|
import os |
|
import sys |
|
os.system("pip install transformers==4.27.0") |
|
os.system("pip install torch") |
|
os.system("pip install openai") |
|
os.system("pip install accelerate") |
|
from transformers import pipeline, WhisperModel, WhisperTokenizer, WhisperFeatureExtractor, AutoFeatureExtractor, AutoProcessor, WhisperConfig |
|
os.system("pip install evaluate") |
|
|
|
|
|
os.system("pip install datasets") |
|
|
|
os.system("pip install spicy==1.8.1") |
|
os.system("pip install soundfile") |
|
os.system("pip install jiwer") |
|
os.system("pip install datasets[audio]") |
|
os.system("pip install numba==0.51.2") |
|
import torch |
|
from evaluate import evaluator |
|
from datasets import load_dataset, Audio, disable_caching, set_caching_enabled |
|
|
|
set_caching_enabled(False) |
|
disable_caching() |
|
|
|
huggingface_token = os.environ["huggingface_token"] |
|
|
|
model = WhisperModel.from_pretrained("mskov/whisper_miso", use_auth_token=huggingface_token) |
|
feature_extractor = AutoFeatureExtractor.from_pretrained("mskov/whisper_miso", use_auth_token=huggingface_token) |
|
|
|
model_config = WhisperConfig.from_pretrained("mskov/whisper_miso", use_auth_token=huggingface_token) |
|
model = WhisperModel(config=model_config) |
|
model.load_state_dict(torch.load("mskov/whisper_miso/pytorch_model.bin", use_auth_token=huggingface_token)) |
|
model.eval() |
|
|
|
dataset = load_dataset("mskov/miso_test", split="test").cast_column("audio", Audio()) |
|
print(dataset) |
|
|
