irthomasthomas
commented
2 months ago import json
import numpy as np
from concurrent.futures import ThreadPoolExecutor
from llm_handler import send_to_llm
from params import OUTPUT_FILE_PATH, NUM_WORKERS, PROVIDER
from system_messages import (
SYSTEM_MESSAGES_ORCA,
SYSTEM_MESSAGES_TESS,
SYSTEM_MESSAGES_CODE,
)
from topics import TOPICS
code_data = False
if code_data:
SYSTEM_MESSAGES = SYSTEM_MESSAGES_CODE
PROMPT_1 = """For the following SUBJECT_AREA, generate a question that covers a very narrow topic in the SUBJECT_AREA, with sufficient depth and breadth. The topic in the question should be important to the SUBJECT_AREA, with known-answers present. The generated question should be detailed, seek true nature of our universe from first principles, curiosity invoking, thought provoking, and also should be able to be answered by an intelligence like yourself. Make sure the question is sufficiently harder and multi-part, like a graduate level course question. The question should seek an answer with computer code."""
else:
SYSTEM_MESSAGES = SYSTEM_MESSAGES_ORCA + SYSTEM_MESSAGES_TESS
PROMPT_1 = """For the following SUBJECT_AREA, generate a question that covers a very narrow topic in the SUBJECT_AREA, with sufficient depth and breadth. The topic in the question should be important to the SUBJECT_AREA, with known-answers present. The generated question should be detailed, seek true nature of our universe from first principles, curiosity invoking, thought provoking, and also should be able to be answered by an intelligence like yourself. Make sure the question is sufficiently harder and multi-part, like a graduate level course question."""
msg_context = {"role": "system", "content": str(PROMPT_1)}
def generate_data(
topic_selected,
system_message_generation,
system_message_selected,
OUTPUT_FILE_PATH,
):
system_contexts = [
system_message_generation,
system_message_selected,
]
user_prompts = [f"SUBJECT_AREA: {topic_selected}"]
gpt_outputs = []
for pp in range(len(system_contexts)):
msg_list = []
msg_system = {"role": "system", "content": str(system_contexts[pp])}
msg_list.append(msg_system)
msg_prompt = {"role": "user", "content": user_prompts[pp]}
msg_list.append(msg_prompt)
llm_response, llm_usage = send_to_llm(PROVIDER, msg_list)
user_prompts.append(llm_response)
gpt_outputs.append(llm_response)
data = {
"system": system_contexts[1],
"instruction": str(user_prompts[1]),
"response": str(gpt_outputs[1]),
}
with open(OUTPUT_FILE_PATH, "a") as output_file:
output_file.write(json.dumps(data) + "\n")
return data, llm_usage
def main():
nn = 0
failed = 0
with ThreadPoolExecutor(max_workers=NUM_WORKERS) as executor:
# Create a list of futures, one for each topic
futures = []
for _ in range(NUM_WORKERS):
topic_number = np.random.randint(0, len(TOPICS))
topic_selected = TOPICS[topic_number]
system_message_number = np.random.randint(0, len(SYSTEM_MESSAGES))
system_message_selected = SYSTEM_MESSAGES[system_message_number]
system_message_generation = PROMPT_1
futures.append(
executor.submit(
generate_data,
topic_selected,
system_message_generation,
system_message_selected,
OUTPUT_FILE_PATH,
)
)
# Wait for all futures to complete
for future in futures:
data, gpt_usage = future.result()
if gpt_usage is not None:
nn += 1
print(data)
print(
f"Generation {nn} Complete, Token usage: {gpt_usage}, Failed: {failed}"
)
else:
failed += 1
print("=" * 132)
while True:
main()
TITLE: Sensei/system_messages.py at main · migtissera/Sensei
SYSTEM_MESSAGES_TESS