zephyr-7b-dpo-qlora / README.md
mashinaalive's picture
End of training
8cd5ed5 verified
metadata
license: apache-2.0
library_name: peft
tags:
  - alignment-handbook
  - trl
  - dpo
  - generated_from_trainer
  - trl
  - dpo
  - generated_from_trainer
base_model: mistralai/Mistral-7B-v0.1
datasets:
  - HuggingFaceH4/ultrafeedback_binarized
model-index:
  - name: zephyr-7b-dpo-qlora
    results: []

zephyr-7b-dpo-qlora

This model is a fine-tuned version of alignment-handbook/zephyr-7b-sft-qlora on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5156
  • Rewards/chosen: -4.0806
  • Rewards/rejected: -5.8791
  • Rewards/accuracies: 0.7495
  • Rewards/margins: 1.7985
  • Logps/rejected: -832.4777
  • Logps/chosen: -672.6758
  • Logits/rejected: -1.1337
  • Logits/chosen: -1.4991

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: 1
  • eval_batch_size: 1
  • seed: 42
  • distributed_type: multi-GPU
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 4
  • 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
0.6914 0.01 100 0.6910 0.0199 0.0156 0.6220 0.0043 -243.0070 -262.6279 -2.9204 -2.9325
0.6877 0.01 200 0.6869 0.0449 0.0321 0.6255 0.0128 -241.3639 -260.1325 -2.9210 -2.9353
0.6841 0.02 300 0.6804 0.0577 0.0306 0.6495 0.0270 -241.5080 -258.8525 -2.9183 -2.9327
0.6737 0.03 400 0.6713 0.0481 -0.0000 0.6550 0.0481 -244.5744 -259.8077 -2.8962 -2.9118
0.6443 0.03 500 0.6547 -0.0859 -0.1788 0.6725 0.0929 -262.4492 -273.2110 -2.8544 -2.8722
0.6257 0.04 600 0.6467 -0.1409 -0.2585 0.6700 0.1176 -270.4241 -278.7132 -2.8252 -2.8436
0.6614 0.05 700 0.6531 -0.4512 -0.5525 0.6560 0.1013 -299.8257 -309.7392 -2.8005 -2.8266
0.618 0.05 800 0.6287 -0.5931 -0.7949 0.6570 0.2018 -324.0607 -323.9273 -2.7680 -2.7835
0.6067 0.06 900 0.6182 -0.4024 -0.6377 0.6735 0.2353 -308.3404 -304.8563 -2.7744 -2.7821
0.6175 0.07 1000 0.6295 -0.9965 -1.2062 0.6510 0.2097 -365.1882 -364.2672 -2.7531 -2.7655
0.7016 0.07 1100 0.5882 -0.5598 -0.9258 0.6855 0.3659 -337.1476 -320.6015 -2.6566 -2.6844
0.6085 0.08 1200 0.5893 -0.9202 -1.3935 0.6845 0.4733 -383.9212 -356.6389 -2.5379 -2.5651
0.6945 0.09 1300 0.5813 -0.7746 -1.2214 0.6855 0.4468 -366.7095 -342.0802 -2.5286 -2.5657
0.5341 0.09 1400 0.6005 -1.5045 -2.0068 0.6700 0.5023 -445.2536 -415.0722 -2.1612 -2.2154
0.5724 0.1 1500 0.5871 -1.2357 -1.9169 0.6890 0.6812 -436.2651 -388.1943 -1.9123 -1.9874
0.5714 0.1 1600 0.6159 -0.6963 -1.0648 0.6615 0.3685 -351.0529 -334.2465 -2.3794 -2.4209
0.5017 0.11 1700 0.6453 -2.5019 -2.9679 0.6430 0.4660 -541.3613 -514.8109 -2.1283 -2.1618
0.6473 0.12 1800 0.5910 -1.2621 -1.9024 0.6975 0.6403 -434.8128 -390.8305 -2.1324 -2.2148
0.6148 0.12 1900 0.5746 -0.9118 -1.5087 0.7015 0.5969 -395.4436 -355.8020 -1.9123 -2.0483
0.7404 0.13 2000 0.5779 -1.7225 -2.5148 0.7015 0.7923 -496.0523 -436.8742 -1.5445 -1.6821
0.4925 0.14 2100 0.5995 -1.8835 -2.6942 0.6915 0.8107 -513.9925 -452.9669 -1.3415 -1.4881
0.6846 0.14 2200 0.6261 -4.8072 -5.5803 0.6810 0.7732 -802.6066 -745.3393 -0.7665 -0.8833
0.4865 0.15 2300 0.7695 -6.1888 -7.5253 0.6670 1.3365 -997.1058 -883.5037 0.1325 -0.0279
0.512 0.16 2400 0.5834 -2.1074 -3.0227 0.7005 0.9153 -546.8382 -475.3564 -1.1445 -1.3016
0.5232 0.16 2500 0.5786 -2.2483 -3.2545 0.7080 1.0061 -570.0168 -489.4522 -0.9055 -1.2295
0.624 0.17 2600 0.5486 -2.3903 -3.2093 0.7210 0.8190 -565.4991 -503.6495 -0.8069 -1.1030
0.7293 0.18 2700 0.5603 -2.3227 -3.0042 0.7025 0.6816 -544.9946 -496.8855 -1.0069 -1.2653
0.4734 0.18 2800 0.5765 -2.5387 -3.5979 0.7100 1.0591 -604.3604 -518.4933 -0.6145 -0.9492
0.5551 0.19 2900 0.5749 -2.9759 -4.1407 0.7105 1.1647 -658.6375 -562.2119 -0.6867 -1.0008
0.7045 0.2 3000 0.5745 -2.7788 -3.8731 0.7210 1.0943 -631.8785 -542.4957 -1.1347 -1.4700
0.732 0.2 3100 0.5703 -3.7406 -4.8298 0.7150 1.0893 -727.5560 -638.6746 -0.8125 -1.2049
0.585 0.21 3200 0.5682 -2.3964 -3.2161 0.7050 0.8197 -566.1844 -504.2575 -1.3892 -1.6495
0.5844 0.22 3300 0.5572 -2.9653 -4.0866 0.7140 1.1213 -653.2316 -561.1476 -0.8307 -1.1810
0.4916 0.22 3400 0.5626 -3.4086 -4.4479 0.7155 1.0393 -689.3580 -605.4802 -0.9139 -1.2400
0.5492 0.23 3500 0.5706 -4.5918 -5.7581 0.7240 1.1663 -820.3834 -723.8027 -0.2622 -0.7195
0.4557 0.24 3600 0.5935 -5.0167 -6.2930 0.7045 1.2763 -873.8727 -766.2865 -0.2562 -0.7418
0.526 0.24 3700 0.5307 -3.0056 -3.9427 0.7205 0.9372 -638.8435 -565.1747 -0.7845 -1.1784
0.5895 0.25 3800 0.5401 -1.5812 -2.3022 0.7160 0.7211 -474.7949 -422.7354 -1.4099 -1.6296
0.7091 0.26 3900 0.5538 -3.7794 -4.8848 0.7200 1.1054 -733.0519 -642.5602 -0.1957 -0.6685
0.504 0.26 4000 0.5234 -1.5416 -2.3895 0.7365 0.8479 -483.5219 -418.7833 -1.4126 -1.6915
0.571 0.27 4100 0.5638 -3.0703 -4.2968 0.7255 1.2264 -674.2473 -571.6520 -0.6805 -1.0519
0.5907 0.27 4200 0.5569 -2.8129 -4.0340 0.7140 1.2211 -647.9714 -545.9053 -0.4486 -0.8569
0.4848 0.28 4300 0.5795 -3.6500 -5.0997 0.7280 1.4497 -754.5433 -629.6202 -0.3192 -0.7815
0.4623 0.29 4400 0.5920 -3.5180 -5.0207 0.7190 1.5027 -746.6427 -616.4236 -0.3936 -0.8598
0.4432 0.29 4500 0.5776 -3.9754 -5.4827 0.7340 1.5074 -792.8453 -662.1547 -0.2694 -0.7167
0.577 0.3 4600 0.5534 -3.6646 -5.0144 0.7220 1.3498 -746.0093 -631.0773 -0.6870 -1.0688
0.4871 0.31 4700 0.5627 -6.1323 -7.4013 0.7125 1.2690 -984.7041 -877.8547 -0.2580 -0.6404
0.5773 0.31 4800 0.5536 -4.0861 -5.4635 0.7245 1.3773 -790.9176 -673.2338 -0.8070 -1.1399
0.429 0.32 4900 0.6206 -4.6994 -6.5033 0.7235 1.8039 -894.9047 -734.5591 -0.4526 -0.8332
0.483 0.33 5000 0.5430 -5.3138 -6.6674 0.7245 1.3537 -911.3127 -795.9951 -0.4096 -0.7938
0.3309 0.33 5100 0.5673 -4.4644 -5.8910 0.7150 1.4266 -833.6697 -711.0602 -0.5042 -0.9408
0.5417 0.34 5200 0.5361 -3.9649 -5.2919 0.7280 1.3269 -773.7585 -661.1136 -0.7978 -1.1458
0.505 0.35 5300 0.5394 -5.1592 -6.4691 0.7340 1.3098 -891.4778 -780.5414 -0.3848 -0.7747
0.2418 0.35 5400 0.5436 -3.7243 -5.0978 0.7320 1.3735 -754.3560 -637.0532 -0.7071 -1.0946
0.5596 0.36 5500 0.5357 -4.1527 -5.4062 0.7355 1.2535 -785.1954 -679.8907 -0.8061 -1.1252
0.6177 0.37 5600 0.5369 -2.9287 -4.1640 0.7315 1.2353 -660.9726 -557.4890 -1.2997 -1.5595
0.563 0.37 5700 0.5817 -3.9459 -5.5034 0.7335 1.5575 -794.9144 -659.2140 -1.1996 -1.4800
0.4282 0.38 5800 0.5350 -2.9337 -4.2877 0.7305 1.3540 -673.3404 -557.9899 -1.3725 -1.6274
0.4219 0.39 5900 0.5515 -3.8227 -5.4619 0.7400 1.6392 -790.7645 -646.8944 -1.1562 -1.4290
0.6167 0.39 6000 0.5245 -3.2679 -4.5975 0.7375 1.3295 -704.3193 -591.4142 -1.3565 -1.5848
0.5634 0.4 6100 0.5366 -3.4000 -4.8133 0.7290 1.4133 -725.9063 -604.6245 -1.2394 -1.4960
0.4555 0.41 6200 0.5346 -2.8800 -4.3275 0.7325 1.4475 -677.3170 -552.6166 -1.3785 -1.6071
0.328 0.41 6300 0.5238 -2.5320 -4.0174 0.7300 1.4854 -646.3101 -517.8212 -1.4532 -1.6986
0.6362 0.42 6400 0.5241 -3.0294 -4.5779 0.7350 1.5485 -702.3620 -567.5569 -1.1700 -1.4758
0.3597 0.43 6500 0.5416 -3.6329 -5.3460 0.7355 1.7131 -779.1708 -627.9059 -0.9547 -1.2830
0.5852 0.43 6600 0.5490 -3.2062 -4.7795 0.7290 1.5734 -722.5227 -585.2350 -1.1807 -1.4797
0.43 0.44 6700 0.5776 -4.0288 -5.9260 0.7295 1.8972 -837.1742 -667.5021 -1.0169 -1.3083
0.4531 0.44 6800 0.5667 -3.4266 -5.1366 0.7385 1.7100 -758.2289 -607.2781 -1.2266 -1.5044
0.4527 0.45 6900 0.5578 -3.1111 -4.7331 0.7275 1.6220 -717.8849 -575.7309 -1.3552 -1.6319
0.5708 0.46 7000 0.5356 -3.2294 -4.8033 0.7355 1.5739 -724.8993 -587.5587 -1.3405 -1.6090
0.6367 0.46 7100 0.5204 -3.6636 -5.2112 0.7390 1.5476 -765.6871 -630.9789 -1.2865 -1.5484
0.7849 0.47 7200 0.5288 -4.0303 -5.6684 0.7380 1.6382 -811.4156 -667.6451 -1.1175 -1.4048
0.3462 0.48 7300 0.5395 -4.2366 -5.9634 0.7345 1.7268 -840.9079 -688.2756 -1.0407 -1.3267
0.4616 0.48 7400 0.5362 -3.5956 -5.2374 0.7355 1.6419 -768.3163 -624.1782 -1.1111 -1.4320
0.4879 0.49 7500 0.5311 -3.9628 -5.5891 0.7400 1.6263 -803.4814 -660.9017 -1.1543 -1.4181
0.6047 0.5 7600 0.5197 -3.6077 -5.1990 0.7440 1.5913 -764.4761 -625.3945 -1.2726 -1.5299
0.5471 0.5 7700 0.5191 -3.4181 -4.9614 0.7380 1.5433 -740.7103 -606.4263 -1.2776 -1.5228
0.3957 0.51 7800 0.5341 -3.5608 -5.2091 0.7355 1.6483 -765.4808 -620.6991 -1.2424 -1.5134
0.5307 0.52 7900 0.5247 -3.6480 -5.2101 0.7375 1.5621 -765.5830 -629.4217 -1.2260 -1.5021
0.6165 0.52 8000 0.5350 -4.5481 -6.1501 0.7385 1.6020 -859.5797 -719.4283 -1.0660 -1.3580
0.4843 0.53 8100 0.5416 -5.3400 -7.0079 0.7345 1.6679 -945.3573 -798.6175 -0.9235 -1.2203
0.3469 0.54 8200 0.5294 -4.3054 -5.9409 0.7360 1.6355 -838.6585 -695.1555 -1.0939 -1.4047
0.6583 0.54 8300 0.5330 -4.5942 -6.3157 0.7425 1.7215 -876.1429 -724.0405 -0.9177 -1.2946
0.3581 0.55 8400 0.5290 -4.4272 -6.1139 0.7430 1.6867 -855.9659 -707.3421 -1.0403 -1.3877
0.4143 0.56 8500 0.5271 -4.2079 -5.9375 0.7505 1.7296 -838.3192 -685.4116 -0.9933 -1.3601
0.6205 0.56 8600 0.5300 -3.9823 -5.7856 0.7490 1.8033 -823.1313 -662.8466 -1.0674 -1.4290
0.5613 0.57 8700 0.5370 -3.6486 -5.4644 0.7405 1.8158 -791.0135 -629.4801 -1.0772 -1.4600
0.3026 0.58 8800 0.5405 -4.1182 -5.9998 0.7480 1.8816 -844.5538 -676.4411 -0.9434 -1.3583
0.6241 0.58 8900 0.5261 -3.5431 -5.2430 0.7415 1.6999 -768.8730 -618.9297 -1.0692 -1.4737
0.5426 0.59 9000 0.5123 -3.4277 -5.0588 0.7415 1.6311 -750.4479 -607.3850 -1.0844 -1.4735
0.7459 0.6 9100 0.5097 -3.6073 -5.1879 0.7470 1.5806 -763.3654 -625.3505 -1.0356 -1.4295
0.4619 0.6 9200 0.5202 -4.1917 -5.8950 0.7415 1.7033 -834.0685 -683.7893 -0.9207 -1.3270
0.3541 0.61 9300 0.5061 -3.4397 -4.9850 0.7480 1.5453 -743.0750 -608.5919 -1.1180 -1.5005
0.4268 0.62 9400 0.5187 -3.9580 -5.7277 0.7465 1.7697 -817.3372 -660.4188 -0.9943 -1.4003
0.6392 0.62 9500 0.5298 -4.1845 -6.0696 0.7385 1.8851 -851.5309 -683.0696 -0.8994 -1.3308
0.6151 0.63 9600 0.5237 -3.8920 -5.7099 0.7440 1.8179 -815.5630 -653.8219 -0.9559 -1.3883
0.4596 0.63 9700 0.5333 -3.7944 -5.6758 0.7470 1.8813 -812.1490 -644.0645 -1.0611 -1.4511
0.6714 0.64 9800 0.5592 -4.4270 -6.5772 0.7385 2.1501 -902.2877 -707.3235 -0.9338 -1.3445
0.6304 0.65 9900 0.5398 -4.4397 -6.4394 0.7410 1.9997 -888.5164 -708.5909 -0.9850 -1.3756
0.463 0.65 10000 0.5291 -4.2047 -6.1080 0.7470 1.9033 -855.3674 -685.0887 -1.0414 -1.4192
0.4455 0.66 10100 0.5431 -4.5725 -6.5907 0.7450 2.0182 -903.6422 -721.8721 -0.9830 -1.3678
0.3541 0.67 10200 0.5516 -4.8037 -6.9155 0.7455 2.1118 -936.1205 -744.9925 -0.9014 -1.3059
0.3868 0.67 10300 0.5256 -4.1702 -6.0539 0.7485 1.8836 -849.9585 -681.6424 -1.0641 -1.4424
0.6851 0.68 10400 0.5218 -4.0721 -5.9151 0.7480 1.8430 -836.0790 -671.8286 -1.1069 -1.4800
0.619 0.69 10500 0.5219 -3.9593 -5.7760 0.7475 1.8167 -822.1694 -660.5464 -1.1250 -1.5018
0.6235 0.69 10600 0.5139 -3.6928 -5.4123 0.7460 1.7195 -785.8032 -633.8964 -1.2033 -1.5598
0.3952 0.7 10700 0.5147 -3.9589 -5.7048 0.7525 1.7459 -815.0552 -660.5131 -1.1463 -1.5122
0.4521 0.71 10800 0.5215 -4.2859 -6.1109 0.7490 1.8250 -855.6591 -693.2052 -1.0765 -1.4514
0.7094 0.71 10900 0.5195 -4.2340 -6.0437 0.7495 1.8097 -848.9450 -688.0204 -1.0678 -1.4484
0.6759 0.72 11000 0.5184 -4.1690 -5.9809 0.7485 1.8119 -842.6664 -681.5213 -1.0737 -1.4573
0.4752 0.73 11100 0.5154 -3.8737 -5.6279 0.7465 1.7542 -807.3627 -651.9897 -1.1638 -1.5326
0.4382 0.73 11200 0.5193 -3.9946 -5.7959 0.75 1.8013 -824.1631 -664.0820 -1.1533 -1.5243
0.5666 0.74 11300 0.5179 -3.9724 -5.7729 0.7510 1.8004 -821.8571 -661.8636 -1.1489 -1.5188
0.6254 0.75 11400 0.5160 -3.8732 -5.6427 0.7510 1.7695 -808.8420 -651.9423 -1.1772 -1.5401
0.5912 0.75 11500 0.5173 -3.9316 -5.7185 0.75 1.7868 -816.4195 -657.7830 -1.1612 -1.5292
0.5279 0.76 11600 0.5231 -4.1317 -5.9844 0.7470 1.8528 -843.0165 -677.7863 -1.1125 -1.4905
0.5654 0.77 11700 0.5235 -4.1005 -5.9425 0.7450 1.8420 -838.8231 -674.6689 -1.1325 -1.5063
0.6573 0.77 11800 0.5228 -4.1344 -5.9811 0.7455 1.8467 -842.6800 -678.0629 -1.1285 -1.5005
0.4045 0.78 11900 0.5222 -4.1607 -6.0027 0.7465 1.8420 -844.8414 -680.6879 -1.1271 -1.4978
0.436 0.79 12000 0.5193 -4.1188 -5.9342 0.7455 1.8154 -837.9908 -676.4965 -1.1403 -1.5061
0.519 0.79 12100 0.5164 -4.0229 -5.8065 0.7495 1.7836 -825.2211 -666.9062 -1.1552 -1.5189
0.5342 0.8 12200 0.5155 -3.9832 -5.7666 0.7485 1.7834 -821.2302 -662.9399 -1.1597 -1.5231
0.3715 0.8 12300 0.5171 -4.0251 -5.8295 0.7465 1.8044 -827.5244 -667.1307 -1.1525 -1.5152
0.7344 0.81 12400 0.5187 -4.1262 -5.9517 0.7470 1.8255 -839.7450 -677.2386 -1.1281 -1.4944
0.4667 0.82 12500 0.5171 -4.0972 -5.9057 0.7475 1.8085 -835.1400 -674.3381 -1.1316 -1.4972
0.5658 0.82 12600 0.5172 -4.1066 -5.9177 0.7470 1.8111 -836.3404 -675.2822 -1.1301 -1.4965
0.6554 0.83 12700 0.5167 -4.1131 -5.9204 0.7490 1.8073 -836.6075 -675.9286 -1.1283 -1.4943
0.5481 0.84 12800 0.5154 -4.0796 -5.8674 0.7490 1.7878 -831.3082 -672.5789 -1.1394 -1.5030
0.3902 0.84 12900 0.5155 -4.0744 -5.8664 0.7485 1.7920 -831.2067 -672.0550 -1.1385 -1.5025
0.3801 0.85 13000 0.5155 -4.0583 -5.8464 0.7480 1.7881 -829.2069 -670.4493 -1.1422 -1.5056
0.6991 0.86 13100 0.5154 -4.0516 -5.8412 0.7495 1.7896 -828.6917 -669.7778 -1.1435 -1.5069
0.472 0.86 13200 0.5151 -4.0533 -5.8454 0.7485 1.7921 -829.1138 -669.9543 -1.1407 -1.5046
0.3055 0.87 13300 0.5151 -4.0433 -5.8344 0.7495 1.7910 -828.0081 -668.9514 -1.1421 -1.5057
0.6737 0.88 13400 0.5151 -4.0448 -5.8347 0.7505 1.7898 -828.0372 -669.1003 -1.1420 -1.5060
0.3819 0.88 13500 0.5151 -4.0549 -5.8467 0.7490 1.7918 -829.2462 -670.1140 -1.1399 -1.5038
0.8034 0.89 13600 0.5154 -4.0637 -5.8586 0.7490 1.7949 -830.4301 -670.9915 -1.1367 -1.5018
0.4371 0.9 13700 0.5157 -4.0796 -5.8779 0.7495 1.7983 -832.3608 -672.5767 -1.1338 -1.4991
0.3428 0.9 13800 0.5155 -4.0754 -5.8733 0.7495 1.7979 -831.8970 -672.1581 -1.1347 -1.5001
0.5029 0.91 13900 0.5156 -4.0734 -5.8709 0.7495 1.7975 -831.6616 -671.9635 -1.1351 -1.5004
0.5905 0.92 14000 0.5155 -4.0760 -5.8741 0.7525 1.7981 -831.9777 -672.2200 -1.1345 -1.4997
0.3965 0.92 14100 0.5157 -4.0782 -5.8761 0.7505 1.7979 -832.1840 -672.4373 -1.1343 -1.4994
0.4038 0.93 14200 0.5156 -4.0795 -5.8779 0.7490 1.7984 -832.3639 -672.5670 -1.1340 -1.4994
0.4043 0.94 14300 0.5156 -4.0807 -5.8792 0.7505 1.7985 -832.4966 -672.6912 -1.1337 -1.4988
0.5662 0.94 14400 0.5155 -4.0814 -5.8804 0.7490 1.7991 -832.6164 -672.7547 -1.1335 -1.4987
0.4828 0.95 14500 0.5157 -4.0810 -5.8796 0.7490 1.7986 -832.5297 -672.7201 -1.1340 -1.4990
0.5555 0.96 14600 0.5157 -4.0805 -5.8787 0.7490 1.7982 -832.4430 -672.6707 -1.1335 -1.4990
0.704 0.96 14700 0.5155 -4.0802 -5.8790 0.7505 1.7988 -832.4694 -672.6378 -1.1338 -1.4989
0.7164 0.97 14800 0.5158 -4.0806 -5.8795 0.7490 1.7990 -832.5262 -672.6747 -1.1340 -1.4991
0.3263 0.97 14900 0.5155 -4.0795 -5.8783 0.7510 1.7988 -832.3969 -672.5685 -1.1339 -1.4994
0.3809 0.98 15000 0.5155 -4.0804 -5.8793 0.7490 1.7989 -832.5026 -672.6627 -1.1337 -1.4992
0.4781 0.99 15100 0.5158 -4.0809 -5.8789 0.7495 1.7980 -832.4585 -672.7083 -1.1336 -1.4991
0.5115 0.99 15200 0.5159 -4.0804 -5.8780 0.7475 1.7976 -832.3694 -672.6617 -1.1337 -1.4991

Framework versions

  • PEFT 0.7.1
  • Transformers 4.38.2
  • Pytorch 2.1.2
  • Datasets 2.14.6
  • Tokenizers 0.15.2