Spaces:
Build error
Build error
File size: 1,624 Bytes
061f670 ad29970 061f670 |
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 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
from datetime import datetime, timedelta
import os
import gradio as gr
import torch
import nemo.collections.asr as nemo_asr
import wandb
MODEL_HISTORY_DAYS = 180
WANDB_ENTITY = os.environ.get("WANDB_ENTITY", "tarteel")
WANDB_PROJECT_NAME = os.environ.get("WANDB_PROJECT_NAME", "nemo-experiments")
MODEL_NAME = os.environ.get("MODEL_NAME", "CfCtcLg-SpeUni1024-DI-EATLDN-CA:v0")
def get_device():
if torch.cuda.is_available():
return "cuda"
else:
return "cpu"
# run = wandb.init(entity=WANDB_ENTITY, project=WANDB_PROJECT_NAME)
wandb_api = wandb.Api(overrides={"entity": WANDB_ENTITY})
artifact = wandb_api.artifact(f"{WANDB_ENTITY}/{WANDB_PROJECT_NAME}/{MODEL_NAME}")
artifact_dir = artifact.download()
# find the model (ending with .nemo) in the artifact directory
model_path = [
os.path.join(root, file)
for root, dirs, files in os.walk(artifact_dir)
for file in files
if file.endswith(".nemo")
][0]
model = nemo_asr.models.EncDecCTCModelBPE.restore_from(
model_path, map_location=get_device()
)
def transcribe(audio_file):
transcription_file = model.transcribe([audio_file])[0]
return transcription_file
demo = gr.Blocks()
with demo:
gr.Markdown(
"""
# ﷽
"""
)
with gr.Row():
audio_file = gr.Audio(source="upload", type="filepath", label="File")
with gr.Row():
output_file = gr.TextArea(label="Audio Transcription")
b1 = gr.Button("Transcribe")
b1.click(
transcribe,
inputs=[audio_file],
outputs=[output_file],
api_name="transcribe",
)
demo.launch()
|