.gpu import load_transcription_model, transcribe
from strolidlib.utils import seconds_to_ydhms
# Check CUDA availability
if is_cuda_available():
print("CUDA is available!")
# Load a model and transcribe
model = load_transcription_model("en_real_quick_0")
result = transcribe(model, audio_tensor)# Install with dev dependencies
poetry install
# Run tests
poetry run pytest