File size: 2,678 Bytes
c7b6a1c 5638f9e c7b6a1c 2419282 608249a 2419282 8506a61 5638f9e 2419282 c7b6a1c 5638f9e 8506a61 5638f9e |
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 69 |
import torch
from typing import Any, Dict
from transformers import AutoModelForCausalLM, AutoTokenizer, AutoConfig
from accelerate import dispatch_model, infer_auto_device_map
from accelerate.utils import get_balanced_memory
class EndpointHandler:
def __init__(self, path=""):
with torch.autocast('cuda'):
path = "oleksandrfluxon/mpt-7b-instruct-evaluate"
config = AutoConfig.from_pretrained(
path,
trust_remote_code=True
)
# config.attn_config['attn_impl'] = 'triton'
config.init_device = 'cuda:0' # For fast initialization directly on GPU!
config.max_seq_len = 4096 # (input + output) tokens can now be up to 4096
# load model and tokenizer from path
self.tokenizer = AutoTokenizer.from_pretrained('EleutherAI/gpt-neox-20b', padding_side="left")
model = AutoModelForCausalLM.from_pretrained(
path,
config,
device_map="auto",
torch_dtype=torch.float16,
trust_remote_code=True,
load_in_8bit=True # Load model in the lowest 4-bit precision quantization
)
max_memory = get_balanced_memory(
model,
max_memory=None,
no_split_module_classes=["MPTBlock"],
dtype='float16',
low_zero=False
)
device_map = infer_auto_device_map(
model,
max_memory=max_memory,
no_split_module_classes=["MPTBlock"],
dtype='float16'
)
self.model = dispatch_model(model, device_map=device_map)
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)
with torch.autocast('cuda'):
# preprocess
# inputs = self.tokenizer(inputs, return_tensors="pt").to(self.device)
inputs = self.tokenizer(inputs, return_tensors="pt")
# pass inputs with all kwargs in data
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}] |