joy-caption-pre-alpha

joy-caption-pre-alpha.py

#import spaces
#import gradio as gr
from huggingface_hub import InferenceClient
from torch import nn
from transformers import AutoModel, AutoProcessor, AutoTokenizer, PreTrainedTokenizer, PreTrainedTokenizerFast, AutoModelForCausalLM
from pathlib import Path
import torch
import torch.amp.autocast_mode
from PIL import Image
import os, time


CLIP_PATH = "google/siglip-so400m-patch14-384"
VLM_PROMPT = "A short caption for this image:\n"
MODEL_PATH = "NousResearch/Meta-Llama-3.1-8B"#"mlabonne/Daredevil-8B"#"meta-llama/Meta-Llama-3.1-8B"
CHECKPOINT_PATH = Path("wpkklhc6")
TITLE = "<h1><center>JoyCaption Pre-Alpha (2024-07-30a)</center></h1>"
HF_TOKEN = os.environ.get("HF_TOKEN", None)

MAX_NEW_TOKENS = 200
IMAGE_FOLDER = "/workspace/1"

class ImageAdapter(nn.Module):
	def __init__(self, input_features: int, output_features: int):
		super().__init__()
		self.linear1 = nn.Linear(input_features, output_features)
		self.activation = nn.GELU()
		self.linear2 = nn.Linear(output_features, output_features)
	
	def forward(self, vision_outputs: torch.Tensor):
		x = self.linear1(vision_outputs)
		x = self.activation(x)
		x = self.linear2(x)
		return x


# Load CLIP
print("Loading CLIP")
clip_processor = AutoProcessor.from_pretrained(CLIP_PATH)
clip_model = AutoModel.from_pretrained(CLIP_PATH)
clip_model = clip_model.vision_model
clip_model.eval()
clip_model.requires_grad_(False)
clip_model.to("cuda")

# Tokenizer
print("Loading tokenizer")
tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH, use_fast=False)
assert isinstance(tokenizer, PreTrainedTokenizer) or isinstance(tokenizer, PreTrainedTokenizerFast), f"Tokenizer is of type {type(tokenizer)}"

# LLM
print("Loading LLM")
text_model = AutoModelForCausalLM.from_pretrained(MODEL_PATH, device_map="auto", torch_dtype=torch.bfloat16)
text_model.eval()

# Image Adapter
print("Loading image adapter")
image_adapter = ImageAdapter(clip_model.config.hidden_size, text_model.config.hidden_size)
image_adapter.load_state_dict(torch.load(CHECKPOINT_PATH / "image_adapter.pt", map_location="cpu"))
image_adapter.eval()
image_adapter.to("cuda")

torch.cuda.empty_cache()
# Tokenize the prompt
prompt = tokenizer.encode(VLM_PROMPT, return_tensors='pt', padding=False, truncation=False, add_special_tokens=False)
# Embed prompt
prompt_embeds = text_model.model.embed_tokens(prompt.to('cuda'))
assert prompt_embeds.shape == (1, prompt.shape[1], text_model.config.hidden_size), f"Prompt shape is {prompt_embeds.shape}, expected {(1, prompt.shape[1], text_model.config.hidden_size)}"
embedded_bos = text_model.model.embed_tokens(torch.tensor([[tokenizer.bos_token_id]], device=text_model.device, dtype=torch.int64))

def caption_get(image_path):
    input_image = Image.open(image_path)
    # Preprocess image
    image = clip_processor(images=input_image, return_tensors='pt').pixel_values
    image = image.to('cuda')
    # Embed image
    with torch.amp.autocast_mode.autocast('cuda', enabled=True):
        vision_outputs = clip_model(pixel_values=image, output_hidden_states=True)
        image_features = vision_outputs.hidden_states[-2]
        embedded_images = image_adapter(image_features)
        embedded_images = embedded_images.to('cuda')

    # Construct prompts
    inputs_embeds = torch.cat([
		embedded_bos.expand(embedded_images.shape[0], -1, -1),
		embedded_images.to(dtype=embedded_bos.dtype),
		prompt_embeds.expand(embedded_images.shape[0], -1, -1),
	], dim=1)

    input_ids = torch.cat([
		torch.tensor([[tokenizer.bos_token_id]], dtype=torch.long),
		torch.zeros((1, embedded_images.shape[1]), dtype=torch.long),
		prompt,
	], dim=1).to('cuda')
    attention_mask = torch.ones_like(input_ids)

    generate_ids = text_model.generate(input_ids, inputs_embeds=inputs_embeds, attention_mask=attention_mask, max_new_tokens=MAX_NEW_TOKENS, do_sample=True, top_k=10, temperature=0.5, suppress_tokens=None)

	# Trim off the prompt
    generate_ids = generate_ids[:, input_ids.shape[1]:]
    if generate_ids[0][-1] == tokenizer.eos_token_id:
        generate_ids = generate_ids[:, :-1]

    caption = tokenizer.batch_decode(generate_ids, skip_special_tokens=False, clean_up_tokenization_spaces=False)[0]

    return caption.strip()

# Чтение всех изображений из папки
image_folder = Path(IMAGE_FOLDER).resolve()
jpeg_images = image_folder.glob('*.jpeg')
jpg_images = image_folder.glob('*.jpg')
png_images = image_folder.glob('*.png')
all_images = list(jpeg_images) + list(png_images) + list(jpg_images)
# Фильтрация списка all_images, чтобы оставить только файлы, начинающиеся с 
#all_images = [img for img in all_images if img.name.startswith('nus')]
all_images = sorted(all_images)
#all_images = all_images[:10]

num_images = len(all_images)
start_time = time.time()
images_processed = 0
for image_path in all_images:  # Для файлов .jpg, измените расширение при необходимости
    images_processed+=1
    caption = caption_get(image_path)
    text_filename = str(image_path.with_suffix('.txt'))
    with open(text_filename, 'w' if Path(text_filename).exists() else 'w') as file_text:
        file_text.write(f"{caption}")
    if images_processed % 1 ==0:
        elapsed_time = time.time() - start_time
        estimated_total_time = (elapsed_time / images_processed) * num_images
        remaining_time = estimated_total_time - elapsed_time
        print(f"File: {image_path}, Caption: {caption}\n")
        print(f"Processed {images_processed}/{num_images} files, approximate remaining time: {time.strftime('%H:%M:%S', time.gmtime(remaining_time))}")