Edit model card

zephyr-7b-dpo-lora-pubmedqa

This model is a fine-tuned version of EllieS/zephyr-7b-sft-qlora on the EllieS/pubmedqa_dpo_data dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0603
  • Rewards/chosen: 0.3150
  • Rewards/rejected: -2.4832
  • Rewards/accuracies: 1.0
  • Rewards/margins: 2.7982
  • Logps/rejected: -284.3339
  • Logps/chosen: -0.1594
  • Logits/rejected: -3.0974
  • Logits/chosen: -3.0713

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: 5e-06
  • train_batch_size: 4
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss Rewards/chosen Rewards/rejected Rewards/accuracies Rewards/margins Logps/rejected Logps/chosen Logits/rejected Logits/chosen
0.6917 0.0 100 0.6905 -0.0001 -0.0055 1.0 0.0054 -36.5632 -31.6725 -2.8964 -2.8908
0.6767 0.01 200 0.6758 0.0015 -0.0337 0.8000 0.0352 -39.3843 -31.5100 -2.8965 -2.8896
0.6195 0.01 300 0.6289 0.0879 -0.0489 0.8000 0.1368 -40.9016 -22.8663 -2.8974 -2.8895
0.5265 0.02 400 0.5498 0.1648 -0.1775 0.8000 0.3423 -53.7617 -15.1788 -2.8898 -2.8791
0.4265 0.02 500 0.4915 0.1800 -0.3847 0.8000 0.5647 -74.4904 -13.6600 -2.8837 -2.8704
0.3822 0.02 600 0.4624 0.1435 -0.5997 0.8000 0.7433 -95.9890 -17.3068 -2.8766 -2.8639
0.3164 0.03 700 0.4349 0.1388 -0.7367 0.8000 0.8755 -109.6869 -17.7766 -2.8752 -2.8643
0.2646 0.03 800 0.4327 0.1176 -0.8370 0.8000 0.9546 -119.7166 -19.9026 -2.8746 -2.8661
0.2358 0.04 900 0.4248 0.1070 -0.9301 0.8000 1.0371 -129.0235 -20.9559 -2.8792 -2.8718
0.2427 0.04 1000 0.4057 0.1150 -1.0071 0.8000 1.1221 -136.7261 -20.1632 -2.8785 -2.8730
0.2167 0.04 1100 0.3903 0.1183 -1.0874 0.8000 1.2057 -144.7528 -19.8280 -2.8804 -2.8753
0.1888 0.05 1200 0.3909 0.1086 -1.1508 0.8000 1.2593 -151.0921 -20.8049 -2.8784 -2.8719
0.17 0.05 1300 0.4013 0.0826 -1.2194 0.8000 1.3020 -157.9523 -23.4004 -2.8826 -2.8756
0.139 0.06 1400 0.3722 0.1104 -1.2731 0.8000 1.3834 -163.3211 -20.6215 -2.8859 -2.8795
0.1525 0.06 1500 0.3450 0.1389 -1.3306 0.8000 1.4695 -169.0768 -17.7689 -2.8962 -2.8891
0.1622 0.06 1600 0.3556 0.1096 -1.3911 0.8000 1.5007 -175.1273 -20.7011 -2.9043 -2.8991
0.1242 0.07 1700 0.3780 0.0727 -1.4364 0.8000 1.5091 -179.6512 -24.3892 -2.9111 -2.9064
0.1232 0.07 1800 0.3637 0.0774 -1.4857 0.8000 1.5631 -184.5862 -23.9169 -2.9201 -2.9152
0.1141 0.08 1900 0.3232 0.1026 -1.5624 0.8000 1.6651 -192.2589 -21.3979 -2.9257 -2.9200
0.1525 0.08 2000 0.3077 0.0918 -1.6428 0.8000 1.7346 -200.2968 -22.4783 -2.9355 -2.9305
0.1236 0.08 2100 0.4024 0.0066 -1.6566 0.8000 1.6632 -201.6738 -31.0022 -2.9432 -2.9440
0.1467 0.09 2200 0.3299 0.0719 -1.7345 0.8000 1.8064 -209.4659 -24.4665 -2.9485 -2.9427
0.1314 0.09 2300 0.3409 0.0424 -1.8004 0.8000 1.8428 -216.0566 -27.4246 -2.9606 -2.9549
0.0827 0.1 2400 0.3526 0.0279 -1.8371 0.8000 1.8651 -219.7282 -28.8650 -2.9690 -2.9643
0.132 0.1 2500 0.2030 0.1635 -1.9493 0.8000 2.1128 -230.9491 -15.3147 -2.9840 -2.9775
0.0851 0.1 2600 0.1688 0.2903 -1.9136 1.0 2.2039 -227.3723 -2.6290 -2.9917 -2.9828
0.1549 0.11 2700 0.1753 0.2794 -1.9249 1.0 2.2043 -228.5043 -3.7234 -2.9974 -2.9883
0.0822 0.11 2800 0.2224 0.1933 -1.9332 0.8000 2.1265 -229.3374 -12.3327 -3.0048 -2.9948
0.0862 0.12 2900 0.2089 0.2191 -1.9430 0.8000 2.1621 -230.3155 -9.7481 -3.0075 -2.9997
0.0702 0.12 3000 0.2735 0.1033 -1.9629 0.8000 2.0662 -232.3102 -21.3305 -3.0130 -3.0048
0.0865 0.12 3100 0.0754 0.3161 -2.2881 1.0 2.6042 -264.8222 -0.0494 -3.0220 -3.0058
0.0763 0.13 3200 0.0745 0.3162 -2.2987 1.0 2.6149 -265.8853 -0.0435 -3.0279 -3.0137
0.11 0.13 3300 0.1464 0.3107 -2.0080 1.0 2.3186 -236.8128 -0.5936 -3.0307 -3.0167
0.1114 0.14 3400 0.1534 0.2652 -2.0338 1.0 2.2989 -239.3922 -5.1433 -3.0296 -3.0148
0.0934 0.14 3500 0.1702 0.2508 -2.0232 1.0 2.2741 -238.3399 -6.5776 -3.0312 -3.0213
0.063 0.14 3600 0.0717 0.3162 -2.3330 1.0 2.6492 -269.3163 -0.0426 -3.0421 -3.0220
0.0673 0.15 3700 0.1843 0.2409 -2.0103 1.0 2.2512 -237.0486 -7.5727 -3.0412 -3.0296
0.1292 0.15 3800 0.0723 0.3161 -2.3246 1.0 2.6407 -268.4736 -0.0498 -3.0497 -3.0278
0.0683 0.16 3900 0.0699 0.3161 -2.3559 1.0 2.6720 -271.6014 -0.0451 -3.0506 -3.0321
0.1472 0.16 4000 0.0697 0.3161 -2.3568 1.0 2.6729 -271.6938 -0.0454 -3.0525 -3.0307
0.104 0.16 4100 0.0977 0.1714 -2.3361 1.0 2.5074 -269.6208 -14.5249 -3.0556 -3.0361
0.103 0.17 4200 0.0802 0.3159 -2.2830 1.0 2.5990 -264.3194 -0.0656 -3.0572 -3.0348
0.0939 0.17 4300 0.0705 0.3161 -2.3567 1.0 2.6728 -271.6851 -0.0492 -3.0577 -3.0352
0.1012 0.18 4400 0.0687 0.3161 -2.3729 1.0 2.6889 -273.3021 -0.0526 -3.0618 -3.0388
0.0928 0.18 4500 0.0678 0.3159 -2.3822 1.0 2.6981 -274.2347 -0.0664 -3.0640 -3.0422
0.092 0.18 4600 0.0681 0.3161 -2.3780 1.0 2.6941 -273.8116 -0.0472 -3.0627 -3.0396
0.0966 0.19 4700 0.0685 0.3159 -2.3792 1.0 2.6952 -273.9405 -0.0664 -3.0634 -3.0394
0.0605 0.19 4800 0.0679 0.3130 -2.3880 1.0 2.7011 -274.8187 -0.3556 -3.0658 -3.0441
0.1031 0.2 4900 0.0748 0.2545 -2.3915 1.0 2.6460 -275.1633 -6.2069 -3.0700 -3.0534
0.093 0.2 5000 0.0670 0.3159 -2.3964 1.0 2.7123 -275.6591 -0.0692 -3.0697 -3.0463
0.1176 0.2 5100 0.0728 0.2609 -2.3988 1.0 2.6597 -275.8960 -5.5717 -3.0660 -3.0456
0.1232 0.21 5200 0.0659 0.3157 -2.4090 1.0 2.7246 -276.9131 -0.0931 -3.0683 -3.0432
0.1606 0.21 5300 0.0654 0.3158 -2.4143 1.0 2.7300 -277.4427 -0.0841 -3.0748 -3.0497
0.0604 0.22 5400 0.0655 0.3159 -2.4142 1.0 2.7301 -277.4360 -0.0707 -3.0763 -3.0505
0.0966 0.22 5500 0.0653 0.3158 -2.4151 1.0 2.7308 -277.5210 -0.0842 -3.0783 -3.0537
0.0609 0.22 5600 0.0662 0.3159 -2.4082 1.0 2.7241 -276.8354 -0.0684 -3.0757 -3.0549
0.0583 0.23 5700 0.0661 0.3157 -2.4106 1.0 2.7263 -277.0789 -0.0922 -3.0779 -3.0542
0.0721 0.23 5800 0.0648 0.3157 -2.4224 1.0 2.7381 -278.2588 -0.0944 -3.0766 -3.0521
0.0935 0.24 5900 0.0649 0.3158 -2.4218 1.0 2.7376 -278.1963 -0.0777 -3.0794 -3.0554
0.1559 0.24 6000 0.0648 0.3158 -2.4243 1.0 2.7401 -278.4456 -0.0830 -3.0778 -3.0567
0.0973 0.24 6100 0.0645 0.3159 -2.4224 1.0 2.7384 -278.2578 -0.0666 -3.0870 -3.0614
0.1298 0.25 6200 0.0640 0.3155 -2.4313 1.0 2.7469 -279.1472 -0.1059 -3.0851 -3.0611
0.1037 0.25 6300 0.0648 0.3153 -2.4239 1.0 2.7393 -278.4086 -0.1257 -3.0885 -3.0585
0.0939 0.26 6400 0.0633 0.3154 -2.4417 1.0 2.7571 -280.1844 -0.1197 -3.0841 -3.0594
0.0588 0.26 6500 0.0635 0.3152 -2.4399 1.0 2.7551 -280.0093 -0.1416 -3.0841 -3.0580
0.1025 0.26 6600 0.0637 0.3143 -2.4371 1.0 2.7514 -279.7303 -0.2316 -3.0867 -3.0576
0.0583 0.27 6700 0.0636 0.3155 -2.4383 1.0 2.7538 -279.8492 -0.1092 -3.0834 -3.0596
0.0882 0.27 6800 0.0630 0.3154 -2.4476 1.0 2.7631 -280.7805 -0.1187 -3.0858 -3.0619
0.0744 0.28 6900 0.0637 0.3152 -2.4389 1.0 2.7541 -279.9033 -0.1358 -3.0874 -3.0588
0.0555 0.28 7000 0.0632 0.3153 -2.4449 1.0 2.7603 -280.5067 -0.1259 -3.0870 -3.0640
0.0745 0.28 7100 0.0644 0.3148 -2.4314 1.0 2.7463 -279.1595 -0.1782 -3.0864 -3.0591
0.112 0.29 7200 0.0634 0.3152 -2.4435 1.0 2.7587 -280.3658 -0.1356 -3.0870 -3.0601
0.0566 0.29 7300 0.0630 0.3154 -2.4478 1.0 2.7632 -280.7963 -0.1159 -3.0888 -3.0650
0.0632 0.3 7400 0.0632 0.3150 -2.4466 1.0 2.7616 -280.6743 -0.1563 -3.0852 -3.0647
0.1134 0.3 7500 0.0626 0.3151 -2.4519 1.0 2.7670 -281.2076 -0.1470 -3.0885 -3.0592
0.0802 0.3 7600 0.0626 0.3153 -2.4532 1.0 2.7685 -281.3376 -0.1320 -3.0896 -3.0624
0.0577 0.31 7700 0.1374 0.3151 -2.0953 1.0 2.4105 -245.5485 -0.1450 -3.0868 -3.0604
0.1363 0.31 7800 0.0626 0.3150 -2.4535 1.0 2.7685 -281.3674 -0.1637 -3.0920 -3.0651
0.1268 0.32 7900 0.0621 0.3151 -2.4577 1.0 2.7727 -281.7813 -0.1512 -3.0910 -3.0626
0.0593 0.32 8000 0.0618 0.3153 -2.4621 1.0 2.7774 -282.2267 -0.1342 -3.0912 -3.0626
0.1578 0.32 8100 0.0621 0.3154 -2.4563 1.0 2.7717 -281.6464 -0.1229 -3.0946 -3.0659
0.1307 0.33 8200 0.0620 0.3153 -2.4588 1.0 2.7741 -281.8946 -0.1311 -3.0971 -3.0699
0.0743 0.33 8300 0.0613 0.3152 -2.4694 1.0 2.7846 -282.9548 -0.1430 -3.0962 -3.0691
0.0892 0.34 8400 0.0614 0.3152 -2.4670 1.0 2.7822 -282.7147 -0.1406 -3.0969 -3.0686
0.091 0.34 8500 0.0625 0.3151 -2.4544 1.0 2.7696 -281.4597 -0.1490 -3.1052 -3.0734
0.1245 0.34 8600 0.0618 0.3153 -2.4570 1.0 2.7723 -281.7121 -0.1273 -3.1034 -3.0742
0.1212 0.35 8700 0.0609 0.3151 -2.4725 1.0 2.7876 -283.2621 -0.1509 -3.1016 -3.0750
0.062 0.35 8800 0.0609 0.3148 -2.4739 1.0 2.7887 -283.4044 -0.1800 -3.1021 -3.0740
0.1407 0.36 8900 0.0609 0.3151 -2.4735 1.0 2.7886 -283.3669 -0.1549 -3.1023 -3.0728
0.0594 0.36 9000 0.0605 0.3150 -2.4794 1.0 2.7944 -283.9577 -0.1585 -3.1003 -3.0703
0.065 0.36 9100 0.0604 0.3153 -2.4803 1.0 2.7956 -284.0407 -0.1291 -3.1035 -3.0733
0.0942 0.37 9200 0.0605 0.3152 -2.4780 1.0 2.7932 -283.8191 -0.1447 -3.1043 -3.0751
0.0863 0.37 9300 0.0608 0.3152 -2.4752 1.0 2.7904 -283.5404 -0.1411 -3.1016 -3.0714
0.084 0.38 9400 0.0606 0.3153 -2.4782 1.0 2.7934 -283.8343 -0.1349 -3.1007 -3.0726
0.1181 0.38 9500 0.0605 0.3150 -2.4798 1.0 2.7948 -283.9931 -0.1562 -3.1032 -3.0748
0.1231 0.38 9600 0.0603 0.3148 -2.4829 1.0 2.7977 -284.3057 -0.1813 -3.1009 -3.0725
0.1187 0.39 9700 0.0603 0.3151 -2.4813 1.0 2.7964 -284.1463 -0.1474 -3.1023 -3.0728
0.1663 0.39 9800 0.0604 0.3149 -2.4799 1.0 2.7948 -284.0080 -0.1717 -3.1038 -3.0729
0.082 0.4 9900 0.0603 0.3150 -2.4808 1.0 2.7958 -284.0976 -0.1608 -3.1028 -3.0724
0.0561 0.4 10000 0.0601 0.3150 -2.4849 1.0 2.7999 -284.5103 -0.1611 -3.1023 -3.0735
0.1215 0.4 10100 0.0602 0.3149 -2.4835 1.0 2.7983 -284.3610 -0.1741 -3.1025 -3.0729
0.0617 0.41 10200 0.0600 0.3152 -2.4852 1.0 2.8004 -284.5347 -0.1360 -3.1022 -3.0739
0.0834 0.41 10300 0.0599 0.3152 -2.4869 1.0 2.8021 -284.7045 -0.1425 -3.1017 -3.0739
0.0931 0.42 10400 0.0604 0.3149 -2.4814 1.0 2.7963 -284.1569 -0.1700 -3.1008 -3.0727
0.0576 0.42 10500 0.0600 0.3152 -2.4850 1.0 2.8002 -284.5181 -0.1442 -3.0986 -3.0707
0.0882 0.42 10600 0.0606 0.3146 -2.4799 1.0 2.7946 -284.0095 -0.1975 -3.0995 -3.0712
0.0893 0.43 10700 0.0599 0.3149 -2.4865 1.0 2.8014 -284.6656 -0.1668 -3.1019 -3.0728
0.0551 0.43 10800 0.0597 0.3145 -2.4914 1.0 2.8059 -285.1605 -0.2109 -3.1032 -3.0748
0.0875 0.44 10900 0.0598 0.3148 -2.4889 1.0 2.8037 -284.9028 -0.1783 -3.1026 -3.0747
0.0902 0.44 11000 0.0597 0.3150 -2.4892 1.0 2.8042 -284.9340 -0.1595 -3.1028 -3.0734
0.0619 0.44 11100 0.0600 0.3153 -2.4867 1.0 2.8019 -284.6834 -0.1342 -3.1034 -3.0740
0.1255 0.45 11200 0.0598 0.3153 -2.4891 1.0 2.8044 -284.9269 -0.1275 -3.1024 -3.0746
0.0564 0.45 11300 0.0600 0.3155 -2.4855 1.0 2.8009 -284.5616 -0.1136 -3.1028 -3.0736
0.0554 0.46 11400 0.0599 0.3151 -2.4881 1.0 2.8032 -284.8240 -0.1496 -3.0997 -3.0722
0.0543 0.46 11500 0.0604 0.3152 -2.4811 1.0 2.7963 -284.1210 -0.1361 -3.1028 -3.0740
0.0568 0.46 11600 0.0598 0.3147 -2.4902 1.0 2.8049 -285.0388 -0.1892 -3.1031 -3.0762
0.0562 0.47 11700 0.0602 0.3148 -2.4845 1.0 2.7993 -284.4625 -0.1755 -3.1001 -3.0733
0.056 0.47 11800 0.0600 0.3146 -2.4882 1.0 2.8028 -284.8389 -0.2018 -3.1000 -3.0738
0.0557 0.48 11900 0.0601 0.3148 -2.4855 1.0 2.8003 -284.5624 -0.1753 -3.0986 -3.0718
0.0913 0.48 12000 0.0600 0.3151 -2.4873 1.0 2.8024 -284.7490 -0.1495 -3.0990 -3.0730
0.1144 0.48 12100 0.0602 0.3150 -2.4849 1.0 2.7999 -284.5060 -0.1564 -3.0945 -3.0693
0.0911 0.49 12200 0.0603 0.3151 -2.4833 1.0 2.7984 -284.3466 -0.1472 -3.0992 -3.0716
0.091 0.49 12300 0.0604 0.3149 -2.4833 1.0 2.7982 -284.3491 -0.1747 -3.0975 -3.0713
0.0548 0.5 12400 0.0606 0.3143 -2.4812 1.0 2.7956 -284.1383 -0.2264 -3.0971 -3.0704
0.0622 0.5 12500 0.0604 0.3146 -2.4819 1.0 2.7964 -284.2023 -0.2027 -3.0979 -3.0709
0.056 0.5 12600 0.0601 0.3148 -2.4851 1.0 2.7999 -284.5273 -0.1777 -3.0984 -3.0740
0.1175 0.51 12700 0.0601 0.3146 -2.4868 1.0 2.8014 -284.7003 -0.2008 -3.0952 -3.0735
0.0714 0.51 12800 0.0605 0.3152 -2.4818 1.0 2.7969 -284.1913 -0.1419 -3.0964 -3.0714
0.0572 0.52 12900 0.0600 0.3151 -2.4863 1.0 2.8014 -284.6450 -0.1463 -3.0980 -3.0730
0.09 0.52 13000 0.0601 0.3152 -2.4854 1.0 2.8006 -284.5557 -0.1361 -3.0992 -3.0733
0.055 0.52 13100 0.0601 0.3147 -2.4865 1.0 2.8012 -284.6667 -0.1921 -3.0975 -3.0725
0.0937 0.53 13200 0.0608 0.3149 -2.4763 1.0 2.7912 -283.6473 -0.1729 -3.0956 -3.0706
0.0543 0.53 13300 0.0604 0.3148 -2.4826 1.0 2.7974 -284.2760 -0.1764 -3.0962 -3.0705
0.0549 0.54 13400 0.0601 0.3152 -2.4860 1.0 2.8012 -284.6145 -0.1399 -3.0965 -3.0708
0.057 0.54 13500 0.0604 0.3148 -2.4821 1.0 2.7969 -284.2305 -0.1809 -3.0939 -3.0694
0.2197 0.54 13600 0.0602 0.3152 -2.4842 1.0 2.7994 -284.4350 -0.1356 -3.0985 -3.0721
0.0871 0.55 13700 0.0601 0.3151 -2.4850 1.0 2.8002 -284.5201 -0.1456 -3.0979 -3.0717
0.1127 0.55 13800 0.0608 0.3151 -2.4771 1.0 2.7922 -283.7219 -0.1474 -3.0968 -3.0709
0.0862 0.56 13900 0.0609 0.3150 -2.4758 1.0 2.7908 -283.5951 -0.1580 -3.0982 -3.0726
0.0862 0.56 14000 0.0613 0.3149 -2.4719 1.0 2.7868 -283.2067 -0.1697 -3.0952 -3.0701
0.0904 0.56 14100 0.0611 0.3148 -2.4745 1.0 2.7893 -283.4654 -0.1784 -3.0964 -3.0712
0.0932 0.57 14200 0.0608 0.3148 -2.4784 1.0 2.7932 -283.8593 -0.1848 -3.0954 -3.0698
0.0765 0.57 14300 0.0610 0.3146 -2.4759 1.0 2.7906 -283.6089 -0.1970 -3.0971 -3.0715
0.076 0.58 14400 0.0604 0.3148 -2.4822 1.0 2.7969 -284.2308 -0.1812 -3.0974 -3.0716
0.0553 0.58 14500 0.0608 0.3144 -2.4788 1.0 2.7932 -283.8933 -0.2187 -3.0962 -3.0722
0.0975 0.58 14600 0.0604 0.3149 -2.4825 1.0 2.7974 -284.2697 -0.1702 -3.0973 -3.0710
0.0544 0.59 14700 0.0607 0.3148 -2.4793 1.0 2.7941 -283.9458 -0.1840 -3.0964 -3.0708
0.0545 0.59 14800 0.0605 0.3151 -2.4812 1.0 2.7963 -284.1385 -0.1515 -3.0972 -3.0717
0.0902 0.6 14900 0.0606 0.3148 -2.4812 1.0 2.7959 -284.1325 -0.1829 -3.0980 -3.0710
0.059 0.6 15000 0.0599 0.3150 -2.4877 1.0 2.8027 -284.7884 -0.1639 -3.0972 -3.0708
0.1524 0.6 15100 0.0601 0.3147 -2.4862 1.0 2.8009 -284.6385 -0.1890 -3.0986 -3.0718
0.0558 0.61 15200 0.0605 0.3149 -2.4815 1.0 2.7964 -284.1689 -0.1694 -3.0972 -3.0715
0.0969 0.61 15300 0.0601 0.3151 -2.4861 1.0 2.8012 -284.6226 -0.1482 -3.0963 -3.0704
0.0569 0.62 15400 0.0600 0.3151 -2.4876 1.0 2.8027 -284.7789 -0.1546 -3.0975 -3.0719
0.0935 0.62 15500 0.0603 0.3151 -2.4838 1.0 2.7989 -284.3983 -0.1488 -3.0961 -3.0698
0.0815 0.62 15600 0.0602 0.3143 -2.4861 1.0 2.8004 -284.6287 -0.2292 -3.0973 -3.0711
0.0556 0.63 15700 0.0603 0.3148 -2.4832 1.0 2.7979 -284.3312 -0.1816 -3.0971 -3.0709
0.0559 0.63 15800 0.0607 0.3147 -2.4795 1.0 2.7942 -283.9677 -0.1932 -3.0964 -3.0704
0.0871 0.64 15900 0.0606 0.3148 -2.4804 1.0 2.7952 -284.0538 -0.1793 -3.0967 -3.0718
0.0551 0.64 16000 0.0607 0.3146 -2.4792 1.0 2.7938 -283.9374 -0.2004 -3.0957 -3.0699
0.1236 0.64 16100 0.0606 0.3149 -2.4805 1.0 2.7954 -284.0651 -0.1657 -3.0981 -3.0712
0.0768 0.65 16200 0.0605 0.3152 -2.4818 1.0 2.7969 -284.1944 -0.1449 -3.0964 -3.0698
0.0923 0.65 16300 0.0606 0.3147 -2.4809 1.0 2.7957 -284.1096 -0.1852 -3.0964 -3.0711
0.0956 0.66 16400 0.0606 0.3150 -2.4811 1.0 2.7962 -284.1297 -0.1552 -3.0961 -3.0700
0.0552 0.66 16500 0.0605 0.3151 -2.4811 1.0 2.7962 -284.1294 -0.1549 -3.0966 -3.0703
0.0921 0.66 16600 0.0606 0.3150 -2.4792 1.0 2.7942 -283.9340 -0.1596 -3.0962 -3.0705
0.1013 0.67 16700 0.0605 0.3150 -2.4812 1.0 2.7962 -284.1385 -0.1641 -3.0952 -3.0693
0.0583 0.67 16800 0.0607 0.3152 -2.4790 1.0 2.7942 -283.9129 -0.1385 -3.0957 -3.0703
0.0564 0.68 16900 0.0605 0.3151 -2.4816 1.0 2.7967 -284.1773 -0.1529 -3.0970 -3.0707
0.0925 0.68 17000 0.0603 0.3149 -2.4838 1.0 2.7987 -284.3918 -0.1673 -3.0960 -3.0705
0.056 0.68 17100 0.0606 0.3151 -2.4814 1.0 2.7964 -284.1519 -0.1524 -3.0951 -3.0697
0.0923 0.69 17200 0.0605 0.3150 -2.4822 1.0 2.7972 -284.2387 -0.1643 -3.0958 -3.0703
0.0832 0.69 17300 0.0602 0.3150 -2.4859 1.0 2.8009 -284.6101 -0.1606 -3.0962 -3.0707
0.171 0.7 17400 0.0605 0.3150 -2.4817 1.0 2.7967 -284.1821 -0.1576 -3.0961 -3.0701
0.0823 0.7 17500 0.0606 0.3149 -2.4806 1.0 2.7954 -284.0712 -0.1717 -3.0964 -3.0699
0.0707 0.7 17600 0.0608 0.3150 -2.4775 1.0 2.7924 -283.7639 -0.1645 -3.0964 -3.0705
0.0906 0.71 17700 0.0607 0.3150 -2.4787 1.0 2.7937 -283.8846 -0.1602 -3.0963 -3.0701
0.0622 0.71 17800 0.0606 0.3148 -2.4811 1.0 2.7959 -284.1299 -0.1805 -3.0950 -3.0692
0.0548 0.72 17900 0.0603 0.3148 -2.4834 1.0 2.7982 -284.3509 -0.1764 -3.0954 -3.0708
0.0557 0.72 18000 0.0603 0.3148 -2.4839 1.0 2.7987 -284.4059 -0.1814 -3.0963 -3.0701
0.0551 0.72 18100 0.0603 0.3147 -2.4834 1.0 2.7982 -284.3599 -0.1855 -3.0952 -3.0702
0.0688 0.73 18200 0.0603 0.3145 -2.4840 1.0 2.7984 -284.4141 -0.2146 -3.0954 -3.0694
0.0897 0.73 18300 0.0604 0.3149 -2.4823 1.0 2.7972 -284.2420 -0.1675 -3.0961 -3.0703
0.1115 0.74 18400 0.0604 0.3147 -2.4827 1.0 2.7974 -284.2878 -0.1883 -3.0952 -3.0703
0.0914 0.74 18500 0.0603 0.3148 -2.4838 1.0 2.7987 -284.4003 -0.1758 -3.0967 -3.0712
0.092 0.74 18600 0.0604 0.3149 -2.4836 1.0 2.7984 -284.3739 -0.1744 -3.0962 -3.0709
0.0641 0.75 18700 0.0607 0.3150 -2.4795 1.0 2.7944 -283.9636 -0.1641 -3.0953 -3.0704
0.0576 0.75 18800 0.0607 0.3147 -2.4797 1.0 2.7944 -283.9855 -0.1860 -3.0961 -3.0708
0.0539 0.76 18900 0.0606 0.3150 -2.4802 1.0 2.7952 -284.0349 -0.1604 -3.0960 -3.0708
0.0935 0.76 19000 0.0606 0.3143 -2.4806 1.0 2.7949 -284.0767 -0.2273 -3.0964 -3.0711
0.0887 0.76 19100 0.0605 0.3150 -2.4812 1.0 2.7962 -284.1394 -0.1649 -3.0951 -3.0697
0.1274 0.77 19200 0.0605 0.3144 -2.4823 1.0 2.7967 -284.2411 -0.2167 -3.0961 -3.0706
0.1333 0.77 19300 0.0604 0.3150 -2.4831 1.0 2.7982 -284.3301 -0.1556 -3.0949 -3.0695
0.0551 0.78 19400 0.0603 0.3148 -2.4839 1.0 2.7987 -284.4022 -0.1777 -3.0967 -3.0713
0.0575 0.78 19500 0.0604 0.3148 -2.4824 1.0 2.7972 -284.2572 -0.1827 -3.0964 -3.0711
0.0581 0.78 19600 0.0604 0.3148 -2.4831 1.0 2.7979 -284.3292 -0.1798 -3.0967 -3.0707
0.0869 0.79 19700 0.0603 0.3146 -2.4841 1.0 2.7987 -284.4286 -0.2042 -3.0962 -3.0708
0.0724 0.79 19800 0.0603 0.3151 -2.4844 1.0 2.7994 -284.4534 -0.1539 -3.0955 -3.0712
0.1063 0.8 19900 0.0604 0.3149 -2.4830 1.0 2.7979 -284.3166 -0.1672 -3.0965 -3.0702
0.0857 0.8 20000 0.0603 0.3149 -2.4843 1.0 2.7992 -284.4483 -0.1739 -3.0958 -3.0702
0.0561 0.8 20100 0.0603 0.3145 -2.4849 1.0 2.7994 -284.5063 -0.2069 -3.0964 -3.0707
0.0556 0.81 20200 0.0604 0.3149 -2.4828 1.0 2.7977 -284.2974 -0.1730 -3.0955 -3.0696
0.091 0.81 20300 0.0603 0.3151 -2.4846 1.0 2.7997 -284.4747 -0.1502 -3.0950 -3.0689
0.0899 0.82 20400 0.0603 0.3147 -2.4847 1.0 2.7994 -284.4896 -0.1901 -3.0964 -3.0700
0.0849 0.82 20500 0.0604 0.3151 -2.4824 1.0 2.7974 -284.2542 -0.1547 -3.0959 -3.0701
0.092 0.82 20600 0.0603 0.3145 -2.4852 1.0 2.7997 -284.5341 -0.2096 -3.0964 -3.0713
0.0871 0.83 20700 0.0604 0.3151 -2.4828 1.0 2.7979 -284.3003 -0.1508 -3.0957 -3.0698
0.0915 0.83 20800 0.0603 0.3148 -2.4838 1.0 2.7987 -284.4004 -0.1760 -3.0960 -3.0707
0.0566 0.84 20900 0.0602 0.3150 -2.4845 1.0 2.7994 -284.4606 -0.1612 -3.0951 -3.0690
0.0552 0.84 21000 0.0603 0.3145 -2.4849 1.0 2.7994 -284.5057 -0.2062 -3.0963 -3.0711
0.164 0.84 21100 0.0603 0.3149 -2.4840 1.0 2.7989 -284.4156 -0.1661 -3.0961 -3.0705
0.0829 0.85 21200 0.0604 0.3145 -2.4834 1.0 2.7979 -284.3568 -0.2073 -3.0961 -3.0707
0.0552 0.85 21300 0.0602 0.3145 -2.4857 1.0 2.8002 -284.5827 -0.2082 -3.0954 -3.0706
0.0797 0.86 21400 0.0603 0.3148 -2.4839 1.0 2.7987 -284.4014 -0.1769 -3.0967 -3.0708
0.0569 0.86 21500 0.0604 0.3149 -2.4833 1.0 2.7982 -284.3410 -0.1666 -3.0953 -3.0704
0.1239 0.86 21600 0.0603 0.3152 -2.4830 1.0 2.7982 -284.3148 -0.1403 -3.0961 -3.0710
0.1285 0.87 21700 0.0602 0.3148 -2.4854 1.0 2.8002 -284.5537 -0.1793 -3.0966 -3.0700
0.0557 0.87 21800 0.0603 0.3146 -2.4843 1.0 2.7989 -284.4450 -0.1955 -3.0965 -3.0704
0.0999 0.88 21900 0.0604 0.3149 -2.4833 1.0 2.7982 -284.3460 -0.1716 -3.0962 -3.0708
0.0704 0.88 22000 0.0604 0.3151 -2.4826 1.0 2.7977 -284.2776 -0.1531 -3.0955 -3.0703
0.0902 0.88 22100 0.0603 0.3148 -2.4837 1.0 2.7984 -284.3843 -0.1848 -3.0950 -3.0700
0.0842 0.89 22200 0.0603 0.3148 -2.4848 1.0 2.7997 -284.5004 -0.1761 -3.0966 -3.0708
0.091 0.89 22300 0.0603 0.3151 -2.4831 1.0 2.7982 -284.3220 -0.1475 -3.0962 -3.0710
0.0612 0.9 22400 0.0604 0.3149 -2.4828 1.0 2.7977 -284.2958 -0.1713 -3.0969 -3.0716
0.0555 0.9 22500 0.0603 0.3148 -2.4846 1.0 2.7994 -284.4779 -0.1785 -3.0965 -3.0706
0.1258 0.9 22600 0.0602 0.3149 -2.4853 1.0 2.8002 -284.5466 -0.1720 -3.0963 -3.0704
0.091 0.91 22700 0.0604 0.3150 -2.4827 1.0 2.7977 -284.2870 -0.1626 -3.0966 -3.0709
0.0939 0.91 22800 0.0604 0.3149 -2.4823 1.0 2.7972 -284.2432 -0.1687 -3.0964 -3.0707
0.057 0.92 22900 0.0603 0.3148 -2.4836 1.0 2.7984 -284.3766 -0.1771 -3.0973 -3.0711
0.1252 0.92 23000 0.0603 0.3149 -2.4835 1.0 2.7984 -284.3660 -0.1666 -3.0956 -3.0703
0.2019 0.92 23100 0.0604 0.3147 -2.4828 1.0 2.7974 -284.2932 -0.1937 -3.0962 -3.0714
0.0905 0.93 23200 0.0603 0.3149 -2.4848 1.0 2.7997 -284.4948 -0.1704 -3.0966 -3.0713
0.1121 0.93 23300 0.0602 0.3149 -2.4850 1.0 2.7999 -284.5165 -0.1670 -3.0955 -3.0710
0.0559 0.94 23400 0.0603 0.3149 -2.4840 1.0 2.7989 -284.4164 -0.1670 -3.0968 -3.0712
0.0555 0.94 23500 0.0603 0.3148 -2.4847 1.0 2.7994 -284.4821 -0.1826 -3.0958 -3.0703
0.093 0.94 23600 0.0602 0.3147 -2.4854 1.0 2.8002 -284.5602 -0.1857 -3.0959 -3.0706
0.0561 0.95 23700 0.0603 0.3149 -2.4838 1.0 2.7987 -284.3966 -0.1721 -3.0959 -3.0700
0.0576 0.95 23800 0.0603 0.3147 -2.4848 1.0 2.7994 -284.4909 -0.1915 -3.0966 -3.0703
0.1255 0.96 23900 0.0603 0.3147 -2.4843 1.0 2.7989 -284.4434 -0.1940 -3.0964 -3.0707
0.094 0.96 24000 0.0603 0.3147 -2.4840 1.0 2.7987 -284.4168 -0.1923 -3.0961 -3.0701
0.0543 0.96 24100 0.0603 0.3148 -2.4844 1.0 2.7992 -284.4561 -0.1817 -3.0963 -3.0710
0.0551 0.97 24200 0.0603 0.3148 -2.4844 1.0 2.7992 -284.4572 -0.1828 -3.0959 -3.0703
0.1155 0.97 24300 0.0603 0.3149 -2.4840 1.0 2.7989 -284.4147 -0.1652 -3.0971 -3.0712
0.0549 0.98 24400 0.0604 0.3149 -2.4822 1.0 2.7972 -284.2398 -0.1654 -3.0956 -3.0701
0.0556 0.98 24500 0.0602 0.3150 -2.4854 1.0 2.8004 -284.5563 -0.1569 -3.0960 -3.0699
0.0693 0.98 24600 0.0604 0.3147 -2.4835 1.0 2.7982 -284.3652 -0.1907 -3.0963 -3.0711
0.0919 0.99 24700 0.0604 0.3147 -2.4835 1.0 2.7982 -284.3639 -0.1894 -3.0962 -3.0710
0.0924 0.99 24800 0.0604 0.3148 -2.4834 1.0 2.7982 -284.3547 -0.1803 -3.0963 -3.0707
0.0575 1.0 24900 0.0603 0.3146 -2.4841 1.0 2.7987 -284.4260 -0.2015 -3.0973 -3.0712
0.0884 1.0 25000 0.0603 0.3150 -2.4832 1.0 2.7982 -284.3339 -0.1594 -3.0974 -3.0713

Framework versions

  • PEFT 0.7.1
  • Transformers 4.36.2
  • Pytorch 2.1.1+cu121
  • Datasets 2.14.6
  • Tokenizers 0.15.0
Downloads last month
0
Unable to determine this model’s pipeline type. Check the docs .

Adapter for

Dataset used to train EllieS/zephyr-7b-dpo-lora-pubmedqa