search

SYSTRAN / faster-whisper

Faster Whisper transcription with CTranslate2
MIT License
11.9k stars 998 forks source link

Could not load library libcublasLt.so.11. #1017

Open nguyenhoanganh2002 opened 3 weeks ago

nguyenhoanganh2002 commented 3 weeks ago

code

from faster_whisper import WhisperModel

model_size = "large-v3"

model = WhisperModel(model_size, device="cuda", device_index=0, compute_type="float16")

segments, info = model.transcribe("rock_vocals.wav", beam_size=5, language="en", condition_on_previous_text=False)

print("hehe")

for segment in segments:
    print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text))

log

hehe
Could not load library libcublasLt.so.11. Error: libcublasLt.so.11: cannot open shared object file: No such file or directory
Aborted (core dumped)
rjwilmsi commented 1 week ago

In terminal for CUDA I first need to do the below, you will need something similar (adjust for your user and your distro's installation path):

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/User/.local/lib/python3.11/site-packages/nvidia/cudnn/lib:/home/User/.local/lib/python3.11/site-packages/nvidia/cublas/lib