File size: 2,164 Bytes
897e575 |
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 |
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
from langchain.llms import HuggingFacePipeline
from langchain import PromptTemplate, LLMChain
template = """{char_name}'s Persona: {char_persona}
<START>
{chat_history}
{char_name}: {char_greeting}
<END>
{user_name}: {user_input}
{char_name}: """
class EndpointHandler():
def __init__(self, path=""):
tokenizer = AutoTokenizer.from_pretrained(path)
model = AutoModelForCausalLM.from_pretrained(path, load_in_8bit = True, device_map = "auto")
local_llm = HuggingFacePipeline(
pipeline = pipeline(
"text-generation",
model = model,
tokenizer = tokenizer,
max_length = 2048,
temperature = 0.5,
top_p = 0.9,
top_k = 0,
repetition_penalty = 1.1,
pad_token_id = 50256,
num_return_sequences = 1
)
)
prompt_template = PromptTemplate(
template = template,
input_variables = [
"user_input",
"user_name",
"char_name",
"char_persona",
"char_greeting",
"chat_history"
],
validate_template = True
)
self.llm_engine = LLMChain(
llm = local_llm,
prompt = prompt_template
)
def __call__(self, data):
inputs = data.pop("inputs", data)
try:
response = self.llm_engine.predict(
user_input = inputs["user_input"],
user_name = inputs["user_name"],
char_name = inputs["char_name"],
char_persona = inputs["char_persona"],
char_greeting = inputs["char_greeting"],
chat_history = inputs["chat_history"]
).split("\n",1)[0]
return {
"inputs": inputs,
"text": response
}
except Exception as e:
return {
"inputs": inputs,
"error": str(e)
} |