README - long-t5-tglobal-base-16384-booksum-V11-big_patent-V2
- this README was added because there wasn't one
- created 2022-07-31_12-14-50
about
An experiment testing some transfer learning with pszemraj/long-t5-tglobal-base-16384-book-summary to evaluate the ability to learn some technical documentation through the big_patent dataset on modeldatabase.
This checkpoint has been trained on dataset subsection y of big_patent for approx 400 steps of functional batch size 128.
- Downloads last month
- 103
Datasets used to train pszemraj/long-t5-tglobal-base-16384-booksum-V11-big_patent-V2
Space using pszemraj/long-t5-tglobal-base-16384-booksum-V11-big_patent-V2 1
Evaluation results
- ROUGE-1 on kmfoda/booksumtest set verified23.144
- ROUGE-2 on kmfoda/booksumtest set verified3.239
- ROUGE-L on kmfoda/booksumtest set verified12.704
- ROUGE-LSUM on kmfoda/booksumtest set verified19.810
- loss on kmfoda/booksumtest set verified2.766
- gen_len on kmfoda/booksumtest set verified63.449
- ROUGE-1 on samsumtest set verified26.803
- ROUGE-2 on samsumtest set verified6.066
- ROUGE-L on samsumtest set verified20.010
- ROUGE-LSUM on samsumtest set verified21.912
- loss on samsumtest set verified2.317
- gen_len on samsumtest set verified19.111
- ROUGE-1 on xsumtest set verified25.206
- ROUGE-2 on xsumtest set verified4.705
- ROUGE-L on xsumtest set verified17.859
- ROUGE-LSUM on xsumtest set verified18.080
- loss on xsumtest set verified3.003
- gen_len on xsumtest set verified27.482
- ROUGE-1 on cnn_dailymailtest set verified27.569
- ROUGE-2 on cnn_dailymailtest set verified6.126
- ROUGE-L on cnn_dailymailtest set verified17.113
- ROUGE-LSUM on cnn_dailymailtest set verified23.007
- loss on cnn_dailymailtest set verified2.219
- gen_len on cnn_dailymailtest set verified39.195
- ROUGE-1 on billsumtest set verified28.063
- ROUGE-2 on billsumtest set verified9.900
- ROUGE-L on billsumtest set verified18.250
- ROUGE-LSUM on billsumtest set verified21.905
- loss on billsumtest set verified2.033
- gen_len on billsumtest set verified48.599
- ROUGE-1 on big_patenttest set verified34.785
- ROUGE-2 on big_patenttest set verified9.755
- ROUGE-L on big_patenttest set verified22.228
- ROUGE-LSUM on big_patenttest set verified28.039
- loss on big_patenttest set verified1.779
- gen_len on big_patenttest set verified71.637
- ROUGE-1 on launch/gov_reportvalidation set verified23.593
- ROUGE-2 on launch/gov_reportvalidation set verified5.676
- ROUGE-L on launch/gov_reportvalidation set verified13.811
- ROUGE-LSUM on launch/gov_reportvalidation set verified20.244
- loss on launch/gov_reportvalidation set verified2.638
- gen_len on launch/gov_reportvalidation set verified64.181
- ROUGE-1 on launch/gov_reporttest set verified23.744
- ROUGE-2 on launch/gov_reporttest set verified5.501
- ROUGE-L on launch/gov_reporttest set verified13.813
- ROUGE-LSUM on launch/gov_reporttest set verified20.462
- loss on launch/gov_reporttest set verified2.638
- gen_len on launch/gov_reporttest set verified64.909