google/seqio

Issues

• Support for "Deterministic Pipelines"

  #243 opened 2 years ago by ruomingp
  5

• What is a number?

  #734 opened 9 months ago by justzh
  0

• unable to train mt5 from t5x using mixtures ValueError: Dataset is missing an expected feature during input_validation validation: 'inputs'

  #310 opened 2 years ago by StephennFernandes
  3

• ValueError: mutable default <class 'seqio.vocabularies.PassThroughVocabulary'> for field vocabulary is not allowed: use default_factory

  #562 opened 2 years ago by jli262
  3

• Unimax sampler implementation?

  #565 opened 2 years ago by StephennFernandes
  0

• unimax sampling ?

  #556 opened 2 years ago by StephennFernandes
  0

• seqio.get_mixture_or_task('bool_q_template_0_no_opt_five_shot') failed

  #553 opened 2 years ago by liuzhiyong01
  0

• Dataset performance

  #547 opened 2 years ago by KeremTurgutlu
  0

• TfdsDataProvider gives error with non-None tfds_data_dir

  #545 opened 2 years ago by lucaslingle
  2

• How to choose minimum sequence length while avoiding truncation

  #535 opened 2 years ago by marcospiau
  0

• Concatenating Tasks?

  #513 opened 2 years ago by gahdritz
  2

• caching tasks goes out of memory due to apache beam

  #514 opened 2 years ago by mayurnewase
  2

• Please include installation instructions

  #261 opened 3 years ago by leiterenato
  2

• import seqio: AttributeError: module 'typing' has no attribute 'get_origin'

  #502 opened 2 years ago by PhoebusSi
  1

• How to apply the huggingface tokenizer in seqio.vocabulary

  #406 opened 2 years ago by nawnoes
  0

• seqio 0.0.13 cannot be installed on Apple Silicon due to transitive tensorflow dependency of clu

  #396 opened 2 years ago by tuzhucheng
  2

• Different preprocessors for each dataset split

  #407 opened 2 years ago by marcospiau
  2

• how to decide ideal mixture rates ?

  #299 opened 2 years ago by StephennFernandes
  1

• FunctionDataSource does not allow function with 3 positional arguments thus shuffling does not work

  #307 opened 2 years ago by marton-avrios
  2

• Tokenizer is not behaving as expected on special tokens (doesn't recognize `pad` and `eos` tokens)

  #311 opened 2 years ago by armancohan
  1

• Using a registered task to add another

  #333 opened 2 years ago by AkshitaB
  1

• import seqio

  #291 opened 2 years ago by Arij-Aladel
  0

• How to just use the mixture functionality in seqio

  #244 opened 3 years ago by StephennFernandes
  9

• HuggingFace Tokenizers compatibility

  #188 opened 3 years ago by gabeorlanski
  3

• Dataset seeking for restarting from a T5X crashed run using HuggingFace datasets

  #224 opened 3 years ago by versae
  8

• Add Evaluator example

  #223 opened 3 years ago by abisee
  0

• Add method to directly add tasks/mixtures.

  #215 opened 3 years ago by salayatana66
  1

• Using seqio for T5X Dataset Generation

  #182 opened 3 years ago by stefan-it
  7

• seqio_cache_tasks fails on DataflowRunner

  #109 opened 3 years ago by bzz
  3

• Possible ByteVocabulary Bug

  #158 opened 3 years ago by xwd
  2

• `tokenize_and_append_eos` needs another requrired input (`output_features`)

  #6 opened 4 years ago by puraminy
  1