metadata
license: apache-2.0
datasets:
- Open-Orca/FLAN
language:
- en
pipeline_tag: text2text-generation
T5-small distilled from T5-flan-base using OpenOrca'a FLAN dataset using Seq2seq methodology. The baseline for Seq2seq distillation on T5 decoder models. It is free under Apache License.
The distilled model is a T5-small model with 60M parameters, not a T5-flan-small model.