Chris Finlayson commited on
Commit
eb2bab3
·
1 Parent(s): 1dfb9d3
application.py ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import PyPDF2
2
+ from transformers import pipeline
3
+ import gradio as gr
4
+
5
+ # Load the summarization pipeline
6
+ summarizer = pipeline("summarization", model="t5-base", tokenizer="t5-base", framework="pt")
7
+
8
+ def load_pdf(file, max_length, min_length, do_sample):
9
+ pdf = PyPDF2.PdfReader(file)
10
+ text = ""
11
+ for page in pdf.pages:
12
+ text += page.extract_text()
13
+ # Summarize the text
14
+
15
+ summary = summarizer(text, max_length=max_length, min_length=min_length, do_sample=do_sample)
16
+ # Return the summary
17
+ return summary[0]['summary_text']
18
+ # return summarize_pdf(pdf)
19
+
20
+
21
+ inputs = [
22
+ gr.File(label="Upload PDF"),
23
+ gr.Slider(500, 1500, value=1000, label="Max Length"),
24
+ gr.Slider(30, 500, value=300, label="Min Length"),
25
+ gr.Checkbox(label="Do Sample", value=False)
26
+ ]
27
+
28
+ iface = gr.Interface(fn=load_pdf, inputs=inputs, outputs="text", title="PDF summariser")
29
+ iface.launch()
flagged/file/905c51933114f0933157/The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf ADDED
Binary file (474 kB). View file
 
flagged/file/e0e08e8da1f761969779/The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf ADDED
Binary file (474 kB). View file
 
flagged/log.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ file,output,flag,username,timestamp
2
+ "{""path"":""flagged/file/905c51933114f0933157/The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf"",""url"":""http://127.0.0.1:7860/file=/private/var/folders/jk/znm3f6kd1xj8w5_2n06stbqm0000gn/T/gradio/2be4325bfe343a3851a3fe2a62720746675b6c90/The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf"",""size"":474394,""orig_name"":""The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf"",""mime_type"":""""}",<PyPDF2._reader.PdfReader object at 0x2aca9eb50>,,,2023-11-10 16:58:42.638294
3
+ "{""path"":""flagged/file/e0e08e8da1f761969779/The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf"",""url"":""http://127.0.0.1:7860/file=/private/var/folders/jk/znm3f6kd1xj8w5_2n06stbqm0000gn/T/gradio/2be4325bfe343a3851a3fe2a62720746675b6c90/The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf"",""size"":474394,""orig_name"":""The_Design_Space_of_Emergent_Scheduling_for_Distributed_Execution_Frameworks.pdf"",""mime_type"":""""}",<PyPDF2._reader.PdfReader object at 0x2b07221d0>,,,2023-11-10 17:05:37.811292
requirements.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ pypdf2
2
+ transformers