Hosting ChatUniVi in Sanic Application: ZeroDivisionError

[ervinweelzwork]

Problem Description

I'm currently working on integrating ChatUniVi into a Sanic web application and encountered an unexpected error during runtime. The error occurs when attempting to load a pretrained model using the load_pretrained_model function from the ChatUniVi package. The traceback provided below outlines the sequence of events leading to the error.

Error Traceback

Traceback (most recent call last):
  File "C:\Users\Ervin\Documents\Chat-UniVi\main.py", line 22, in <module>
    tokenizer, _, _, _ = load_pretrained_model(model_path, None, model_name)
  File "C:\Users\Ervin\Documents\Chat-UniVi\ChatUniVi\model\builder.py", line 75, in load_pretrained_model
    model = AutoModelForCausalLM.from_pretrained(model_path, low_cpu_mem_usage=True, **kwargs)
  File "C:\Users\Ervin\anaconda3\envs\chatunivi\lib\site-packages\transformers\models\auto\auto_factory.py", line 563, in from_pretrained
    return model_class.from_pretrained(
  File "C:\Users\Ervin\anaconda3\envs\chatunivi\lib\site-packages\transformers\modeling_utils.py", line 3685, in from_pretrained
    max_memory = get_balanced_memory(
  File "C:\Users\Ervin\anaconda3\envs\chatunivi\lib\site-packages\accelerate\utils\modeling.py", line 753, in get_balanced_memory
    per_gpu = module_sizes[""] // (num_devices - 1 if low_zero else num_devices)
ZeroDivisionError: integer division or modulo by zero

Steps to Reproduce

1. Ensure all dependencies are installed and up-to-date.
2. Set the model_path and model_name variables in main.py.
3. Run the application. Error immediately appears

Expected Behavior
The application should successfully load the pretrained model and serve it through the defined endpoints (/img and /vid) without throwing errors.

Environment
Python version: 3.10.4
Sanic version: 24.6.0
Operating System: Windows 10

Requested Action
Could you please provide guidance on how to resolve this issue? Thank you for your assistance.

main.py

import torch
import os
from ChatUniVi.constants import *
from ChatUniVi.conversation import conv_templates, SeparatorStyle
from ChatUniVi.model.builder import load_pretrained_model
from ChatUniVi.utils import disable_torch_init
from ChatUniVi.mm_utils import tokenizer_image_token, get_model_name_from_path, KeywordsStoppingCriteria
from PIL import Image
from decord import VideoReader, cpu
import numpy as np

from sanic import Sanic, json, response
from sanic.log import logger

app = Sanic("ChatUnivi_SANIC")


global model, tokenizer, image_processor
model_path = "C:/Users/Ervin/Documents/Chat-UniVi/model"  # Adjust the path as necessary
model_name = "ChatUniVi"
_, model, image_processor, _ = load_pretrained_model(model_path, None, model_name , low_cpu_mem_usage=False)
tokenizer, _, _, _ = load_pretrained_model(model_path, None, model_name, low_cpu_mem_usage=False)
model.eval()  # Set the model to evaluation mode




def _get_rawvideo_dec(video_path, image_processor, max_frames=MAX_IMAGE_LENGTH, image_resolution=224, video_framerate=1, s=None, e=None):
    # speed up video decode via decord.

    if s is None:
        start_time, end_time = None, None
    else:
        start_time = int(s)
        end_time = int(e)
        start_time = start_time if start_time >= 0. else 0.
        end_time = end_time if end_time >= 0. else 0.
        if start_time > end_time:
            start_time, end_time = end_time, start_time
        elif start_time == end_time:
            end_time = start_time + 1

    if os.path.exists(video_path):
        vreader = VideoReader(video_path, ctx=cpu(0))
    else:
        print(video_path)
        raise FileNotFoundError

    fps = vreader.get_avg_fps()
    f_start = 0 if start_time is None else int(start_time * fps)
    f_end = int(min(1000000000 if end_time is None else end_time * fps, len(vreader) - 1))
    num_frames = f_end - f_start + 1
    if num_frames > 0:
        # T x 3 x H x W
        sample_fps = int(video_framerate)
        t_stride = int(round(float(fps) / sample_fps))

        all_pos = list(range(f_start, f_end + 1, t_stride))
        if len(all_pos) > max_frames:
            sample_pos = [all_pos[_] for _ in np.linspace(0, len(all_pos) - 1, num=max_frames, dtype=int)]
        else:
            sample_pos = all_pos

        patch_images = [Image.fromarray(f) for f in vreader.get_batch(sample_pos).asnumpy()]

        patch_images = torch.stack([image_processor.preprocess(img, return_tensors='pt')['pixel_values'][0] for img in patch_images])
        slice_len = patch_images.shape[0]

        return patch_images, slice_len
    else:
        print("video path: {} error.".format(video_path))


@app.post('/img')
async def img(request):
        # Model Parameter
        # model_path = "C:/Users/Ervin/Documents/Chat-UniVi/model"  # or "Chat-UniVi/Chat-UniVi-13B"
        # model_path = "Chat-UniVi/Chat-UniVi"  # or "Chat-UniVi/Chat-UniVi-13B"
        image_path = "sam.jpg"

        # Input Text
        qs = "Describe the image."

        # Sampling Parameter
        conv_mode = "simple"
        temperature = 0.2
        top_p = None
        num_beams = 1

        disable_torch_init()
        # model_path = os.path.expanduser(model_path)
        # model_name = "ChatUniVi"
        # tokenizer, model, image_processor, context_len = load_pretrained_model(model_path, None, model_name)

        mm_use_im_start_end = getattr(model.config, "mm_use_im_start_end", False)
        mm_use_im_patch_token = getattr(model.config, "mm_use_im_patch_token", True)
        if mm_use_im_patch_token:
            tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)
        if mm_use_im_start_end:
            tokenizer.add_tokens([DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True)
        model.resize_token_embeddings(len(tokenizer))

        vision_tower = model.get_vision_tower()
        if not vision_tower.is_loaded:
            vision_tower.load_model()
        image_processor = vision_tower.image_processor

        # Check if the video exists
        if image_path is not None:
            cur_prompt = qs
            if model.config.mm_use_im_start_end:
                qs = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_TOKEN + DEFAULT_IM_END_TOKEN + '\n' + qs
            else:
                qs = DEFAULT_IMAGE_TOKEN + '\n' + qs

            conv = conv_templates[conv_mode].copy()
            conv.append_message(conv.roles[0], qs)
            conv.append_message(conv.roles[1], None)
            prompt = conv.get_prompt()

            input_ids = tokenizer_image_token(prompt, tokenizer, IMAGE_TOKEN_INDEX, return_tensors='pt').unsqueeze(0).cuda()

            image = Image.open(image_path)
            image_tensor = image_processor.preprocess(image, return_tensors='pt')['pixel_values'][0]

            stop_str = conv.sep if conv.sep_style != SeparatorStyle.TWO else conv.sep2
            keywords = [stop_str]
            stopping_criteria = KeywordsStoppingCriteria(keywords, tokenizer, input_ids)

            with torch.inference_mode():
                output_ids = model.generate(
                    input_ids,
                    images=image_tensor.unsqueeze(0).half().cuda(),
                    do_sample=True,
                    temperature=temperature,
                    top_p=top_p,
                    num_beams=num_beams,
                    max_new_tokens=1024,
                    use_cache=True,
                    stopping_criteria=[stopping_criteria])

            input_token_len = input_ids.shape[1]
            n_diff_input_output = (input_ids != output_ids[:, :input_token_len]).sum().item()
            if n_diff_input_output > 0:
                print(f'[Warning] {n_diff_input_output} output_ids are not the same as the input_ids')
            outputs = tokenizer.batch_decode(output_ids[:, input_token_len:], skip_special_tokens=True)[0]
            outputs = outputs.strip()
            if outputs.endswith(stop_str):
                outputs = outputs[:-len(stop_str)]
            outputs = outputs.strip()
            print(outputs)
        return json(outputs)
    
@app.post("/vid")
async def vid(request):
    # model_path = "C:/Users/Ervin/Documents/Chat-UniVi/model"  # or "Chat-UniVi/Chat-UniVi-13B"
    video_path = "Indo.mp4"

    # The number of visual tokens varies with the length of the video. "max_frames" is the maximum number of frames.
    # When the video is long, we will uniformly downsample the video to meet the frames when equal to the "max_frames".
    max_frames = 100

    # The number of frames retained per second in the video.
    video_framerate = 1

    # Input Text
    qs = "Describe the video."

    # Sampling Parameter
    conv_mode = "simple"
    temperature = 0.2
    top_p = None
    num_beams = 1

    disable_torch_init()
    # model_path = os.path.expanduser(model_path)
    # model_name = "ChatUniVi"
    # tokenizer, model, image_processor, context_len = load_pretrained_model(model_path, None, model_name)

    mm_use_im_start_end = getattr(model.config, "mm_use_im_start_end", False)
    mm_use_im_patch_token = getattr(model.config, "mm_use_im_patch_token", True)
    if mm_use_im_patch_token:
        tokenizer.add_tokens([DEFAULT_IMAGE_PATCH_TOKEN], special_tokens=True)
    if mm_use_im_start_end:
        tokenizer.add_tokens([DEFAULT_IM_START_TOKEN, DEFAULT_IM_END_TOKEN], special_tokens=True)
    model.resize_token_embeddings(len(tokenizer))

    vision_tower = model.get_vision_tower()
    if not vision_tower.is_loaded:
        vision_tower.load_model()
    image_processor = vision_tower.image_processor

    if model.config.config["use_cluster"]:
        for n, m in model.named_modules():
            m = m.to(dtype=torch.bfloat16)

    # Check if the video exists
    if video_path is not None:
        video_frames, slice_len = _get_rawvideo_dec(video_path, image_processor, max_frames=max_frames, video_framerate=video_framerate)

        cur_prompt = qs
        if model.config.mm_use_im_start_end:
            qs = DEFAULT_IM_START_TOKEN + DEFAULT_IMAGE_TOKEN * slice_len + DEFAULT_IM_END_TOKEN + '\n' + qs
        else:
            qs = DEFAULT_IMAGE_TOKEN * slice_len + '\n' + qs

        conv = conv_templates[conv_mode].copy()
        conv.append_message(conv.roles[0], qs)
        conv.append_message(conv.roles[1], None)
        prompt = conv.get_prompt()

        input_ids = tokenizer_image_token(prompt, tokenizer, IMAGE_TOKEN_INDEX, return_tensors='pt').unsqueeze(
            0).cuda()

        stop_str = conv.sep if conv.sep_style != SeparatorStyle.TWO else conv.sep2
        keywords = [stop_str]
        stopping_criteria = KeywordsStoppingCriteria(keywords, tokenizer, input_ids)

        with torch.inference_mode():
            output_ids = model.generate(
                input_ids,
                images=video_frames.half().cuda(),
                do_sample=True,
                temperature=temperature,
                top_p=top_p,
                num_beams=num_beams,
                output_scores=True,
                return_dict_in_generate=True,
                max_new_tokens=1024,
                use_cache=True,
                stopping_criteria=[stopping_criteria])

        output_ids = output_ids.sequences
        input_token_len = input_ids.shape[1]
        n_diff_input_output = (input_ids != output_ids[:, :input_token_len]).sum().item()
        if n_diff_input_output > 0:
            print(f'[Warning] {n_diff_input_output} output_ids are not the same as the input_ids')
        outputs = tokenizer.batch_decode(output_ids[:, input_token_len:], skip_special_tokens=True)[0]
        outputs = outputs.strip()
        if outputs.endswith(stop_str):
            outputs = outputs[:-len(stop_str)]
        outputs = outputs.strip()
        print(outputs)
        return json(outputs)
    

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=8888, debug=True, access_log=True)