|
|
|
|
from swarms.models import OpenAIChat
|
|
|
|
|
from swarms.structs import Flow
|
|
|
|
|
|
|
|
|
|
api_key = "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"
|
|
|
|
|
|
|
|
|
|
# Initialize the language model, this model can be swapped out with Anthropic, ETC, Huggingface Models like Mistral, ETC
|
|
|
|
|
llm = OpenAIChat(
|
|
|
|
|
# model_name="gpt-4"
|
|
|
|
|
openai_api_key=api_key,
|
|
|
|
|
temperature=0.5,
|
|
|
|
|
# max_tokens=100,
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
## Initialize the workflow
|
|
|
|
|
flow = Flow(
|
|
|
|
|
llm=llm,
|
|
|
|
|
max_loops=2,
|
|
|
|
|
dashboard=True,
|
|
|
|
|
# stopping_condition=None, # You can define a stopping condition as needed.
|
|
|
|
|
# loop_interval=1,
|
|
|
|
|
# retry_attempts=3,
|
|
|
|
|
# retry_interval=1,
|
|
|
|
|
# interactive=False, # Set to 'True' for interactive mode.
|
|
|
|
|
# dynamic_temperature=False, # Set to 'True' for dynamic temperature handling.
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
# out = flow.load_state("flow_state.json")
|
|
|
|
|
# temp = flow.dynamic_temperature()
|
|
|
|
|
# filter = flow.add_response_filter("Trump")
|
|
|
|
|
out = flow.run("Generate a 10,000 word blog on health and wellness.")
|
|
|
|
|
# out = flow.validate_response(out)
|
|
|
|
|
# out = flow.analyze_feedback(out)
|
|
|
|
|
# out = flow.print_history_and_memory()
|
|
|
|
|
# # out = flow.save_state("flow_state.json")
|
|
|
|
|
# print(out)
|
