-
Notifications
You must be signed in to change notification settings - Fork 1
/
models.py
192 lines (166 loc) · 6.84 KB
/
models.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
from transformers import GPT2TokenizerFast
import openai
import json
with open('./secrets.json') as file:
secrets = json.loads(file.read())
openai.api_key = secrets["apiKey"]
class ChatGPTModel():
def __init__(self):
self.history = [
{
"role": "system",
"content": "Clippy is an AI assistant"
}
]
self.tipModifier = ''
self.tokenizer = GPT2TokenizerFast.from_pretrained("gpt2")
self.sessionCost = 0
# Define AI Options
self.defaultConfig = {
"engine": "gpt-3.5-turbo",
"temperature": 0.75,
"max_tokens": 128,
"top_p": 1,
"frequency_penalty": 0,
"presence_penalty": 0,
"useMemory": True
}
self.config = self.defaultConfig
self.costPerKTokens = 0.002
def reset(self):
self.history = self.history[:1]
def generateTip(self, prompt):
try:
completion = openai.Completion.create(engine=self.model, prompt='\n'.join(self.history))
response = completion.choices[0].text
response = response.strip()
except openai.error.RateLimitError as e:
response = "[RATELIMITED]\nOur server are overloaded!"
def prompt(self, prompt):
if not self.config["useMemory"]:
self.reset()
prompt = prompt.strip()
self.history.append({
"role": "user",
"content": prompt
})
repetitions = 3
while True:
repetitions -= 1
if (repetitions == 0):
response = "[ERROR] Max Retries Exceeded"
break
try:
promptTokenCount = len(self.tokenizer(self.history[-1]["content"])['input_ids'])
completion = openai.ChatCompletion.create(
model=self.config["engine"],
messages=self.history, #prompt
temperature=self.config["temperature"],
max_tokens=self.config["max_tokens"],
top_p=self.config["top_p"],
frequency_penalty=self.config["frequency_penalty"],
presence_penalty=self.config["presence_penalty"],
)
#completion = openai.Completion.create(engine=self.model, prompt='\n'.join(self.history))
responseText = completion.choices[0].message.content
responseText = responseText.strip()
self.history.append({
"role": "assistant",
"content": responseText
})
responseTokenCount = len(self.tokenizer(responseText)['input_ids'])
totalTokenCount = promptTokenCount + responseTokenCount
self.sessionCost += self.costPerKTokens * (totalTokenCount/1000)
print("=========================")
print("Prompt:", prompt)
print("Response:", responseText)
print("Total Tokens",totalTokenCount)
print("Total Cost: $" + '{:.6f}'.format(self.sessionCost))
print("=========================")
break
except openai.error.RateLimitError as e:
responseText = "[RATELIMITED]\nOur server are overloaded! Please wait a bit."
break
except openai.error.InvalidRequestError as e:
self.history = self.history[3:-3]
print(e)
exit()
except Exception as e:
print("Fatal Exception")
print(e)
exit()
return responseText
class OpenAIModel():
def __init__(self):
self.history = ["Clippy is an AI assistant"]
self.tipModifier = ''
self.tokenizer = GPT2TokenizerFast.from_pretrained("gpt2")
# Define AI Options
self.defaultConfig = {
"engine": "text-davinci-003",
"temperature": 0.75,
"max_tokens": 128,
"top_p": 1,
"frequency_penalty": 0,
"presence_penalty": 0,
"useMemory": True
}
self.config = self.defaultConfig
self.costPerKTokens = 0.02
def reset(self):
self.history = self.history[:1]
def generateTip(self, prompt):
try:
completion = openai.Completion.create(engine=self.model, prompt='\n'.join(self.history))
response = completion.choices[0].text
response = response.strip()
except openai.error.RateLimitError as e:
response = "[RATELIMITED]\nOur server are overloaded!"
def prompt(self, prompt):
if not self.config["useMemory"]:
self.reset()
prompt = prompt.strip()
self.history.append("User: " + prompt)
self.history.append("Response: ")
repetitions = 3
while True:
repetitions -= 1
if (repetitions == 0):
response = "[ERROR] Max Retries Exceeded"
break
try:
promptTokenCount = len(self.tokenizer('\n'.join(self.history))['input_ids'])
completion = openai.Completion.create(
engine=self.config["engine"],
prompt='\n'.join(self.history), #prompt
temperature=self.config["temperature"],
max_tokens=self.config["max_tokens"],
top_p=self.config["top_p"],
frequency_penalty=self.config["frequency_penalty"],
presence_penalty=self.config["presence_penalty"],
)
#completion = openai.Completion.create(engine=self.model, prompt='\n'.join(self.history))
response = completion.choices[0].text
response = response.strip()
self.history[-1] = "Response: " + response
responseTokenCount = len(self.tokenizer(response)['input_ids'])
totalTokenCount = promptTokenCount + responseTokenCount
totalCost = totalTokenCount * (self.costPerKTokens/1000)
print("=========================")
print("Prompt:", prompt)
print("Response:", response)
print("Total Tokens",totalTokenCount)
print("Total Cost: $" + '{:.6f}'.format(totalCost))
print("=========================")
break
except openai.error.RateLimitError as e:
response = "[RATELIMITED]\nOur server are overloaded! Please wait a bit."
break
except openai.error.InvalidRequestError as e:
self.history = self.history[3:-3]
print(e)
exit()
except Exception as e:
print(e)
exit()
return response