File size: 1,715 Bytes
6ebccb5 13ad266 6ebccb5 13ad266 6ebccb5 8e4e321 6ebccb5 8e4e321 6ebccb5 |
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 |
import torch
from typing import Any, Dict
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers.models.auto import modeling_auto
modeling_auto.MODEL_FOR_CAUSAL_LM_MAPPING_NAMES['falcon'] = 'FalconForCausalLM'
class EndpointHandler:
def __init__(self, path=""):
# load model and tokenizer from path
self.tokenizer = AutoTokenizer.from_pretrained(path)
self.model = AutoModelForCausalLM.from_pretrained(
path, device_map="auto", torch_dtype=torch.float16, trust_remote_code=True
)
self.device = "cuda" if torch.cuda.is_available() else "cpu"
def __call__(self, data: Dict[str, Any]) -> Dict[str, str]:
# process input
inputs = data.pop("inputs", data)
parameters = data.pop("parameters", None)
# preprocess
print(print("inputs......", inputs))
inputs = self.tokenizer(inputs, return_tensors="pt").to(self.device)
t=0
for j in range(len(inputs['token_type_ids'][0])):
if inputs['input_ids'][0][j]==39 and inputs['input_ids'][0][j+1]== 5584:
t=0
if inputs['input_ids'][0][j]==39 and inputs['input_ids'][0][j+1]== 13359:
t=1
inputs['token_type_ids'][0][j]=t
# pass inputs with all kwargs in data
print("inputs......", inputs)
if parameters is not None:
outputs = self.model.generate(**inputs, **parameters)
else:
outputs = self.model.generate(**inputs)
# postprocess the prediction
prediction = self.tokenizer.decode(outputs[0], skip_special_tokens=True)
return [{"generated_text": prediction}] |