Update app.py

app.py

@@ -1,5 +1,6 @@
+#!/usr/bin/env python
 ### -----------------------------------------------------------------------
-### (BASE, Revised) version_1.07 ALPHA, app.py
+### (test_BASE, Revised) version_1.07 ALPHA, app.py
 ### -----------------------------------------------------------------------
 
 # -------------------------------------------------------------------------
@@ -16,96 +17,105 @@
 # limitations under the License.
 # -------------------------------------------------------------------------
 
-import spaces
-import gradio as gr
-from PIL import Image
-#from pydub import AudioSegment
-#from scipy.io import wavfile
-
 import os
 import re
+import uuid
 import time
-import warnings
-#import datetime
-#import pandas as pd
-#import csv
+import psutil
 import subprocess
-from pathlib import Path
+from tqdm import tqdm
+
 import tempfile
 from fpdf import FPDF
+from pathlib import Path
+
+import numpy as np
+import torch
+from transformers import pipeline 
 
-import psutil
 from gpuinfo import GPUInfo
 
-#import numpy as np
-import torch
-#import torchaudio
-#import torchaudio.transforms as transforms
+import gradio as gr
 
-from transformers import pipeline #AutoModel
 
-#import spacy
-#import networkx as nx
-#from sklearn.feature_extraction.text import TfidfVectorizer
-#from sklearn.metrics.pairwise import cosine_similarity
+###############################################################################
+# Configuration.
+###############################################################################
 
-warnings.filterwarnings("ignore")
+#if not torch.cuda.is_available():
+    #DESCRIPTION += "\n<p>⚠️Running on CPU, This may not work on CPU.</p>"
 
-# ------------header section------------
-HEADER_INFO = """
-    # SWITCHVOX ✨|🇳🇴 *Transkribering av lydfiler til norsk skrift*
-""".strip()
-LOGO = "https://cdn-lfs-us-1.huggingface.co/repos/fe/3b/fe3bd7c8beece8b087fddcc2278295e7f56c794c8dcf728189f4af8bddc585e1/24ad06a03a5bc66f3eba361b94e45ad17e46f98b76632f2d17faf8a0b4f9ab6b?response-content-disposition=inline%3B+filename*%3DUTF-8%27%27banner_trans.png%3B+filename%3D%22banner_trans.png%22%3B&response-content-type=image%2Fpng&Expires=1725145079&Policy=eyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTcyNTE0NTA3OX19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy11cy0xLmh1Z2dpbmdmYWNlLmNvL3JlcG9zL2ZlLzNiL2ZlM2JkN2M4YmVlY2U4YjA4N2ZkZGNjMjI3ODI5NWU3ZjU2Yzc5NGM4ZGNmNzI4MTg5ZjRhZjhiZGRjNTg1ZTEvMjRhZDA2YTAzYTViYzY2ZjNlYmEzNjFiOTRlNDVhZDE3ZTQ2Zjk4Yjc2NjMyZjJkMTdmYWY4YTBiNGY5YWI2Yj9yZXNwb25zZS1jb250ZW50LWRpc3Bvc2l0aW9uPSomcmVzcG9uc2UtY29udGVudC10eXBlPSoifV19&Signature=PCB1BZeLzsZXnn4lRi1Fj50%7E0E4G39u6-GKLNLLgxhDyhXlqb3BJkR7IOgdyjuNyBh8Iz2d7QqhzNSsOlQmqR30cJLl6aDM5eJO01OlWXoD3Z0KcphoVBFyyrkoxe2oS8i2mdlbFRYn7oc%7EhyOcW46zR6HtqAB91iEydhEa5WTyz3C9nWasgMZevb0vRJtzwhplM9e-%7EbRrZTm2fMzkL14IGWpTpUOGBe93BDSAYbPhrZK1jvuY8p0Tmy1iEKVP3Zdzix5U5lrbxit5luitEhK8x6q2t63Gdv7F0CZvjQtTh7MYkB5GNiru8bTGKAgCdHGiZbG7VCGfhlX3UKvUTPg__&Key-Pair-Id=K24J24Z295AEI9"
-SIDEBAR_INFO = f"""
-<div align="center">
-    <img src="{LOGO}" style="width: 100%; height: auto;"/>
-</div>
-"""
+CACHE_EXAMPLES = torch.device('cuda') and os.getenv("CACHE_EXAMPLES", "0") == "1"
+#CACHE_EXAMPLES = torch.cuda.is_available() and os.getenv("CACHE_EXAMPLES", "0") == "1"
+#USE_TORCH_COMPILE = os.getenv("USE_TORCH_COMPILE", "0") == "1"
+#ENABLE_CPU_OFFLOAD = os.getenv("ENABLE_CPU_OFFLOAD", "0") == "1"
+
+device = torch.device('cuda')
+#device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-@spaces.GPU(duration=120)
-def transcribe(microphone, file_upload):
+#@spaces.GPU
+def transcribe(microphone, file_upload, progress=gr.Progress(track_tqdm=True)):
     
     file = microphone if microphone is not None else file_upload
     start_time = time.time()
     
     #--------------____________________________________________--------------"
 
-    device = "cuda" if torch.cuda.is_available() else "cpu"
-    pipe = pipeline("automatic-speech-recognition", model="NbAiLab/nb-whisper-large", device=device)
-    text = pipe(file)["text"]
+    #if torch.cuda.is_available():
+        #with torch.no_grad():
+            #pipe = pipeline("automatic-speech-recognition", model="NbAiLab/nb-whisper-large", device=device)
+        
+    with torch.no_grad():
+        pipe = pipeline("automatic-speech-recognition", model="NbAiLab/nb-whisper-large", device=device)
     
-    #--------------____________________________________________--------------"
+    
+    text = pipe(file)["text"] 
 
+    #--------------____________________________________________--------------"
+    
     end_time = time.time()
     output_time = end_time - start_time
+    
+    # --Word count
     word_count = len(text.split())
 
-    # --GPU metrics
+    # --Memory metrics
     memory = psutil.virtual_memory()
     
-    # Default GPU utilization and memory to 0 in case of an error
-    gpu_utilization = 0
-    gpu_memory = 0
-    try:
-        gpu_utilization, gpu_memory = GPUInfo.gpu_usage()
-        gpu_utilization = gpu_utilization[0] if len(gpu_utilization) > 0 else 0
-        gpu_memory = gpu_memory[0] if len(gpu_memory) > 0 else 0
-    except Exception as e:
-        print(f"Error retrieving GPU info: {e}")
-
     # --CPU metric
     cpu_usage = psutil.cpu_percent(interval=1)
+
+    # --GPU metric
+    gpu_utilization, gpu_memory = GPUInfo.gpu_usage()
     
     # --system info string
     system_info = f"""
     Processing time: {output_time:.2f} seconds.
     Number of words: {word_count}
+    Memory used: {memory.percent}%
+    GPU Utilization: {gpu_utilization}%
+    GPU Memory: {gpu_memory}
+    CPU Usage: {cpu_usage}%
     """
-    # *Memory: {memory.total / (1024 * 1024 * 1024):.2f}GB, used: {memory.percent}%, available: {memory.available / (1024 * 1024 * 1024):.2f}GB.*
-    # *GPU Utilization: {gpu_utilization}%, GPU Memory: {gpu_memory}*
-    # *CPU Usage: {cpu_usage}%*
+    #--------------____________________________________________--------------"
+
     return text, system_info
 
+
+###############################################################################
+# Interface.
+###############################################################################
+
+HEADER_INFO = """
+    # SWITCHVOX ✨|🇳🇴 *Transkribering av lydfiler til Norsk skrift.*
+""".strip()
+LOGO = "https://cdn-lfs-us-1.huggingface.co/repos/fe/3b/fe3bd7c8beece8b087fddcc2278295e7f56c794c8dcf728189f4af8bddc585e1/24ad06a03a5bc66f3eba361b94e45ad17e46f98b76632f2d17faf8a0b4f9ab6b?response-content-disposition=inline%3B+filename*%3DUTF-8%27%27banner_trans.png%3B+filename%3D%22banner_trans.png%22%3B&response-content-type=image%2Fpng&Expires=1725145079&Policy=eyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTcyNTE0NTA3OX19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy11cy0xLmh1Z2dpbmdmYWNlLmNvL3JlcG9zL2ZlLzNiL2ZlM2JkN2M4YmVlY2U4YjA4N2ZkZGNjMjI3ODI5NWU3ZjU2Yzc5NGM4ZGNmNzI4MTg5ZjRhZjhiZGRjNTg1ZTEvMjRhZDA2YTAzYTViYzY2ZjNlYmEzNjFiOTRlNDVhZDE3ZTQ2Zjk4Yjc2NjMyZjJkMTdmYWY4YTBiNGY5YWI2Yj9yZXNwb25zZS1jb250ZW50LWRpc3Bvc2l0aW9uPSomcmVzcG9uc2UtY29udGVudC10eXBlPSoifV19&Signature=PCB1BZeLzsZXnn4lRi1Fj50%7E0E4G39u6-GKLNLLgxhDyhXlqb3BJkR7IOgdyjuNyBh8Iz2d7QqhzNSsOlQmqR30cJLl6aDM5eJO01OlWXoD3Z0KcphoVBFyyrkoxe2oS8i2mdlbFRYn7oc%7EhyOcW46zR6HtqAB91iEydhEa5WTyz3C9nWasgMZevb0vRJtzwhplM9e-%7EbRrZTm2fMzkL14IGWpTpUOGBe93BDSAYbPhrZK1jvuY8p0Tmy1iEKVP3Zdzix5U5lrbxit5luitEhK8x6q2t63Gdv7F0CZvjQtTh7MYkB5GNiru8bTGKAgCdHGiZbG7VCGfhlX3UKvUTPg__&Key-Pair-Id=K24J24Z295AEI9"
+SIDEBAR_INFO = f"""
+<div align="center">
+    <img src="{LOGO}" style="width: 100%; height: auto;"/>
+</div>
+"""
+
 def save_to_pdf(text, summary):
     pdf = FPDF()
     pdf.add_page()
@@ -127,13 +137,14 @@ css = """
 #transcription_output textarea {
     background-color: #000000;  /* black */
     color: #00FF00 !important;  /* text color */
-    font-size: 16px;  /* font size */
+    font-size: 18px;  /* font size */
 }
 
 #system_info_box textarea {
     background-color: #ffe0b3;  /* orange */
     color: black !important;  /* text color */
-    font-size: 14px;  /* font size */
+    font-size: 16px;  /* font size */
+    font-weight: bold;  /* bold font */
 }
 """
 
@@ -159,12 +170,14 @@ with iface:
     with gr.Row():   
         with gr.Column(scale=3):
             text_output = gr.Textbox(label="Transkribert Tekst", elem_id="transcription_output")
-        system_info = gr.Textbox(label="Antall sekunder, ord:", elem_id="system_info_box")
+        with gr.Column(scale=1):
+            system_info = gr.Textbox(label="Antall sekunder, ord:", elem_id="system_info_box")
+        
     
     with gr.Tabs():
         with gr.TabItem("Download PDF"):
             pdf_text_only = gr.Button("Last ned pdf med resultat")
-            pdf_output = gr.File(label="/.docx?")
+            pdf_output = gr.File(label="/.pdf")
 
             pdf_text_only.click(fn=lambda text: save_to_pdf(text, ""), inputs=[text_output], outputs=[pdf_output])
 
@@ -180,9 +193,14 @@ with iface:
         </a>
         </div>
         ''')
-        
-    transcribe_btn.click(fn=transcribe, inputs=[microphone, upload], outputs=[text_output, system_info])
+    transcribe_btn.click(
+        fn=transcribe, 
+        inputs=[microphone, upload], 
+        outputs=[text_output, system_info]
+    )    
+    
+    #transcribe_btn.click(fn=transcribe, inputs=[microphone, upload], outputs=[text_output, system_info])
 
 
 
-iface.launch(share=True, debug=True)
+iface.launch(share=True,debug=True, enable_queue=True)