File size: 2,230 Bytes
0b5b7f4 973bb39 de07127 a8d3864 2996449 664eb76 0b5b7f4 664eb76 973bb39 590e8a9 74f5766 e7380d9 e4a4e02 8fdbf39 fd1d55f b1e43e6 dcf6504 e4a4e02 4da2268 196da5c e4a4e02 74f5766 973bb39 74f5766 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 |
import transformers
from transformers import pipeline
import gradio as gr
import os
import sys
os.system("pip install evaluate")
os.system("pip install datasets")
os.system("pip install spicy")
os.system("pip install soundfile")
os.system("pip install datasets[audio]")
from evaluate import evaluator
from datasets import load_dataset, Audio
p = pipeline("automatic-speech-recognition")
task_evaluator = evaluator("automatic-speech-recognition")
#url = {"test" : "https://huggingface.co/datasets/mskov/miso_test/blob/main/test_set.parquet"}
#data = load_dataset("audiofolder", data_dir="mskov/miso_test")
# data = load_dataset("audiofolder", data_files=["datasets/mskov/miso_test/test_set/and.wav","mskov/miso_test/test_set/chew1.wav","mskov/miso_test/test_set/chew3.wav", "mskov/miso_test/test_set/chew3.wav","mskov/miso_test/test_set/chew4.wav","mskov/miso_test/test_set/cough1.wav","mskov/miso_test/test_set/cough2.wav","mskov/miso_test/test_set/cough3.wav","mskov/miso_test/test_set/hi.wav","mskov/miso_test/test_set/knock_knock.wav","mskov/miso_test/test_set/mouth_sounds1.wav","mskov/miso_test/test_set/mouth_sounds2.wav","mskov/miso_test/test_set/no.wav","mskov/miso_test/test_set/not_bad.wav","mskov/miso_test/test_set/oh_i_wish.wav","mskov/miso_test/test_set/pop1.wav","mskov/miso_test/test_set/really.wav","mskov/miso_test/test_set/sigh1.wav","mskov/miso_test/test_set/sigh2.wav","mskov/miso_test/test_set/slurp1.wav","mskov/miso_test/test_set/slurp2.wav","mskov/miso_test/test_set/sneeze1.wav","mskov/miso_test/test_set/sneeze2.wav","mskov/miso_test/test_set/so_i_did_it_again.wav"])
dataset = load_dataset("mskov/miso_test", split="test").cast_column("audio", Audio())
results = task_evaluator.compute(
model_or_pipeline="https://huggingface.co/mskov/whisper_miso",
data=dataset,
input_column="audio",
label_column="category",
metric="wer",
)
print(results)
def transcribe(audio, state=""):
text = p(audio)["text"]
state += text + " "
return state, state
gr.Interface(
fn=transcribe,
inputs=[
gr.Audio(source="microphone", type="filepath", streaming=True),
"state"
],
outputs=[
"textbox",
"state"
],
live=True).launch()
|