Blaise-g
/

led_pubmed_sumpubmed_1

text2text-generation

Eval Results (legacy)

Model card Files Files and versions

Validation Metrics

Loss: 2.133
Rouge1: 45.861
Rouge2: 14.179
RougeL: 23.565
RougeLsum: 40.908
Gen Len: 195.334

Downloads last month: 7

Safetensors

Model size

0.5B params

Tensor type

F32

·

Dataset used to train Blaise-g/led_pubmed_sumpubmed_1

Evaluation results

ROUGE-1 on Blaise-g/SumPubmed
test set self-reported

41.252
ROUGE-2 on Blaise-g/SumPubmed
test set self-reported

11.129
ROUGE-L on Blaise-g/SumPubmed
test set self-reported

20.253
ROUGE-LSUM on Blaise-g/SumPubmed
test set self-reported

37.150
loss on Blaise-g/SumPubmed
test set self-reported

6.371
gen_len on Blaise-g/SumPubmed
test set self-reported

193.374