Edit model card

codet5p-770m-py-sanitized-codebleu-1-True-1e-06-0.1-prefix-tuning

This model is a fine-tuned version of Salesforce/codet5p-770m-py on the mbpp dataset. It achieves the following results on the evaluation set:

  • Loss: 9.3690
  • Codebleu: 0.0440
  • Ngram Match Score: 0.0003
  • Weighted Ngram Match Score: 0.0003
  • Syntax Match Score: 0.0013
  • Dataflow Match Score: 0.1084

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: 1e-06
  • 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
  • lr_scheduler_warmup_steps: 100
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Codebleu Ngram Match Score Weighted Ngram Match Score Syntax Match Score Dataflow Match Score
9.311 1.0 8 9.4023 0.0488 0.0003 0.0003 0.0013 0.1205
9.1929 2.0 16 9.4021 0.0488 0.0003 0.0003 0.0013 0.1205
9.2718 3.0 24 9.4019 0.0480 0.0003 0.0003 0.0013 0.1185
9.1752 4.0 32 9.4015 0.0480 0.0003 0.0003 0.0013 0.1185
9.1592 5.0 40 9.4010 0.0480 0.0003 0.0003 0.0013 0.1185
9.1841 6.0 48 9.4005 0.0480 0.0003 0.0003 0.0013 0.1185
9.2778 7.0 56 9.3998 0.0480 0.0003 0.0003 0.0013 0.1185
9.2436 8.0 64 9.3991 0.0480 0.0004 0.0003 0.0013 0.1185
9.1964 9.0 72 9.3983 0.0480 0.0004 0.0003 0.0013 0.1185
9.1678 10.0 80 9.3974 0.0480 0.0004 0.0003 0.0013 0.1185
9.2327 11.0 88 9.3963 0.0480 0.0004 0.0003 0.0013 0.1185
9.2025 12.0 96 9.3950 0.0480 0.0004 0.0003 0.0013 0.1185
9.202 13.0 104 9.3937 0.0480 0.0004 0.0003 0.0013 0.1185
9.2192 14.0 112 9.3924 0.0480 0.0003 0.0003 0.0013 0.1185
9.1797 15.0 120 9.3911 0.0480 0.0003 0.0003 0.0013 0.1185
9.1952 16.0 128 9.3901 0.0480 0.0003 0.0003 0.0013 0.1185
9.1932 17.0 136 9.3891 0.0480 0.0003 0.0003 0.0013 0.1185
9.2271 18.0 144 9.3880 0.0464 0.0003 0.0003 0.0013 0.1145
9.2241 19.0 152 9.3869 0.0464 0.0003 0.0003 0.0013 0.1145
9.2655 20.0 160 9.3857 0.0464 0.0003 0.0003 0.0013 0.1145
9.1865 21.0 168 9.3846 0.0464 0.0003 0.0003 0.0013 0.1145
9.1816 22.0 176 9.3834 0.0464 0.0003 0.0003 0.0013 0.1145
9.1961 23.0 184 9.3824 0.0464 0.0003 0.0003 0.0013 0.1145
9.1911 24.0 192 9.3814 0.0464 0.0003 0.0003 0.0013 0.1145
9.1785 25.0 200 9.3805 0.0464 0.0003 0.0003 0.0013 0.1145
9.1715 26.0 208 9.3797 0.0464 0.0003 0.0003 0.0013 0.1145
9.2293 27.0 216 9.3788 0.0464 0.0003 0.0003 0.0013 0.1145
9.2066 28.0 224 9.3780 0.0464 0.0003 0.0003 0.0013 0.1145
9.2515 29.0 232 9.3772 0.0464 0.0003 0.0003 0.0013 0.1145
9.198 30.0 240 9.3764 0.0464 0.0003 0.0003 0.0013 0.1145
9.2351 31.0 248 9.3758 0.0464 0.0003 0.0003 0.0013 0.1145
9.228 32.0 256 9.3751 0.0440 0.0003 0.0003 0.0013 0.1084
9.2185 33.0 264 9.3744 0.0440 0.0003 0.0003 0.0013 0.1084
9.1991 34.0 272 9.3738 0.0440 0.0003 0.0003 0.0013 0.1084
9.2657 35.0 280 9.3732 0.0440 0.0003 0.0003 0.0013 0.1084
9.1871 36.0 288 9.3726 0.0440 0.0003 0.0003 0.0013 0.1084
9.1871 37.0 296 9.3721 0.0440 0.0003 0.0003 0.0013 0.1084
9.2122 38.0 304 9.3716 0.0440 0.0003 0.0003 0.0013 0.1084
9.2057 39.0 312 9.3712 0.0440 0.0003 0.0003 0.0013 0.1084
9.2578 40.0 320 9.3708 0.0440 0.0003 0.0003 0.0013 0.1084
9.1719 41.0 328 9.3704 0.0440 0.0003 0.0003 0.0013 0.1084
9.1942 42.0 336 9.3701 0.0440 0.0003 0.0003 0.0013 0.1084
9.2253 43.0 344 9.3698 0.0440 0.0003 0.0003 0.0013 0.1084
9.1445 44.0 352 9.3696 0.0440 0.0003 0.0003 0.0013 0.1084
9.2078 45.0 360 9.3694 0.0440 0.0003 0.0003 0.0013 0.1084
9.23 46.0 368 9.3692 0.0440 0.0003 0.0003 0.0013 0.1084
9.1418 47.0 376 9.3691 0.0440 0.0003 0.0003 0.0013 0.1084
9.2528 48.0 384 9.3690 0.0440 0.0003 0.0003 0.0013 0.1084
9.1605 49.0 392 9.3690 0.0440 0.0003 0.0003 0.0013 0.1084
9.1709 50.0 400 9.3690 0.0440 0.0003 0.0003 0.0013 0.1084

Framework versions

  • Transformers 4.31.0
  • Pytorch 2.0.1
  • Datasets 2.14.4
  • Tokenizers 0.13.3
Downloads last month
0
Unable to determine this model's library. Check the docs .

Finetuned from

Dataset used to train vichyt/codet5p-770m-py-sanitized-codebleu-1-True-1e-06-0.1-prefix-tuning