Korean T5ForConditionalGeneration Base Cased model (from KETI-AIR)

Description

Pretrained T5ForConditionalGeneration model, adapted from Hugging Face and curated to provide scalability and production-readiness using Spark NLP. ke-t5-base-ko is a Korean model originally trained by KETI-AIR.

Live Demo Open in Colab Download Copy S3 URI

How to use

documentAssembler = DocumentAssembler() \
    .setInputCols("text") \
    .setOutputCols("document")

t5 = T5Transformer.pretrained("t5_ke_base","ko") \
    .setInputCols("document") \
    .setOutputCol("answers")
    
pipeline = Pipeline(stages=[documentAssembler, t5])

data = spark.createDataFrame([["PUT YOUR STRING HERE"]]).toDF("text")

result = pipeline.fit(data).transform(data)

val documentAssembler = new DocumentAssembler() 
      .setInputCols("text")
      .setOutputCols("document")
       
val t5 = T5Transformer.pretrained("t5_ke_base","ko") 
    .setInputCols("document")
    .setOutputCol("answers")
   
val pipeline = new Pipeline().setStages(Array(documentAssembler, t5))

val data = Seq("PUT YOUR STRING HERE").toDS.toDF("text")

val result = pipeline.fit(data).transform(data)

Model Information

References

• https://huggingface.co/KETI-AIR/ke-t5-base-ko
• https://github.com/google-research/text-to-text-transfer-transformer#released-model-checkpoints
• https://github.com/AIRC-KETI/ke-t5
• https://aclanthology.org/2021.findings-emnlp.33/
• https://jmlr.org/papers/volume21/20-074/20-074.pdf
• https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html
• https://aclanthology.org/2021.acl-long.330.pdf
• https://dl.acm.org/doi/pdf/10.1145/3442188.3445922
• https://www.tensorflow.org/datasets/catalog/c4
• https://jmlr.org/papers/volume21/20-074/20-074.pdf
• https://mlco2.github.io/impact#compute
• https://arxiv.org/abs/1910.09700