from typing import Dict, List, Any from transformers import pipeline from PIL import Image import requests from transformers import AutoModelForCausalLM, LlamaTokenizer import torch # from accelerate import ( # init_empty_weights, # infer_auto_device_map, # load_checkpoint_and_dispatch, # ) import os import logging # from transformers import logging as hf_logging # hf_logging.set_verbosity_debug() logging.basicConfig(level=logging.INFO) class EndpointHandler: def __init__(self, path=""): # Preload all the elements you are going to need at inference. # self.pipeline = pipeline( # "text-generation", model="THUDM/cogvlm-chat-hf", trust_remote_code=True # ) # self.model = AutoModelForCausalLM.from_pretrained( # "THUDM/cogvlm-chat-hf", trust_remote_code=True # ) self.tokenizer = LlamaTokenizer.from_pretrained("lmsys/vicuna-7b-v1.5") self.model = ( AutoModelForCausalLM.from_pretrained( "THUDM/cogvlm-chat-hf", torch_dtype=torch.bfloat16, low_cpu_mem_usage=True, trust_remote_code=True, ) .to("cuda") .eval() ) # DISTRIBUTED GPUS # with init_empty_weights(): # self.model = AutoModelForCausalLM.from_pretrained( # "THUDM/cogvlm-chat-hf", # torch_dtype=torch.bfloat16, # low_cpu_mem_usage=True, # trust_remote_code=True, # ) # # print("LISTING FILES IN ", "/root/.cache/huggingface") # # list_files("/root/.cache/huggingface", 0, 5) # device_map = infer_auto_device_map( # self.model, # max_memory={ # 0: "12GiB", # 1: "12GiB", # 2: "12GiB", # 3: "12GiB", # "cpu": "180GiB", # }, # no_split_module_classes=["CogVLMDecoderLayer"], # ) # self.model = load_checkpoint_and_dispatch( # self.model, # "/root/.cache/huggingface/hub/models--THUDM--cogvlm-chat-hf/snapshots/8abca878c4257412c4c38eeafaed3fe27a036730", # device_map=device_map, # no_split_module_classes=["CogVLMDecoderLayer"], # ) # self.model = self.model.eval() ## DISTRIBUTED GPUS def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]: """ data args: inputs (:obj: `str` | `PIL.Image` | `np.array`) kwargs Return: A :obj:`list` | `dict`: will be serialized and returned """ query = data["inputs"] img_uri = data["img_uri"] image = Image.open( requests.get( img_uri, stream=True, ).raw ).convert("RGB") inputs = self.model.build_conversation_input_ids( self.tokenizer, query=query, history=[], images=[image], template_version="vqa", ) # vqa mode inputs = { "input_ids": inputs["input_ids"].unsqueeze(0).to("cuda"), "token_type_ids": inputs["token_type_ids"].unsqueeze(0).to("cuda"), "attention_mask": inputs["attention_mask"].unsqueeze(0).to("cuda"), "images": [[inputs["images"][0].to("cuda").to(torch.bfloat16)]], } gen_kwargs = {"max_length": 2048, "do_sample": False} with torch.no_grad(): outputs = self.model.generate(**inputs, **gen_kwargs) logging.info(f"OUTPUTS 1: {outputs} length: {outputs.shape}") outputs = outputs[:, inputs["input_ids"].shape[1] :] logging.info(f"OUTPUTS 2: {outputs.shape}") response = self.tokenizer.decode(outputs[0]) return response # query = "How many houses are there in this cartoon?" # image = Image.open( # requests.get( # "https://github.com/THUDM/CogVLM/blob/main/examples/3.jpg?raw=true", stream=True # ).raw # ).convert("RGB") # inputs = model.build_conversation_input_ids( # tokenizer, query=query, history=[], images=[image], template_version="vqa" # ) # vqa mode # inputs = { # "input_ids": inputs["input_ids"].unsqueeze(0).to("cuda"), # "token_type_ids": inputs["token_type_ids"].unsqueeze(0).to("cuda"), # "attention_mask": inputs["attention_mask"].unsqueeze(0).to("cuda"), # "images": [[inputs["images"][0].to("cuda").to(torch.bfloat16)]], # } # gen_kwargs = {"max_length": 2048, "do_sample": False} # with torch.no_grad(): # outputs = model.generate(**inputs, **gen_kwargs) # outputs = outputs[:, inputs["input_ids"].shape[1] :] # print(tokenizer.decode(outputs[0]))