dependencies requirements.txt step

[patrickocal]

Successfully built overrides jsonnet word2number future sacremoses
Installing collected packages: python-dateutil, murmurhash, jmespath, cymem, wasabi, srsly, preshed, plac, joblib, click, catalogue, botocore, blis, zipp, tomli, tokenizers, threadpoolctl, thinc, sentencepiece, scipy, sacremoses, s3transfer, protobuf, pluggy, iniconfig, greenlet, future, filelock, exceptiongroup, wcwidth, transformers, torch, tensorboardX, sqlalchemy, spacy, scikit-learn, rapidfuzz, pytest, overrides, nltk, Mako, jsonpickle, jsonnet, importlib-resources, importlib-metadata, h5py, boto3, word2number, soupsieve, pytz, py-rouge, Levenshtein, ftfy, conllu, colorlog, allennlp, alembic, python-Levenshtein, pandas, optuna, lxml, beautifulsoup4, allennlp-models
  Attempting uninstall: tokenizers
    Found existing installation: tokenizers 0.14.1
    Uninstalling tokenizers-0.14.1:
      Successfully uninstalled tokenizers-0.14.1
  Attempting uninstall: filelock
    Found existing installation: filelock 3.12.4
    Uninstalling filelock-3.12.4:
      Successfully uninstalled filelock-3.12.4
  Attempting uninstall: transformers
    Found existing installation: transformers 4.34.0
    Uninstalling transformers-4.34.0:
      Successfully uninstalled transformers-4.34.0
  Attempting uninstall: torch
    Found existing installation: torch 2.1.0
    Uninstalling torch-2.1.0:
      Successfully uninstalled torch-2.1.0
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.
accelerate 0.23.0 requires torch>=1.10.0, but you have torch 1.6.0 which is incompatible.
Successfully installed Levenshtein-0.23.0 Mako-1.2.4 alembic-1.12.0 allennlp-1.1.0 allennlp-models-1.1.0 beautifulsoup4-4.8.1 blis-0.7.11 boto3-1.28.67 botocore-1.31.67 catalogue-1.0.2 click-8.1.7 colorlog-6.7.0 conllu-4.1 cymem-2.0.8 exceptiongroup-1.1.3 filelock-3.0.12 ftfy-6.1.1 future-0.18.3 greenlet-3.0.0 h5py-3.10.0 importlib-metadata-6.8.0 importlib-resources-6.1.0 iniconfig-2.0.0 jmespath-1.0.1 joblib-1.3.2 jsonnet-0.20.0 jsonpickle-3.0.2 lxml-4.9.3 murmurhash-1.0.10 nltk-3.8.1 optuna-3.4.0 overrides-3.1.0 pandas-0.25.2 plac-1.1.3 pluggy-1.3.0 preshed-3.0.9 protobuf-4.24.4 py-rouge-1.1 pytest-7.4.2 python-Levenshtein-0.23.0 python-dateutil-2.8.2 pytz-2023.3.post1 rapidfuzz-3.4.0 s3transfer-0.7.0 sacremoses-0.0.53 scikit-learn-1.3.1 scipy-1.10.1 sentencepiece-0.1.99 soupsieve-2.5 spacy-2.3.9 sqlalchemy-2.0.22 srsly-1.0.7 tensorboardX-2.6.2.2 thinc-7.4.6 threadpoolctl-3.2.0 tokenizers-0.8.1rc1 tomli-2.0.1 torch-1.6.0 transformers-3.0.2 wasabi-0.10.1 wcwidth-0.2.8 word2number-1.1 zipp-3.17.0

I have now added the following line to requirements.txt: torch==1.10.0

[patrickocal]

but then I get

Requirement already satisfied: scipy in /home/uqpocall/.local/lib/python3.8/site-packages (from allennlp==1.1.0->-r requirements.txt (line 5)) (1.10.1)
ERROR: Cannot install allennlp==1.1.0 and torch==1.10.0 because these package versions have conflicting dependencies.

The conflict is caused by:
    The user requested torch==1.10.0
    allennlp 1.1.0 depends on torch<1.7.0 and >=1.6.0

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/user_guide/#fixing-conflicting-dependencies

[dwadden]

Hi Patrick,

Sorry about this issue. I looked into this, and at this point you're right we're running into a lot of dependency issues -- it looks like python 3.7 isn't available on conda anymore, and pytorch 1.6 isn't available from pip.

Unfortunately, I don't have the bandwidth to resolve this at this point. If you (or someone else) is able to figure out the dependencies to get things working again, I'd very much appreciate a PR. I'll add a note on the README mentioning this as an issue.

Dave

[ItamarLB]

Hi Dave, do you happen to have a recommendation for a follow up work that also incorporate graphs in a similar manner?

Itamar

[dwadden]

Hi Itamar,

Sorry, I don't have any particular references on the latest work on this. You might be able to just check for recent papers that cite this one? Sorry I can't help more!

Best,

Dave