project.yml

title: "Weights & Biases for xlm-roberta model for ner-uk"
description: "Use [Weights & Biases](https://www.wandb.com/) for logging of training experiments. This project template uses the NER_UK and includes two workflows: `log` for training a simple text classification model and logging the results to Weights & Biases (works out-of-the-box and only requires the `[training.logger]` to be set in the config) and `parameter-search` for running a hyperparameter search using [Weights & Biases Sweeps](https://docs.wandb.ai/guides/sweeps), running the experiments and logging the results."

# Variables can be referenced across the project.yml using ${vars.var_name}
vars:
  default_config: "ner_uk_xlm_roberta_wandb.config"

# These are the directories that the project needs. The project CLI will make
# sure that they always exist.
directories: ["scripts", "assets", "training", "configs"]

# Assets that should be downloaded or available in the directory. You can replace
# this with your own input data.
# assets:
#  - dest: "assets/aclImdb_v1.tar.gz"
#    url: "http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz"
#    description: "Movie Review Dataset for sentiment analysis by Maas et al., ACL 2011."

workflows:
  log:
    - data
    - train

  parameter-search:
    - data
    - train-search

# Project commands, specified in a style similar to CI config files (e.g. Azure
# pipelines). The name is the command name that lets you trigger the command
# via "spacy project run [command] [path]". The help message is optional and
# shown when executing "spacy project run [optional command] [path] --help".
commands:
  - name: "install"
    help: "Install dependencies and log in to Weights & Biases"
    script:
      - "pip install -r requirements.txt"
      - "pip install --upgrade spacy spacy-transformers"
      - "wandb login"
    deps:
      - "requirements.txt"

  - name: "data"
    help: "Extract the gold-standard annotations"
    script:
      - "mkdir -p assets/NER_UK/test/"
      - "mkdir -p assets/NER_UK/train/"
      - "python -m spacy convert /gdrive/MyDrive/NER_UK/fixed-split/test.iob assets/NER_UK/test/"
      - "python -m spacy convert /gdrive/MyDrive/NER_UK/fixed-split/train.iob assets/NER_UK/train/"
    deps:
      - "/gdrive/MyDrive/NER_UK/fixed-split/test.iob"
      - "/gdrive/MyDrive/NER_UK/fixed-split/train.iob"
    outputs:
      - "assets/NER_UK/train/"
      - "assets/NER_UK/test/"

  - name: "train"
    help: "Train a model using the default config"
    script:
      - "python -m spacy train configs/${vars.default_config} --output training/ --gpu-id 0"
    deps:
      - "assets/NER_UK/train/"
      - "assets/NER_UK/test/"

  - name: "train-search"
    help: "Run customized training runs for hyperparameter search using [Weights & Biases Sweeps](https://docs.wandb.ai/guides/sweeps)"
    script:
      - "python ./scripts/sweeps_using_config.py configs/${vars.default_config} scripts/sweep.yml training/"
    deps:
      - "assets/NER_UK/train/"
      - "assets/NER_UK/test/"
      - "configs/"

  - name: "clean"
    help: "Remove intermediate files."
    script:
      - "rm -rf training/*"
      - "rm -rf wandb/*"