InternVL2_5-1B

!pip install -U bitsandbytes -q

!pip install decord

import numpy as np import torch import torchvision.transforms as T from decord import VideoReader, cpu from PIL import Image from torchvision.transforms.functional import InterpolationMode from transformers import AutoModel, AutoTokenizer

Load the model and tokenizer

path = "OpenGVLab/InternVL2_5-1B" model = AutoModel.from_pretrained( path, torch_dtype=torch.bfloat16, load_in_8bit=True, low_cpu_mem_usage=True, use_flash_attn=True, trust_remote_code=True ).eval() # Remove .cuda() here

tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)

import numpy as np import torch import torchvision.transforms as T from decord import VideoReader, cpu from PIL import Image from torchvision.transforms.functional import InterpolationMode from transformers import AutoModel, AutoTokenizer

Define build_transform and dynamic_preprocess functions

def build_transform(input_size=448): """Builds a transformation pipeline for image preprocessing.""" transform = T.Compose([ T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC), T.ToTensor(), T.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]), ]) return transform

def dynamic_preprocess(image, image_size=448, use_thumbnail=True, max_num=12): """Dynamically preprocesses the image.""" # Placeholder for dynamic preprocessing logic (replace with actual implementation) # This example simply returns the original image in a list return [image]

Load the model and tokenizer

path = "OpenGVLab/InternVL2_5-1B" model = AutoModel.from_pretrained( path, torch_dtype=torch.bfloat16, load_in_8bit=True, low_cpu_mem_usage=True, use_flash_attn=True, trust_remote_code=True ).eval() # Remove .cuda() here

tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)

def load_image(image_file, input_size=448, max_num=None): # Parameter max_num set to None. """Loads an image and preprocesses it. If 'max_num' is not None, it is passed to dynamic_preprocess, otherwise a default of 12 is used."""

image = Image.open(image_file).convert('RGB')
transform = build_transform(input_size=input_size)
if max_num:  # Check if max_num is provided
    images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num)
else:
    images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=12) # Default max_num to 12
pixel_values = [transform(image) for image in images]
pixel_values = torch.stack(pixel_values)
return pixel_values

image_path = '/content/1.png' def load_image(image_file, input_size=448, max_num=None): # Parameter max_num set to None. """Loads an image and preprocesses it. If 'max_num' is not None, it is passed to dynamic_preprocess, otherwise a default of 12 is used."""

image = Image.open(image_file).convert('RGB')
transform = build_transform(input_size=input_size)
if max_num:  # Check if max_num is provided
    images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=max_num)
else:
    images = dynamic_preprocess(image, image_size=input_size, use_thumbnail=True, max_num=12) # Default max_num to 12
pixel_values = [transform(image) for image in images]
pixel_values = torch.stack(pixel_values)
return pixel_values

pixel_values = load_image(image_path).to(torch.bfloat16) # Keep pixel_values on CPU generation_config = dict(max_new_tokens=1024, do_sample=True)

question = '\nPlease describe the image shortly.' response = model.chat(tokenizer, pixel_values, question, generation_config) print(f'User: {question}\nAssistant: {response}')