Spaces:
Runtime error
Runtime error
File size: 5,005 Bytes
d3ac18a 90387d3 9aff0f1 bd0fb97 7c2b33c 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 418e706 90387d3 0922798 5b38c2c 418e706 76e7966 7651b2f 1461d74 d57b019 88feec1 1e9d3b4 418e706 1a29ab3 b6517de 90387d3 fb05b43 90387d3 fb05b43 07b0800 03881b3 fd4e200 90387d3 418e706 90387d3 b369e83 76e7966 |
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 |
import openai
import time
import logging
import gradio as gr
import os
from src.llm_boilers import llm_boiler
import configparser
logging.basicConfig(format="%(asctime)s - %(message)s", level=logging.INFO)
logging.warning("READY. App started...")
class Chat:
default_system_prompt = "A conversation between a user and an LLM-based AI assistant. The assistant gives helpful and honest answers."
system_format = "system\n{}\n"
def __init__(self, system: str = None, user: str = None, assistant: str = None):
if system is not None:
self.set_system_prompt(system)
else:
self.reset_system_prompt()
self.user = user if user else "user\n{}\n"
self.assistant = assistant if assistant else "assistant\n{}\n"
self.response_prefix = self.assistant.split("{}")[0]
def set_system_prompt(self, system_prompt):
return system_prompt
def reset_system_prompt(self):
return self.set_system_prompt(self.default_system_prompt)
def history_as_formatted_str(self, system, history):
system = self.system_format.format(system)
text = system + "".join(
[
"\n".join(
[
self.user.format(item[0]),
self.assistant.format(item[1]),
]
)
for item in history[:-1]
]
)
text += self.user.format(history[-1][0])
text += self.response_prefix
# Truncate text if it exceeds the limit
if len(text) > 4096:
text = text[-4096:]
return text
def clear_history(self, history):
return []
def turn(self, user_input: str, history):
self.user_turn(user_input, history)
return self.bot_turn()
def user_turn(self, user_input: str, history):
history.append([user_input, ""])
return user_input, history
def bot_turn(self, system, history, openai_key):
conversation = self.history_as_formatted_str(system, history)
assistant_response = call_inf_server(conversation, openai_key)
history[-1][1] = ""
for chunk in assistant_response:
try:
decoded_output = chunk["choices"][0]["delta"]["content"]
history[-1][1] += decoded_output
yield history
except KeyError:
pass
def call_inf_server(prompt, openai_key):
model_id = "gpt-3.5-turbo"
model = llm_boiler(model_id, openai_key)
logging.warning(f'Inf via "{model_id}"" for prompt "{prompt}"')
try:
# Run text generation
response = model.run(prompt, temperature=1.0)
logging.warning(f"Result of text generation: {response}")
return response
except Exception as e:
# Wait and try one more time
print(e)
time.sleep(2)
response = model.run(prompt, temperature=1.0)
logging.warning(f"Result of text generation: {response}")
return response
# Get the OpenAI key from the environment variable
openai_key = os.getenv("API_KEY")
with gr.Blocks(theme='HaleyCH/HaleyCH_Theme') as demo:
gr.Markdown(
"""
<br><h1><center>Chat with gpt-3.5-turbo</center></h1>
This is a lightweight gpt-3.5-turbo conversation completion.
"""
)
conversation = Chat()
chatbot = gr.Chatbot().style(height=400)
with gr.Row():
with gr.Column():
msg = gr.Textbox(
label="Chat Message Box",
placeholder="Chat Message Box",
show_label=False,
).style(container=False)
with gr.Column():
with gr.Row():
submit = gr.Button("Submit")
stop = gr.Button("Stop")
clear = gr.Button("Clear")
with gr.Row():
with gr.Accordion("Advanced Options:", open=False):
with gr.Row():
with gr.Column(scale=2):
system = gr.Textbox(
label="System Prompt",
value=Chat.default_system_prompt,
show_label=False,
).style(container=False)
with gr.Column():
with gr.Row():
change = gr.Button("Change System Prompt")
reset = gr.Button("Reset System Prompt")
with gr.Row():
gr.Markdown(
"Disclaimer: The gpt-3.5-turbo model can produce factually incorrect output and should not be solely relied on to produce "
"factually accurate information. The gpt-3.5-turbo model was trained on various public datasets; while great efforts "
"have been taken to clean the pretraining data, it is possible that this model could generate lewd, "
"biased, or otherwise offensive outputs.",
elem_classes=["disclaimer"],
)
with gr.Row():
|