Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,117 @@
|
|
1 |
---
|
2 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: mit
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
tags:
|
6 |
+
- text-2-text
|
7 |
+
- natural-language
|
8 |
+
- nlp
|
9 |
+
- classification
|
10 |
+
- call center
|
11 |
+
- IT
|
12 |
+
- summarization
|
13 |
+
- text-generation
|
14 |
---
|
15 |
+
# SITCC-T5-Classifier Model Card
|
16 |
+
|
17 |
+
## Model Description
|
18 |
+
The SITCC-T5-Classifier model is a fine-tuned version of the google/flan-t5-base model. It has been specifically trained to process IT ticket descriptions and extract the request/issue and the software/system that the ticket is about. The model was fine-tuned using 5716 synthetically generated input/output pairs generated with OpenAI GPT-4 Turbo.
|
19 |
+
|
20 |
+
## Model Details
|
21 |
+
- Base Model: google/flan-t5-base
|
22 |
+
- Fine-tuning Data: 5716 synthetic IT ticket description pairs generated by OpenAI GPT-4 Turbo
|
23 |
+
|
24 |
+
## Intended Use
|
25 |
+
The SITCC-T5-Classifier model is designed to be used for IT ticket classification and information extraction tasks. It can be used to automatically identify the request/issue and the software/system mentioned in an IT ticket description.
|
26 |
+
|
27 |
+
## Limitations and Known Issues
|
28 |
+
- The model's performance may vary depending on the quality and diversity of the input IT ticket descriptions.
|
29 |
+
- The model may struggle with understanding complex or ambiguous ticket descriptions.
|
30 |
+
- The model may not perform well on ticket descriptions that are significantly different from the training data.
|
31 |
+
|
32 |
+
## Example Usage
|
33 |
+
This example is running on cpu
|
34 |
+
``` python
|
35 |
+
import re
|
36 |
+
import pandas as pd
|
37 |
+
from transformers import T5Tokenizer, T5ForConditionalGeneration
|
38 |
+
|
39 |
+
from time import perf_counter
|
40 |
+
|
41 |
+
class SITCC_T5_Classifier:
|
42 |
+
"""
|
43 |
+
A class for classifying text using the SITCC T5 model.
|
44 |
+
|
45 |
+
Attributes:
|
46 |
+
tokenizer (T5Tokenizer): The tokenizer for the T5 model.
|
47 |
+
model (T5ForConditionalGeneration): The T5 model for classification.
|
48 |
+
"""
|
49 |
+
|
50 |
+
def __init__(self):
|
51 |
+
# Load the tokenizer and model from the fine-tuned model directory
|
52 |
+
self.tokenizer = T5Tokenizer.from_pretrained("KameronB/sitcc-t5-classifier")
|
53 |
+
self.model = T5ForConditionalGeneration.from_pretrained("KameronB/sitcc-t5-classifier", device_map="cpu")
|
54 |
+
|
55 |
+
def process_response(self, response:str) -> dict:
|
56 |
+
"""
|
57 |
+
Process the response and extract the software/system and issue/request.
|
58 |
+
|
59 |
+
Args:
|
60 |
+
response (str): The response text.
|
61 |
+
|
62 |
+
Returns:
|
63 |
+
dict: A dictionary containing the software/system and issue/request.
|
64 |
+
"""
|
65 |
+
matches = re.search(r'Software/System: (.*) Issue/Request: (.*)</s>', response, re.DOTALL)
|
66 |
+
return {
|
67 |
+
"Software/System": matches.group(1),
|
68 |
+
"Issue/Request": matches.group(2)
|
69 |
+
}
|
70 |
+
|
71 |
+
def classify_entry(self, entry:str, max_new_tokens=60) -> dict:
|
72 |
+
"""
|
73 |
+
Classify the input text and return the classification results.
|
74 |
+
|
75 |
+
Args:
|
76 |
+
entry (str): The input text to be classified.
|
77 |
+
max_new_tokens (int): The maximum number of tokens to generate.
|
78 |
+
|
79 |
+
Returns:
|
80 |
+
dict: The classification results.
|
81 |
+
"""
|
82 |
+
# Tokenize the input text
|
83 |
+
input_ids = self.tokenizer(entry, return_tensors="pt").input_ids.to("cpu")
|
84 |
+
|
85 |
+
# Generate the output text
|
86 |
+
outputs = self.model.generate(input_ids, max_new_tokens=max_new_tokens)
|
87 |
+
|
88 |
+
# Decode and return the output text
|
89 |
+
return self.process_response(self.tokenizer.decode(outputs[0]))
|
90 |
+
|
91 |
+
# Create the SITCC T5 Classifier wrapper class for the fine-tuned T5 model
|
92 |
+
sitcc_t5 = SITCC_T5_Classifier()
|
93 |
+
|
94 |
+
# Define the input text
|
95 |
+
|
96 |
+
input_text = [
|
97 |
+
"The customer is getting the following error when using rSATS:\nERROR: 'Failed to connect'. \nI have tried restarting the application and the computer, but the issue persists. \nEscalating to Team",
|
98 |
+
"The customer is experiencing issues with their network connectivity, which is causing slow internet speeds and frequent disconnections.",
|
99 |
+
"The customer is unable to access the shared drive on the network. They receive an error message stating 'Network path not found'. \nEscalating to Network Team",
|
100 |
+
"The customer is unable to print from their computer. They have checked the printer connections and restarted the printer, but the issue persists. \nEscalating to Printer Support Team",
|
101 |
+
]
|
102 |
+
|
103 |
+
# measure the time performance of the model
|
104 |
+
start = perf_counter()
|
105 |
+
for i in range(len(input_text)):
|
106 |
+
# Classify the input text
|
107 |
+
print(sitcc_t5.classify_entry(input_text[i]))
|
108 |
+
|
109 |
+
# measure the time performance of the model
|
110 |
+
end = perf_counter()
|
111 |
+
print(f"Time taken: {end - start} seconds")
|
112 |
+
|
113 |
+
|
114 |
+
|
115 |
+
|
116 |
+
|
117 |
+
```
|