<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
training_df_fullctxt_and_sent_split_filtered_0_15_PubMedBert
This model is a fine-tuned version of dmis-lab/TinyPubMedBERT-v1.0 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3031
- Rouge1: 0.8717
- Rouge2: 0.6989
- Rougel: 0.6336
- Rougelsum: 0.6336
- Exact Match: 0.0
- Precision: [0.8712936639785767, 0.9647811651229858]
- Recall: [0.8689576387405396, 0.9682695865631104]
- F1: [0.8701240420341492, 0.9665222764015198]
- Hashcode: roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0)
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Exact Match | Precision | Recall | F1 | Hashcode |
---|---|---|---|---|---|---|---|---|---|---|---|---|
0.4001 | 1.0 | 5881 | 0.3415 | 0.6842 | 0.6047 | 0.6120 | 0.6120 | 0.0 | [0.8383916616439819, 0.960318922996521] | [0.7912731170654297, 0.963049054145813] | [0.8141512274742126, 0.9616820812225342] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.3165 | 2.0 | 11762 | 0.3255 | 0.7947 | 0.6870 | 0.6369 | 0.6369 | 0.0 | [0.8562091588973999, 0.9591262340545654] | [0.841107964515686, 0.9619568586349487] | [0.8485913872718811, 0.9605394601821899] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.2971 | 3.0 | 17643 | 0.3178 | 0.8168 | 0.6965 | 0.6365 | 0.6365 | 0.0 | [0.8633116483688354, 0.978273868560791] | [0.8504236936569214, 0.9788444638252258] | [0.856819212436676, 0.9785590767860413] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.2853 | 4.0 | 23524 | 0.2934 | 0.8134 | 0.7020 | 0.6328 | 0.6328 | 0.0 | [0.8643838167190552, 0.9647811651229858] | [0.8536887764930725, 0.9682695865631104] | [0.859002947807312, 0.9665222764015198] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.2744 | 5.0 | 29405 | 0.2968 | 0.8664 | 0.7077 | 0.6357 | 0.6357 | 0.0 | [0.8695193529129028, 0.9638710021972656] | [0.8581283688545227, 0.9666727185249329] | [0.8637862205505371, 0.9652698636054993] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.2669 | 6.0 | 35286 | 0.3027 | 0.8472 | 0.6949 | 0.6378 | 0.6378 | 0.0 | [0.8685003519058228, 0.9665455222129822] | [0.8652210235595703, 0.9689881801605225] | [0.8668575882911682, 0.9677652716636658] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.2595 | 7.0 | 41167 | 0.2996 | 0.8840 | 0.7193 | 0.6447 | 0.6447 | 0.0 | [0.8698508143424988, 0.9638710021972656] | [0.8639194965362549, 0.9666727185249329] | [0.8668749332427979, 0.9652698636054993] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.253 | 8.0 | 47048 | 0.2972 | 0.8518 | 0.6891 | 0.6363 | 0.6363 | 0.0 | [0.8666473031044006, 0.9638710021972656] | [0.863062858581543, 0.9666727185249329] | [0.8648514151573181, 0.9652698636054993] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.2481 | 9.0 | 52929 | 0.2985 | 0.8533 | 0.6843 | 0.6309 | 0.6309 | 0.0 | [0.8691736459732056, 0.9647811651229858] | [0.8661415576934814, 0.9682695865631104] | [0.8676549196243286, 0.9665222764015198] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
0.243 | 10.0 | 58810 | 0.3031 | 0.8717 | 0.6989 | 0.6336 | 0.6336 | 0.0 | [0.8712936639785767, 0.9647811651229858] | [0.8689576387405396, 0.9682695865631104] | [0.8701240420341492, 0.9665222764015198] | roberta-large_L17_no-idf_version=0.3.12(hug_trans=4.28.0) |
Framework versions
- Transformers 4.28.0
- Pytorch 1.12.1+cu113
- Datasets 2.7.1
- Tokenizers 0.13.2