installation issues in colab example notebook

[bocanegrah]

❓ Questions and Help

I am trying to the example notebook https://colab.research.google.com/github/facebookresearch/mmf/blob/notebooks/notebooks/mmf_hm_example.ipynb but keep having issues with the installation. When I run

!pip install mmf@https://github.com/facebookresearch/mmf/tarball/master

many packages do install, but there seem to be some conflicts with dependencies

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
yfinance 0.2.14 requires requests>=2.26, but you have requests 2.23.0 which is incompatible.
tweepy 4.13.0 requires requests<3,>=2.27.0, but you have requests 2.23.0 which is incompatible.
torchdata 0.6.0 requires torch==2.0.0, but you have torch 1.11.0 which is incompatible.
tensorflow 2.12.0 requires numpy<1.24,>=1.22, but you have numpy 1.21.4 which is incompatible.
plotnine 0.10.1 requires matplotlib>=3.5.0, but you have matplotlib 3.3.4 which is incompatible.
pandas-profiling 3.2.0 requires requests>=2.24.0, but you have requests 2.23.0 which is incompatible.
mizani 0.8.1 requires matplotlib>=3.5.0, but you have matplotlib 3.3.4 which is incompatible.
google-colab 1.0.0 requires requests>=2.27.0, but you have requests 2.23.0 which is incompatible.

I try to continue running the subsequent cells, and when I reach this cell

from mmf.common.registry import registry
from mmf.models.mmbt import MMBT
from mmf.utils.build import build_dataset
from mmf.utils.env import setup_imports

setup_imports()
dataset = build_dataset("hateful_memes")

import matplotlib.pyplot as plt
plt.rcParams["figure.figsize"] = (20, 20)
dataset.visualize(num_samples=8, size=(512, 512), nrow=4)

I get the following error

---------------------------------------------------------------------------

ImportError                               Traceback (most recent call last)

[<ipython-input-3-15337222f728>](https://localhost:8080/#) in <cell line: 1>()
----> 1 from mmf.common.registry import registry
      2 from mmf.models.mmbt import MMBT
      3 from mmf.utils.build import build_dataset
      4 from mmf.utils.env import setup_imports
      5 

17 frames

[/usr/local/lib/python3.9/dist-packages/mmf/__init__.py](https://localhost:8080/#) in <module>
      6 patch_transformers()
      7 
----> 8 from mmf import common, datasets, models, modules, utils
      9 from mmf.modules import losses, metrics, optimizers, poolers, schedulers
     10 from mmf.version import __version__

[/usr/local/lib/python3.9/dist-packages/mmf/datasets/__init__.py](https://localhost:8080/#) in <module>
      1 # Copyright (c) Facebook, Inc. and its affiliates.
----> 2 from . import processors
      3 from .base_dataset import BaseDataset
      4 from .base_dataset_builder import BaseDatasetBuilder
      5 from .concat_dataset import ConcatDataset

[/usr/local/lib/python3.9/dist-packages/mmf/datasets/processors/__init__.py](https://localhost:8080/#) in <module>
      1 # Copyright (c) Facebook, Inc. and its affiliates.
      2 
----> 3 from mmf.datasets.processors.bert_processors import MaskedTokenProcessor
      4 from mmf.datasets.processors.frcnn_processor import FRCNNPreprocess
      5 from mmf.datasets.processors.image_processors import TorchvisionTransforms

[/usr/local/lib/python3.9/dist-packages/mmf/datasets/processors/bert_processors.py](https://localhost:8080/#) in <module>
      8 from mmf.common.registry import registry
      9 from mmf.common.sample import Sample, SampleList
---> 10 from mmf.datasets.processors.processors import BaseProcessor
     11 
     12 try:

[/usr/local/lib/python3.9/dist-packages/mmf/datasets/processors/processors.py](https://localhost:8080/#) in <module>
     88 from mmf.utils.logger import log_class_usage
     89 from mmf.utils.text import VocabDict
---> 90 from mmf.utils.vocab import Vocab, WordToVectorDict
     91 from omegaconf import DictConfig
     92 

[/usr/local/lib/python3.9/dist-packages/mmf/utils/vocab.py](https://localhost:8080/#) in <module>
     10 from mmf.utils.file_io import PathManager
     11 from mmf.utils.general import get_absolute_path
---> 12 from torchtext import vocab
     13 
     14 

[/usr/local/lib/python3.9/dist-packages/torchtext/__init__.py](https://localhost:8080/#) in <module>
      5 from . import data
      6 from . import nn
----> 7 from . import datasets
      8 from . import utils
      9 from . import vocab

[/usr/local/lib/python3.9/dist-packages/torchtext/datasets/__init__.py](https://localhost:8080/#) in <module>
      1 import importlib
      2 
----> 3 from .ag_news import AG_NEWS
      4 from .amazonreviewfull import AmazonReviewFull
      5 from .amazonreviewpolarity import AmazonReviewPolarity

[/usr/local/lib/python3.9/dist-packages/torchtext/datasets/ag_news.py](https://localhost:8080/#) in <module>
      3 
      4 from torchtext._internal.module_utils import is_module_available
----> 5 from torchtext.data.datasets_utils import (
      6     _wrap_split_argument,
      7     _create_dataset_directory,

[/usr/local/lib/python3.9/dist-packages/torchtext/data/datasets_utils.py](https://localhost:8080/#) in <module>
      4 import io
      5 import torch
----> 6 from torchtext.utils import (
      7     validate_file,
      8     download_from_url,

[/usr/local/lib/python3.9/dist-packages/torchtext/utils.py](https://localhost:8080/#) in <module>
      8 import zipfile
      9 import gzip
---> 10 from ._download_hooks import _DATASET_DOWNLOAD_MANAGER
     11 from torchtext import _CACHE_DIR
     12 

[/usr/local/lib/python3.9/dist-packages/torchtext/_download_hooks.py](https://localhost:8080/#) in <module>
      7 
      8 if is_module_available("torchdata"):
----> 9     from torchdata.datapipes.iter import HttpReader # noqa F401
     10 
     11 

[/usr/local/lib/python3.9/dist-packages/torchdata/__init__.py](https://localhost:8080/#) in <module>
      7 from torchdata import _extension  # noqa: F401
      8 
----> 9 from . import datapipes
     10 
     11 janitor = datapipes.utils.janitor

[/usr/local/lib/python3.9/dist-packages/torchdata/datapipes/__init__.py](https://localhost:8080/#) in <module>
      7 from torch.utils.data import DataChunk, functional_datapipe
      8 
----> 9 from . import iter, map, utils
     10 
     11 __all__ = ["DataChunk", "functional_datapipe", "iter", "map", "utils"]

[/usr/local/lib/python3.9/dist-packages/torchdata/datapipes/iter/__init__.py](https://localhost:8080/#) in <module>
     30     Zipper,
     31 )
---> 32 from torchdata.datapipes.iter.load.aisio import (
     33     AISFileListerIterDataPipe as AISFileLister,
     34     AISFileLoaderIterDataPipe as AISFileLoader,

[/usr/local/lib/python3.9/dist-packages/torchdata/datapipes/iter/load/aisio.py](https://localhost:8080/#) in <module>
      9 from torchdata.datapipes import functional_datapipe
     10 from torchdata.datapipes.iter import IterDataPipe
---> 11 from torchdata.datapipes.utils import StreamWrapper
     12 
     13 try:

[/usr/local/lib/python3.9/dist-packages/torchdata/datapipes/utils/__init__.py](https://localhost:8080/#) in <module>
      7 from torch.utils.data.datapipes.utils.common import StreamWrapper
      8 
----> 9 from torchdata.datapipes.utils._visualization import to_graph
     10 from torchdata.datapipes.utils.janitor import janitor
     11 from torchdata.datapipes.utils.pin_memory import pin_memory_fn

[/usr/local/lib/python3.9/dist-packages/torchdata/datapipes/utils/_visualization.py](https://localhost:8080/#) in <module>
     11 
     12 from torch.utils.data.datapipes.iter.combining import _ChildDataPipe, IterDataPipe
---> 13 from torch.utils.data.graph import traverse_dps
     14 
     15 if TYPE_CHECKING:

ImportError: cannot import name 'traverse_dps' from 'torch.utils.data.graph' (/usr/local/lib/python3.9/dist-packages/torch/utils/data/graph.py)

Have dependencies changed since the publication of that colab notebook? Am I using the correct git branch for installation? Thanks.

[ankitade]

can you try once outside of colab if the same thing work. The recommended way is to follow the installation instruction so colab support may not be great