Update app.py

app.py

@@ -1,4 +1,11 @@
 
+
+#             -----------------COPY OF NEW EDITION[app.py]-----------------
+
+# check if still the case...........??*********************************************
+# "The attention mask is not set and cannot be inferred from input because pad token is same as eos token. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results."
+
+
 import time
 import os
 import spaces
@@ -19,26 +26,28 @@ from transformers import pipeline # AutoProcessor, AutoModelForSpeechSeq2Seq
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 torch_dtype = torch.float32
 
-asr = pipeline("automatic-speech-recognition", model="NbAiLabBeta/nb-whisper-large", device=device, torch_dtype=torch_dtype)
+pipe = pipeline("automatic-speech-recognition", model="NbAiLabBeta/nb-whisper-large", device=device, torch_dtype=torch_dtype)
 
-@spaces.GPU(queue=True)
-def transcribe_audio(audio_file):
+# @spaces.GPU(queue=True)
+def transcribe_audio(audio_file, forced_decoder_ids):
     if audio_file.endswith(".m4a"):
         audio_file = convert_to_wav(audio_file)
 
     start_time = time.time()
-
+    forced_decoder_ids = processor.get_decoder_prompt_ids(language=language, task=task)
+    # check if still the case...........??*********************************************
+    # "You have passed task=transcribe, but also have set `forced_decoder_ids` to [[1, 50288], [2, 50360], [3, 50364]] which creates a conflict. `forced_decoder_ids` will be ignored in favor of task=transcribe."
     with torch.no_grad():
-        output = asr(audio_file, chunk_length_s=30, generate_kwargs={"num_beams": 8}) # "task": "transcribe", "language": "no"
+        output = pipe(audio_file, chunk_length_s=30, generate_kwargs={"forced_decoder_ids”: forced_decoder_ids}", "num_beams": 8, "language": "norwegian"}) # "task": "transcribe",
 
-    transcription = output["text"]
+    text = output["text"]
     end_time = time.time()
     output_time = end_time - start_time
-    word_count = len(transcription.split())
+    word_count = len(text.split())
 
     result = f"Time taken: {output_time:.2f} seconds\nNumber of words: {word_count}"
 
-    return transcription.strip(), result
+    return text, result
 
 # [VERSION 3: full-on w/ 3 styles for summarization]
 import nltk
@@ -56,11 +65,6 @@ nltk.download('stopwords')
 
 WHITESPACE_HANDLER = lambda k: re.sub('\s+', ' ', re.sub('\n+', ' ', k.strip()))
 
-def transcribe(audio_file):
-    transcription, result = transcribe_audio(audio_file)
-    text = transcription
-    return text, result
-
 def clean_text(text):
     text = re.sub(r'https?:\/\/.*[\r\n]*', '', str(text), flags=re.MULTILINE)
     text = re.sub(r'\<a href', ' ', str(text))
@@ -96,7 +100,7 @@ summarization_model = AutoModelForSeq2SeqLM.from_pretrained("t5-base", return_di
 summarization_tokenizer = AutoTokenizer.from_pretrained("t5-base")
 summarization_model.to(device)
 
-@spaces.GPU(queue=True)
+# @spaces.GPU(queue=True)
 def summarize_text(text):
     preprocessed_text = preprocess_text(text)
     if preprocessed_text is None:
@@ -174,13 +178,13 @@ import gradio as gr
 from fpdf import FPDF
 from PIL import Image
 
-def save_to_pdf(transcription, summary):
+def save_to_pdf(text, summary):
     pdf = FPDF()
     pdf.add_page()
     pdf.set_font("Arial", size=12)
 
-    if transcription:
-        pdf.multi_cell(0, 10, "Transcription:\n" + transcription)
+    if text:
+        pdf.multi_cell(0, 10, "text:\n" + text)
 
     # paragraph space
     pdf.ln(10)
@@ -194,16 +198,16 @@ def save_to_pdf(transcription, summary):
 
 banner_html = """
 <div style="text-align: center;">
-    <img src="https://huggingface.co/spaces/camparchimedes/transcription_app/raw/main/picture.png" alt="Banner" width="100%" height="auto">
+    <img src="https://huggingface.co/spaces/camparchimedes/text_app/raw/main/picture.png" alt="Banner" width="100%" height="auto">
 </div>
 """
 
 iface = gr.Interface(
     fn=transcribe_audio,
     inputs=gr.Audio(type="filepath"),
-    outputs="text",
+    outputs="transcription",
     title="SW Transcription App",
-    description="Upload an audio file to get the transcription",
+    description="Upload an audio file to get the text",
     theme="default",
     live=False
 )
@@ -218,17 +222,18 @@ with iface:
 
         with gr.TabItem("Transcription"):
             audio_input = gr.Audio(type="filepath")
-            transcription_output = gr.Textbox(label="Transcription")
+            text_output = gr.Textbox(label="text")
             result_output = gr.Textbox(label="Time taken and Number of words")
             transcribe_button = gr.Button("Transcribe")
 
             def transcribe(audio_file):
-                transcription, result = transcribe_audio(audio_file)
-                return transcription, result
+                text, result = transcribe_audio(audio_file)
+                return text, result
+
             transcribe_button.click(
                 fn=transcribe,
                 inputs=[audio_input],
-                outputs=[transcription_output, result_output]
+                outputs=[text_output, result_output]
             )
 
 
@@ -236,15 +241,15 @@ with iface:
             summary_output = gr.Textbox(label="Summary | Graph-based")
             summarize_button = gr.Button("Summarize")
 
-            def summarize(transcription):
-                if not transcription:
-                    return "Warning: a transcription must be available."
-                summary = graph_based_summary(transcription)
+            def summarize(text):
+                if not text:
+                    return "Warning: a text must be available."
+                summary = graph_based_summary(text)
                 return summary
 
             summarize_button.click(
                 fn=summarize,
-                inputs=[transcription_output],
+                inputs=[text_output],
                 outputs=summary_output
             )
 
@@ -252,15 +257,15 @@ with iface:
             summary_output = gr.Textbox(label="Summary | LexRank")
             summarize_button = gr.Button("Summarize")
 
-            def summarize(transcription):
-                if not transcription:
-                    return "Warning: a transcription must be available."
-                summary = lex_rank_summary(transcription)
+            def summarize(text):
+                if not text:
+                    return "Warning: a text must be available."
+                summary = lex_rank_summary(text)
                 return summary
 
             summarize_button.click(
                 fn=summarize,
-                inputs=[transcription_output],
+                inputs=[text_output],
                 outputs=summary_output
             )
 
@@ -268,40 +273,40 @@ with iface:
             summary_output = gr.Textbox(label="Summary | TextRank")
             summarize_button = gr.Button("Summarize")
 
-            def summarize(transcription):
-                if not transcription:
-                    return "Warning: a transcription must be available."
-                summary = text_rank_summary(transcription)
+            def summarize(text):
+                if not text:
+                    return "Warning: a text must be available."
+                summary = text_rank_summary(text)
                 return summary
 
             summarize_button.click(
                 fn=summarize,
-                inputs=[transcription_output],
+                inputs=[text_output],
                 outputs=summary_output
             )
 
         with gr.TabItem("Download PDF"):
-            pdf_transcription_only = gr.Button("Download PDF with Transcription Only")
+            pdf_text_only = gr.Button("Download PDF with text Only")
             pdf_summary_only = gr.Button("Download PDF with Summary Only")
             pdf_both = gr.Button("Download PDF with Both")
 
-            pdf_output_transcription_only = gr.File(label="Download PDF")
+            pdf_output_text_only = gr.File(label="Download PDF")
             pdf_output_summary_only = gr.File(label="Download PDF")
             pdf_output_both = gr.File(label="Download PDF")
 
-            def generate_pdf_transcription_only(transcription):
-                return save_to_pdf(transcription, "")
+            def generate_pdf_text_only(text):
+                return save_to_pdf(text, "")
 
             def generate_pdf_summary_only(summary):
                 return save_to_pdf("", summary)
 
-            def generate_pdf_both(transcription, summary):
-                return save_to_pdf(transcription, summary)
+            def generate_pdf_both(text, summary):
+                return save_to_pdf(text, summary)
 
-            pdf_transcription_only.click(
-                fn=generate_pdf_transcription_only,
-                inputs=[transcription_output],
-                outputs=[pdf_output_transcription_only]
+            pdf_text_only.click(
+                fn=generate_pdf_text_only,
+                inputs=[text_output],
+                outputs=[pdf_output_text_only]
             )
 
             pdf_summary_only.click(
@@ -312,9 +317,8 @@ with iface:
 
             pdf_both.click(
                 fn=generate_pdf_both,
-                inputs=[transcription_output, summary_output],
+                inputs=[text_output, summary_output],
                 outputs=[pdf_output_both]
             )
 
 iface.launch(share=True, debug=True)
-