Can someone provide your succeeded conda environment (The whole package list) for Demo in InternVideo2?

[Shuaicong97]

I built a conda environment using requirments.txt, I ran the demo_video_text_retrieval.ipynb in InternVideo2/multi_modality.

import numpy as np
import os
import io
import cv2

import torch

from demo_config import (Config,
                    eval_dict_leaf)

from demo.utils import (retrieve_text,
                  _frame_from_video,
                  setup_internvideo2)

In the first part, the following error occurred:

ModuleNotFoundError                       Traceback (most recent call last)
File ~/forkProjects/InternVideo/InternVideo2/multi_modality/models/backbones/internvideo2/internvl_clip_vision.py:12
     11 try:
---> 12     from .flash_attention_class import FlashAttention
     13 except:

File ~/forkProjects/InternVideo/InternVideo2/multi_modality/models/backbones/internvideo2/flash_attention_class.py:6
      4 from einops import rearrange
----> 6 from flash_attn.flash_attn_interface import flash_attn_varlen_qkvpacked_func
      7 from flash_attn.bert_padding import unpad_input, pad_input

ModuleNotFoundError: No module named 'flash_attn'

During handling of the above exception, another exception occurred:

ModuleNotFoundError                       Traceback (most recent call last)
Cell In[4], line 11
      6 import torch
      8 from demo_config import (Config,
      9                     eval_dict_leaf)
---> 11 from demo.utils import (retrieve_text,
     12                   _frame_from_video,
     13                   setup_internvideo2)

File ~/forkProjects/InternVideo/InternVideo2/multi_modality/demo/utils.py:9
      6 import torch
      7 from torch import nn
----> 9 from models.backbones.internvideo2 import pretrain_internvideo2_1b_patch14_224
     10 from models.backbones.bert.builder import build_bert
     11 from models.criterions import get_sim

File ~/forkProjects/InternVideo/InternVideo2/multi_modality/models/backbones/internvideo2/__init__.py:1
----> 1 from .internvl_clip_vision import internvl_clip_6b
      2 from .internvideo2 import pretrain_internvideo2_1b_patch14_224, pretrain_internvideo2_6b_patch14_224
      3 from .internvideo2_clip_vision import InternVideo2

File ~/forkProjects/InternVideo/InternVideo2/multi_modality/models/backbones/internvideo2/internvl_clip_vision.py:14
     12     from .flash_attention_class import FlashAttention
     13 except:
---> 14     from flash_attention_class import FlashAttention
     15 from flash_attn.modules.mlp import FusedMLP
     16 from flash_attn.ops.rms_norm import DropoutAddRMSNorm

ModuleNotFoundError: No module named 'flash_attention_class'

However, in File ~/forkProjects/InternVideo/InternVideo2/multi_modality/models/backbones/internvideo2/internvl_clip_vision.py, using command+click .flash_attention_class, it can navigate to the right place. So I'm not sure how to solve it. Does someone have the same issue?

[Shuaicong97]

Could you give a new WeChat group QR code? The current one is full of members. Thank you in advance! @shepnerd @yinanhe