Spaces:
Sleeping
Sleeping
# RESPONSE | |
import requests | |
# Define the URL | |
url = "https://wcza44xtt6.execute-api.us-west-2.amazonaws.com/default/llama-osu" | |
def new_data(): | |
data = { | |
"inputs": [ | |
[ | |
] | |
], | |
"parameters": { | |
"max_new_tokens": 500, | |
"top_p": 0.9, # if you set top p to 0.9, the model will only consider the most likely words that make up 90% of the probability mass. | |
"temperature": 0.2 # creative level from 0 to 1 (the higher the more creative) | |
} | |
} | |
return data | |
def func_trim_data(data): | |
trimmed_data = new_data() | |
trimmed_data['inputs'][0] = data['inputs'][0][-9:] | |
return trimmed_data | |
data = new_data() | |
def get_response(prompt: str) -> str: | |
if(prompt.lower() == 'reset'): | |
global data | |
data = new_data() | |
return "You can start a new conversation" | |
else: | |
_dict = {"role": "user", "content": f"{prompt}" + " (Make your answer brief with several sentences)"} | |
data["inputs"][0].append(_dict) | |
response = requests.post(url, json=func_trim_data(data)) | |
response_dict = response.json()[0]['generation'] | |
data["inputs"][0].append(response_dict) | |
return response.json()[0]['generation']['content'] |