Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,34 +1,125 @@
|
|
1 |
import gradio as gr
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
|
4 |
-
#
|
5 |
-
|
6 |
-
tokenizer = LEDTokenizer.from_pretrained(model_name)
|
7 |
-
model = LEDForConditionalGeneration.from_pretrained(model_name)
|
8 |
|
9 |
-
|
10 |
-
|
11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
13 |
-
|
14 |
-
|
15 |
-
inputs["input_ids"],
|
16 |
-
num_beams=4,
|
17 |
-
max_length=512, # Can be adjusted based on summary size needs
|
18 |
-
min_length=100,
|
19 |
-
early_stopping=True
|
20 |
-
)
|
21 |
|
22 |
-
|
23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
|
25 |
# Gradio Interface
|
26 |
iface = gr.Interface(
|
27 |
fn=summarize_text,
|
28 |
inputs="text",
|
29 |
outputs="text",
|
30 |
-
title="
|
31 |
-
description="
|
32 |
)
|
33 |
|
34 |
if __name__ == "__main__":
|
|
|
1 |
import gradio as gr
|
2 |
+
import os
|
3 |
+
import requests
|
4 |
+
import torch
|
5 |
+
from transformers import (
|
6 |
+
LEDTokenizer, LEDForConditionalGeneration,
|
7 |
+
BartTokenizer, BartForConditionalGeneration,
|
8 |
+
PegasusTokenizer, PegasusForConditionalGeneration,
|
9 |
+
AutoTokenizer, AutoModelForSeq2SeqLM
|
10 |
+
)
|
11 |
|
12 |
+
# OpenAI API Key
|
13 |
+
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY") # Ensure this is set in your environment variables
|
|
|
|
|
14 |
|
15 |
+
# List of models in priority order
|
16 |
+
MODELS = [
|
17 |
+
{
|
18 |
+
"name": "allenai/led-large-16384",
|
19 |
+
"tokenizer_class": LEDTokenizer,
|
20 |
+
"model_class": LEDForConditionalGeneration
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"name": "facebook/bart-large-cnn",
|
24 |
+
"tokenizer_class": BartTokenizer,
|
25 |
+
"model_class": BartForConditionalGeneration
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"name": "Falconsai/text_summarization",
|
29 |
+
"tokenizer_class": AutoTokenizer,
|
30 |
+
"model_class": AutoModelForSeq2SeqLM
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"name": "google/pegasus-xsum",
|
34 |
+
"tokenizer_class": PegasusTokenizer,
|
35 |
+
"model_class": PegasusForConditionalGeneration
|
36 |
+
}
|
37 |
+
]
|
38 |
+
|
39 |
+
# Load models sequentially
|
40 |
+
loaded_models = []
|
41 |
+
for model_info in MODELS:
|
42 |
+
try:
|
43 |
+
tokenizer = model_info["tokenizer_class"].from_pretrained(model_info["name"])
|
44 |
+
model = model_info["model_class"].from_pretrained(model_info["name"])
|
45 |
+
loaded_models.append({"name": model_info["name"], "tokenizer": tokenizer, "model": model})
|
46 |
+
print(f"Loaded model: {model_info['name']}")
|
47 |
+
except Exception as e:
|
48 |
+
print(f"Failed to load {model_info['name']}: {e}")
|
49 |
+
|
50 |
+
def summarize_with_transformers(text):
|
51 |
+
"""
|
52 |
+
Try summarizing with locally loaded Transformer models in order of priority.
|
53 |
+
"""
|
54 |
+
for model_data in loaded_models:
|
55 |
+
try:
|
56 |
+
tokenizer = model_data["tokenizer"]
|
57 |
+
model = model_data["model"]
|
58 |
+
|
59 |
+
# Tokenize input with truncation
|
60 |
+
inputs = tokenizer([text], max_length=16384, return_tensors="pt", truncation=True)
|
61 |
+
|
62 |
+
# Generate summary
|
63 |
+
summary_ids = model.generate(
|
64 |
+
inputs["input_ids"],
|
65 |
+
num_beams=4,
|
66 |
+
max_length=512,
|
67 |
+
min_length=100,
|
68 |
+
early_stopping=True
|
69 |
+
)
|
70 |
|
71 |
+
summary = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
|
72 |
+
return summary # Return the first successful response
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
|
74 |
+
except Exception as e:
|
75 |
+
print(f"Error using {model_data['name']}: {e}")
|
76 |
+
|
77 |
+
return None # Indicate failure
|
78 |
+
|
79 |
+
def summarize_with_chatgpt(text):
|
80 |
+
"""
|
81 |
+
Fallback to OpenAI ChatGPT API if all other models fail.
|
82 |
+
"""
|
83 |
+
if not OPENAI_API_KEY:
|
84 |
+
return "Error: No OpenAI API key provided."
|
85 |
+
|
86 |
+
headers = {
|
87 |
+
"Authorization": f"Bearer {OPENAI_API_KEY}",
|
88 |
+
"Content-Type": "application/json"
|
89 |
+
}
|
90 |
+
|
91 |
+
payload = {
|
92 |
+
"model": "gpt-3.5-turbo",
|
93 |
+
"messages": [{"role": "user", "content": f"Summarize this article: {text}"}],
|
94 |
+
"max_tokens": 512
|
95 |
+
}
|
96 |
+
|
97 |
+
response = requests.post("https://api.openai.com/v1/chat/completions", headers=headers, json=payload)
|
98 |
+
|
99 |
+
if response.status_code == 200:
|
100 |
+
return response.json()["choices"][0]["message"]["content"]
|
101 |
+
else:
|
102 |
+
return f"Error: Failed to summarize with ChatGPT (status {response.status_code})"
|
103 |
+
|
104 |
+
def summarize_text(text):
|
105 |
+
"""
|
106 |
+
Main function to summarize text, trying Transformer models first, then ChatGPT if needed.
|
107 |
+
"""
|
108 |
+
summary = summarize_with_transformers(text)
|
109 |
+
|
110 |
+
if summary:
|
111 |
+
return summary # Return successful summary from a Transformer model
|
112 |
+
|
113 |
+
print("All Transformer models failed. Falling back to ChatGPT...")
|
114 |
+
return summarize_with_chatgpt(text) # Use ChatGPT as last resort
|
115 |
|
116 |
# Gradio Interface
|
117 |
iface = gr.Interface(
|
118 |
fn=summarize_text,
|
119 |
inputs="text",
|
120 |
outputs="text",
|
121 |
+
title="Multi-Model Summarizer with Fallback",
|
122 |
+
description="Tries multiple models for summarization, falling back to ChatGPT if needed."
|
123 |
)
|
124 |
|
125 |
if __name__ == "__main__":
|