Description
This Text Generation model has been fine-tuned on FLANT5 Using SEC filings data. FLAN-T5 is a state-of-the-art language model developed by Google AI that utilizes the T5 architecture for text generation tasks.
Predicted Entities
How to use
document_assembler = nlp.DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("question")
flant5 = finance.TextGenerator.pretrained('fingen_flant5_finetuned_sec10k','en','finance/models')\
.setInputCols(["question"])\
.setOutputCol("generated_text")
.setMaxNewTokens(150)\
.setStopAtEos(True)
pipeline = nlp.Pipeline(stages=[document_assembler, flant5])
data = spark.createDataFrame([
[1, """Deferred revenue primarily consists of customer billings or payments received in advance of revenues being recognized from the company’s subscription and services contracts"""]
]).toDF('id', 'text')
results = pipeline.fit(data).transform(data)
results.select("generated_text.result").show(truncate=False)
Results
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|result |
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|[The company’s deferred revenue is recognized ratably over the term of the contract, which is generally one year or less, based on the estimated useful lives of the customer and the expected life of the customer’s subscription or services contract, and the estimated useful lives of the customer’s subscription or services contract, if any, if the company determines that the estimated useful lives of the customer’s subscription or services contract are less than the estimated useful lives of the customer’s subscription or services contract, the company recognizes revenue ratably over the term of the contract, which is generally one year or less, based on the estimated useful lives of the customer’s subscription or services contract]|
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
Model Information
| Model Name: | fingen_flant5_finetuned_sec10k |
| Compatibility: | Finance NLP 1.0.0+ |
| License: | Licensed |
| Edition: | Official |
| Language: | en |
| Size: | 1.6 GB |
References
In house annotated dataset