pszemraj/long-t5-tglobal-base-16384-booksum-V12

this checkpoint has some further training and is exists separately to confirm metrics before merging to main

training metadata data in this json
the main model can be found here

Downloads last month: 9

Safetensors

Model size

248M params

Tensor type

F32

Summarization

This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train pszemraj/long-t5-tglobal-base-16384-booksum-V12

Evaluation results

ROUGE-1 on samsum
test set verified

30.003
ROUGE-2 on samsum
test set verified

7.267
ROUGE-L on samsum
test set verified

21.878
ROUGE-LSUM on samsum
test set verified

26.437
loss on samsum
test set verified

2.638
gen_len on samsum
test set verified

54.236
ROUGE-1 on launch/gov_report
test set verified

37.054
ROUGE-2 on launch/gov_report
test set verified

8.151
ROUGE-L on launch/gov_report
test set verified

17.665
ROUGE-LSUM on launch/gov_report
test set verified

33.428

View on Papers With Code