loubnabnl's picture
loubnabnl HF Staff
Model save
c478f81 verified
|
raw
history blame
16.2 kB
metadata
base_model: bigcode/starencoder
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - accuracy
model-index:
  - name: classifier-llama3-shell-500k
    results: []

classifier-llama3-shell-500k

This model is a fine-tuned version of bigcode/starencoder on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3628
  • Precision: 0.6557
  • Recall: 0.5002
  • F1 Macro: 0.5093
  • Accuracy: 0.6202
  • F1 Binary Minimum3: 0.8655
  • F1 Binary Minimum2: 0.9307

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: 0.0001
  • train_batch_size: 16
  • eval_batch_size: 256
  • seed: 0
  • distributed_type: multi-GPU
  • num_devices: 8
  • total_train_batch_size: 128
  • total_eval_batch_size: 2048
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 200
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Macro Accuracy F1 Binary Minimum3 F1 Binary Minimum2
No log 0 0 6.9862 0.0476 0.2 0.0769 0.2382 0 0
0.4093 0.2928 1000 0.4109 0.5009 0.4593 0.4662 0.5836 0.8543 0.9213
0.4015 0.5857 2000 0.4013 0.5537 0.4682 0.4762 0.5887 0.8605 0.9208
0.4088 0.8785 3000 0.3942 0.5063 0.4684 0.4749 0.5941 0.8566 0.9245
0.399 1.1713 4000 0.3928 0.5838 0.4809 0.4902 0.5979 0.8622 0.9245
0.3884 1.4641 5000 0.3934 0.5752 0.4867 0.4955 0.6011 0.8632 0.9253
0.3962 1.7570 6000 0.3869 0.6081 0.4831 0.4927 0.6019 0.8621 0.9259
0.393 2.0498 7000 0.3999 0.5715 0.4686 0.4776 0.5826 0.8639 0.9175
0.4009 2.3426 8000 0.3823 0.6038 0.4872 0.4965 0.6056 0.8603 0.9271
0.3898 2.6354 9000 0.3895 0.5640 0.4893 0.4986 0.6018 0.8646 0.9252
0.3784 2.9283 10000 0.3813 0.6028 0.4817 0.4898 0.6052 0.8587 0.9278
0.4001 3.2211 11000 0.3806 0.6109 0.4842 0.4934 0.6051 0.8624 0.9263
0.3802 3.5139 12000 0.3828 0.6116 0.4782 0.4876 0.5992 0.8623 0.9244
0.3939 3.8067 13000 0.3832 0.6118 0.4774 0.4871 0.5980 0.8623 0.9234
0.3809 4.0996 14000 0.3800 0.6068 0.4884 0.4958 0.6098 0.8573 0.9287
0.3857 4.3924 15000 0.3786 0.6104 0.4880 0.4979 0.6063 0.8631 0.9275
0.401 4.6852 16000 0.3770 0.6241 0.4858 0.4949 0.6076 0.8613 0.9284
0.3737 4.9780 17000 0.3814 0.5728 0.4897 0.4991 0.6050 0.8647 0.9262
0.3792 5.2709 18000 0.3828 0.5930 0.4836 0.4936 0.6008 0.8651 0.9245
0.3882 5.5637 19000 0.3811 0.5800 0.4877 0.4973 0.6036 0.8645 0.9259
0.3683 5.8565 20000 0.3777 0.5816 0.4915 0.5012 0.6080 0.8643 0.9276
0.3816 6.1493 21000 0.3757 0.6353 0.4862 0.4945 0.6098 0.8600 0.9291
0.3857 6.4422 22000 0.3751 0.5801 0.4950 0.5039 0.6120 0.8627 0.9291
0.3948 6.7350 23000 0.3742 0.6232 0.4909 0.5004 0.6104 0.8617 0.9285
0.3817 7.0278 24000 0.3754 0.5829 0.4954 0.5041 0.6119 0.8626 0.9291
0.375 7.3206 25000 0.3735 0.6155 0.4916 0.4993 0.6129 0.8611 0.9291
0.3773 7.6135 26000 0.3730 0.6185 0.4923 0.5014 0.6119 0.8610 0.9291
0.3774 7.9063 27000 0.3735 0.6334 0.4879 0.4970 0.6100 0.8625 0.9288
0.3788 8.1991 28000 0.3736 0.6012 0.4934 0.5028 0.6114 0.8634 0.9288
0.3601 8.4919 29000 0.3771 0.5904 0.4921 0.5022 0.6081 0.8641 0.9276
0.3779 8.7848 30000 0.3724 0.6157 0.4934 0.5019 0.6133 0.8626 0.9293
0.3732 9.0776 31000 0.3820 0.5773 0.4939 0.5038 0.6067 0.8656 0.9272
0.3729 9.3704 32000 0.3725 0.6023 0.4934 0.5030 0.6117 0.8637 0.9289
0.378 9.6633 33000 0.3742 0.6092 0.4903 0.5002 0.6089 0.8638 0.9279
0.3794 9.9561 34000 0.3713 0.6021 0.4971 0.5063 0.6146 0.8628 0.9297
0.3703 10.2489 35000 0.3718 0.6222 0.4923 0.5016 0.6121 0.8642 0.9288
0.3614 10.5417 36000 0.3711 0.6300 0.4930 0.5023 0.6125 0.8633 0.9289
0.3822 10.8346 37000 0.3729 0.6324 0.4881 0.4951 0.6127 0.8591 0.9299
0.3775 11.1274 38000 0.3914 0.5661 0.4887 0.4986 0.5967 0.8657 0.9234
0.3709 11.4202 39000 0.3741 0.6303 0.4897 0.4958 0.6138 0.8569 0.9301
0.3749 11.7130 40000 0.3770 0.5928 0.4919 0.5021 0.6071 0.8650 0.9269
0.3683 12.0059 41000 0.3707 0.6075 0.4951 0.5037 0.6141 0.8645 0.9295
0.3772 12.2987 42000 0.3756 0.5952 0.4942 0.5042 0.6096 0.8655 0.9279
0.3829 12.5915 43000 0.3705 0.5929 0.4980 0.5070 0.6147 0.8643 0.9293
0.3736 12.8843 44000 0.3688 0.6151 0.4971 0.5059 0.6161 0.8636 0.9300
0.3795 13.1772 45000 0.3704 0.6011 0.4975 0.5071 0.6141 0.864 0.9295
0.3756 13.4700 46000 0.3704 0.6181 0.4967 0.5041 0.6168 0.8621 0.9296
0.3711 13.7628 47000 0.3684 0.6255 0.4955 0.5038 0.6160 0.8626 0.9301
0.3764 14.0556 48000 0.3696 0.6068 0.4973 0.5069 0.6150 0.8643 0.9295
0.3666 14.3485 49000 0.3688 0.6085 0.4982 0.5071 0.6158 0.8642 0.9298
0.3617 14.6413 50000 0.3738 0.6002 0.4948 0.5048 0.6104 0.8657 0.9282
0.381 14.9341 51000 0.3681 0.6233 0.4978 0.5074 0.6158 0.8642 0.9299
0.3678 15.2269 52000 0.3674 0.6279 0.4985 0.5076 0.6172 0.8645 0.9300
0.3818 15.5198 53000 0.3684 0.6307 0.4966 0.5043 0.6175 0.8626 0.9297
0.3831 15.8126 54000 0.3674 0.6156 0.4979 0.5067 0.6166 0.8639 0.9301
0.3737 16.1054 55000 0.3694 0.5929 0.4994 0.5086 0.6156 0.8651 0.9300
0.3603 16.3982 56000 0.3675 0.6336 0.4958 0.5053 0.6161 0.8643 0.9296
0.3696 16.6911 57000 0.3676 0.6210 0.4972 0.5068 0.6155 0.8648 0.9296
0.3726 16.9839 58000 0.3671 0.6501 0.4952 0.5035 0.6169 0.8627 0.9302
0.3817 17.2767 59000 0.3696 0.6388 0.4924 0.5023 0.6119 0.8648 0.9282
0.3651 17.5695 60000 0.3660 0.6414 0.4972 0.5061 0.6183 0.8639 0.9303
0.3649 17.8624 61000 0.3667 0.6346 0.4984 0.5061 0.6187 0.8628 0.9297
0.3767 18.1552 62000 0.3658 0.6303 0.5000 0.5095 0.6186 0.8641 0.9304
0.3619 18.4480 63000 0.3675 0.6397 0.4933 0.5000 0.6170 0.8608 0.9302
0.3735 18.7408 64000 0.3708 0.6168 0.4959 0.5060 0.6122 0.8663 0.9283
0.3673 19.0337 65000 0.3679 0.6050 0.5000 0.5093 0.6165 0.8654 0.9298
0.3671 19.3265 66000 0.3671 0.6489 0.4945 0.5040 0.6148 0.8649 0.9294
0.372 19.6193 67000 0.3654 0.6490 0.4975 0.5064 0.6183 0.8650 0.9302
0.3711 19.9122 68000 0.3649 0.6327 0.4997 0.5085 0.6188 0.8639 0.9304
0.36 20.2050 69000 0.3648 0.6307 0.4995 0.5077 0.6190 0.8648 0.9303
0.3718 20.4978 70000 0.3651 0.6370 0.4985 0.5079 0.6183 0.8650 0.9304
0.3762 20.7906 71000 0.3656 0.6448 0.4976 0.5067 0.6172 0.8653 0.9299
0.3602 21.0835 72000 0.3645 0.6468 0.4976 0.5059 0.6187 0.8640 0.9303
0.3624 21.3763 73000 0.3679 0.6015 0.5009 0.5105 0.6165 0.8662 0.9296
0.3774 21.6691 74000 0.3649 0.6171 0.5005 0.5094 0.6187 0.8649 0.9304
0.3724 21.9619 75000 0.3646 0.6370 0.5010 0.5102 0.6194 0.8654 0.9304
0.3762 22.2548 76000 0.3642 0.6594 0.4977 0.5068 0.6182 0.8645 0.9300
0.3795 22.5476 77000 0.3674 0.6267 0.4968 0.5067 0.6151 0.8659 0.9291
0.3774 22.8404 78000 0.3660 0.6146 0.4995 0.5090 0.6171 0.8658 0.9297
0.351 23.1332 79000 0.3648 0.6453 0.4993 0.5087 0.6180 0.8653 0.9302
0.3645 23.4261 80000 0.3643 0.6314 0.4953 0.5023 0.6193 0.8629 0.9304
0.3664 23.7189 81000 0.3644 0.6284 0.5017 0.5101 0.6202 0.8654 0.9303
0.3641 24.0117 82000 0.3655 0.6297 0.4991 0.5090 0.6174 0.8655 0.9299
0.3588 24.3045 83000 0.3646 0.6656 0.4986 0.5084 0.6179 0.8656 0.9300
0.3721 24.5974 84000 0.3633 0.6598 0.4988 0.5078 0.6197 0.8643 0.9307
0.3644 24.8902 85000 0.3636 0.6465 0.4980 0.5053 0.6202 0.8636 0.9304
0.3555 25.1830 86000 0.3631 0.6518 0.4985 0.5064 0.6203 0.8638 0.9306
0.3607 25.4758 87000 0.3641 0.6353 0.4997 0.5091 0.6186 0.8654 0.9303
0.3643 25.7687 88000 0.3629 0.6528 0.4991 0.5072 0.6210 0.8645 0.9308
0.3715 26.0615 89000 0.3644 0.6259 0.5000 0.5092 0.6184 0.8660 0.9303
0.3662 26.3543 90000 0.3642 0.6294 0.5007 0.5102 0.6190 0.8656 0.9303
0.36 26.6471 91000 0.3640 0.6215 0.5011 0.5104 0.6194 0.8657 0.9305
0.3676 26.9400 92000 0.3632 0.6607 0.4987 0.5079 0.6192 0.8656 0.9306
0.3614 27.2328 93000 0.3644 0.6186 0.5002 0.5094 0.6187 0.8659 0.9304
0.3622 27.5256 94000 0.3627 0.6478 0.4987 0.5062 0.6210 0.8642 0.9307
0.3431 27.8184 95000 0.3633 0.6520 0.5003 0.5097 0.6195 0.8654 0.9305
0.3625 28.1113 96000 0.3628 0.6578 0.5007 0.5099 0.6199 0.8656 0.9305
0.3795 28.4041 97000 0.3625 0.6436 0.4978 0.5054 0.6200 0.8646 0.9308
0.3626 28.6969 98000 0.3638 0.6261 0.5008 0.5099 0.6193 0.8657 0.9304
0.3701 28.9898 99000 0.3625 0.6597 0.4995 0.5083 0.6203 0.8647 0.9308
0.3643 29.2826 100000 0.3630 0.6583 0.5005 0.5098 0.6199 0.8654 0.9306
0.3594 29.5754 101000 0.3626 0.6603 0.4997 0.5086 0.6205 0.8651 0.9308
0.3569 29.8682 102000 0.3628 0.6557 0.5002 0.5093 0.6202 0.8655 0.9307

Framework versions

  • Transformers 4.43.4
  • Pytorch 2.4.0+cu121
  • Datasets 2.21.0
  • Tokenizers 0.19.1