tasknet is an interface between Huggingface datasets and Huggingface Trainer.

tasknet relies on task templates to avoid boilerplate codes. The task templates correspond to Transformers AutoClasses:

• SequenceClassification
• TokenClassification
• MultipleChoice
• Seq2SeqLM (experimental support)

The task templates follow the same interface. They implement preprocess_function, a data collator and compute_metrics. Look at tasks.py and use existing templates as a starting point to implement a custom task template.

pip install tasknet

Each task template has fields that should be matched with specific dataset columns. Classification has two text fields s1,s2, and a label y. Pass a dataset to a template, and fill in the mapping between the template fields and the dataset columns to instantiate a task.

import tasknet as tn; from datasets import load_dataset

rte = tn.Classification(
    dataset=load_dataset("glue", "rte"),
    s1="sentence1", s2="sentence2", y="label") #s2 is optional # See AutoTask for shorter code

class hparams:
  model_name='microsoft/deberta-v3-base' # deberta models have the best results (and tasknet support)
  learning_rate = 3e-5 # see hf.co/docs/transformers/en/main_classes/trainer#transformers.TrainingArguments
 
tasks = [rte]
model = tn.Model(tasks, hparams)
trainer = tn.Trainer(model, tasks, hparams)
trainer.train()
trainer.evaluate()
p = trainer.pipeline()
p([{'text':x.premise,'text_pair': x.hypothesis}]) # HuggingFace pipeline for inference

Tasknet is multitask by design. model.task_models_list contains one model per task, with a shared encoder.

tn.Classification(dataset, nrows=5000, nrows_eval=500, oversampling=2)

You can balance multiple datasets with nrows and oversampling. nrows is the maximal number of examples. If a dataset has less than nrows, it will be oversampled at most oversampling times.

You can also leverage tasksource with tn.AutoTask and have one-line access to 600+ datasets, see implemented tasks.

rte = tn.AutoTask("glue/rte", nrows=5000)

AutoTask guesses a template based on the dataset structure. It also accepts a dataset as input, if it fits the template (e.g. after tasksource custom preprocessing).

Minimal-ish example:

https://colab.research.google.com/drive/15Xf4Bgs3itUmok7XlAK6EEquNbvjD9BD?usp=sharing

More complex example, where tasknet was scaled to 600 tasks:

https://colab.research.google.com/drive/1iB4Oxl9_B5W3ZDzXoWJN-olUbqLBxgQS?usp=sharing

jiant is another library comparable to tasknet. tasknet is a minimal extension of Trainer centered on task templates, while jiant builds a Trainer equivalent from scratch called runner. tasknet is leaner and closer to Huggingface native tools. Jiant is config-based and command line focused while tasknet is designed for interactive use and python scripting.

This code uses some part of the examples of the transformers library and some code from multitask-learning-transformers.

You can request features on github or reach me at [email protected]

@misc{sileod22-tasknet,
  author = {Sileo, Damien},
  doi = {10.5281/zenodo.561225781},
  month = {11},
  title = {{tasknet, multitask interface between Trainer and datasets}},
  url = {https://github.com/sileod/tasknet},
  version = {1.5.0},
  year = {2022}}