--- license: apache-2.0 library_name: peft tags: - trl - sft - generated_from_trainer base_model: TheBloke/Mistral-7B-v0.1-GPTQ model-index: - name: mistral-rand-300k results: [] --- # mistral-rand-300k This model is a fine-tuned version of [TheBloke/Mistral-7B-v0.1-GPTQ](https://huggingface.co/TheBloke/Mistral-7B-v0.1-GPTQ) on the None dataset. It achieves the following results on the evaluation set: - Loss: 0.3545 ## 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: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - training_steps: 12451 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:-----:|:---------------:| | 1.6373 | 0.01 | 50 | 0.8699 | | 0.7891 | 0.02 | 100 | 0.7454 | | 0.7135 | 0.02 | 150 | 0.6984 | | 0.6931 | 0.03 | 200 | 0.6722 | | 0.6592 | 0.08 | 250 | 0.6407 | | 0.634 | 0.1 | 300 | 0.6212 | | 0.61 | 0.11 | 350 | 0.6015 | | 0.5961 | 0.13 | 400 | 0.5892 | | 0.5795 | 0.14 | 450 | 0.5773 | | 0.5765 | 0.16 | 500 | 0.5683 | | 0.5619 | 0.18 | 550 | 0.5612 | | 0.5558 | 0.19 | 600 | 0.5519 | | 0.5572 | 0.21 | 650 | 0.5438 | | 0.5467 | 0.22 | 700 | 0.5348 | | 0.5303 | 0.24 | 750 | 0.5248 | | 0.5307 | 0.26 | 800 | 0.5135 | | 0.508 | 0.27 | 850 | 0.5036 | | 0.5087 | 0.29 | 900 | 0.5009 | | 0.5073 | 0.3 | 950 | 0.4945 | | 0.5078 | 0.32 | 1000 | 0.4916 | | 0.5089 | 0.34 | 1050 | 0.4893 | | 0.4829 | 0.35 | 1100 | 0.4862 | | 0.4872 | 0.37 | 1150 | 0.4832 | | 0.4845 | 0.39 | 1200 | 0.4803 | | 0.4993 | 0.4 | 1250 | 0.4774 | | 0.475 | 0.42 | 1300 | 0.4746 | | 0.4836 | 0.43 | 1350 | 0.4735 | | 0.4748 | 0.45 | 1400 | 0.4708 | | 0.4809 | 0.47 | 1450 | 0.4693 | | 0.4755 | 0.48 | 1500 | 0.4668 | | 0.4679 | 0.5 | 1550 | 0.4644 | | 0.4685 | 0.51 | 1600 | 0.4622 | | 0.4706 | 0.53 | 1650 | 0.4611 | | 0.4673 | 0.55 | 1700 | 0.4605 | | 0.4539 | 0.56 | 1750 | 0.4577 | | 0.4501 | 0.58 | 1800 | 0.4560 | | 0.4638 | 0.59 | 1850 | 0.4542 | | 0.4663 | 0.61 | 1900 | 0.4521 | | 0.4638 | 0.63 | 1950 | 0.4509 | | 0.4562 | 0.64 | 2000 | 0.4501 | | 0.4535 | 0.66 | 2050 | 0.4496 | | 0.4548 | 0.67 | 2100 | 0.4470 | | 0.442 | 0.69 | 2150 | 0.4453 | | 0.4543 | 0.71 | 2200 | 0.4449 | | 0.4435 | 0.72 | 2250 | 0.4428 | | 0.4633 | 0.74 | 2300 | 0.4418 | | 0.4438 | 0.75 | 2350 | 0.4416 | | 0.4443 | 0.77 | 2400 | 0.4392 | | 0.4424 | 0.79 | 2450 | 0.4386 | | 0.4341 | 0.8 | 2500 | 0.4367 | | 0.4329 | 0.82 | 2550 | 0.4353 | | 0.4356 | 0.83 | 2600 | 0.4349 | | 0.4384 | 0.85 | 2650 | 0.4351 | | 0.4327 | 0.87 | 2700 | 0.4321 | | 0.4356 | 0.88 | 2750 | 0.4323 | | 0.4428 | 0.9 | 2800 | 0.4310 | | 0.4358 | 0.91 | 2850 | 0.4304 | | 0.4322 | 0.93 | 2900 | 0.4293 | | 0.4336 | 0.95 | 2950 | 0.4280 | | 0.4296 | 0.96 | 3000 | 0.4269 | | 0.4365 | 0.98 | 3050 | 0.4267 | | 0.4313 | 0.99 | 3100 | 0.4250 | | 0.4256 | 1.01 | 3150 | 0.4251 | | 0.4258 | 1.03 | 3200 | 0.4241 | | 0.4245 | 1.04 | 3250 | 0.4225 | | 0.4161 | 1.06 | 3300 | 0.4223 | | 0.4228 | 1.07 | 3350 | 0.4215 | | 0.4194 | 1.09 | 3400 | 0.4205 | | 0.4331 | 1.11 | 3450 | 0.4193 | | 0.4246 | 1.12 | 3500 | 0.4191 | | 0.4246 | 1.14 | 3550 | 0.4173 | | 0.4229 | 1.16 | 3600 | 0.4175 | | 0.4128 | 1.17 | 3650 | 0.4160 | | 0.4189 | 1.19 | 3700 | 0.4156 | | 0.4154 | 1.2 | 3750 | 0.4148 | | 0.4263 | 1.22 | 3800 | 0.4140 | | 0.4124 | 1.24 | 3850 | 0.4139 | | 0.4201 | 1.25 | 3900 | 0.4132 | | 0.4224 | 1.27 | 3950 | 0.4122 | | 0.4114 | 1.28 | 4000 | 0.4122 | | 0.4169 | 1.3 | 4050 | 0.4112 | | 0.4167 | 1.32 | 4100 | 0.4107 | | 0.4019 | 1.33 | 4150 | 0.4095 | | 0.4142 | 1.35 | 4200 | 0.4087 | | 0.4086 | 1.36 | 4250 | 0.4080 | | 0.406 | 1.38 | 4300 | 0.4075 | | 0.4091 | 1.4 | 4350 | 0.4069 | | 0.4149 | 1.41 | 4400 | 0.4062 | | 0.4078 | 1.43 | 4450 | 0.4054 | | 0.3997 | 1.44 | 4500 | 0.4052 | | 0.3985 | 1.46 | 4550 | 0.4040 | | 0.4035 | 1.48 | 4600 | 0.4035 | | 0.3982 | 1.49 | 4650 | 0.4025 | | 0.4018 | 1.51 | 4700 | 0.4030 | | 0.4078 | 1.52 | 4750 | 0.4021 | | 0.3991 | 1.54 | 4800 | 0.4010 | | 0.4033 | 1.56 | 4850 | 0.4003 | | 0.3964 | 1.57 | 4900 | 0.4005 | | 0.3965 | 1.59 | 4950 | 0.3993 | | 0.407 | 1.6 | 5000 | 0.3994 | | 0.4036 | 1.62 | 5050 | 0.3983 | | 0.4063 | 1.64 | 5100 | 0.3980 | | 0.3857 | 1.65 | 5150 | 0.3973 | | 0.3949 | 1.67 | 5200 | 0.3973 | | 0.3872 | 1.68 | 5250 | 0.3965 | | 0.393 | 1.7 | 5300 | 0.3959 | | 0.3891 | 1.72 | 5350 | 0.3955 | | 0.3903 | 1.73 | 5400 | 0.3950 | | 0.3941 | 1.75 | 5450 | 0.3947 | | 0.3879 | 1.76 | 5500 | 0.3941 | | 0.4016 | 1.78 | 5550 | 0.3937 | | 0.3936 | 1.8 | 5600 | 0.3929 | | 0.4005 | 1.81 | 5650 | 0.3932 | | 0.3939 | 1.83 | 5700 | 0.3923 | | 0.4032 | 1.85 | 5750 | 0.3921 | | 0.3921 | 1.86 | 5800 | 0.3921 | | 0.3903 | 1.88 | 5850 | 0.3905 | | 0.3983 | 1.89 | 5900 | 0.3910 | | 0.3806 | 1.91 | 5950 | 0.3897 | | 0.3964 | 1.93 | 6000 | 0.3906 | | 0.3866 | 1.94 | 6050 | 0.3890 | | 0.3882 | 1.96 | 6100 | 0.3888 | | 0.3835 | 1.97 | 6150 | 0.3885 | | 0.3921 | 1.99 | 6200 | 0.3875 | | 0.388 | 2.01 | 6250 | 0.3878 | | 0.3829 | 2.02 | 6300 | 0.3872 | | 0.3814 | 2.04 | 6350 | 0.3867 | | 0.3818 | 2.05 | 6400 | 0.3862 | | 0.3802 | 2.07 | 6450 | 0.3860 | | 0.3739 | 2.09 | 6500 | 0.3853 | | 0.3771 | 2.1 | 6550 | 0.3852 | | 0.3732 | 2.12 | 6600 | 0.3846 | | 0.385 | 2.13 | 6650 | 0.3849 | | 0.3767 | 2.15 | 6700 | 0.3833 | | 0.3802 | 2.17 | 6750 | 0.3836 | | 0.3844 | 2.18 | 6800 | 0.3828 | | 0.3761 | 2.2 | 6850 | 0.3826 | | 0.3765 | 2.21 | 6900 | 0.3826 | | 0.3787 | 2.23 | 6950 | 0.3825 | | 0.378 | 2.25 | 7000 | 0.3815 | | 0.3792 | 2.26 | 7050 | 0.3815 | | 0.3908 | 2.28 | 7100 | 0.3811 | | 0.3757 | 2.29 | 7150 | 0.3810 | | 0.376 | 2.31 | 7200 | 0.3804 | | 0.3785 | 2.33 | 7250 | 0.3805 | | 0.3744 | 2.34 | 7300 | 0.3797 | | 0.3984 | 2.36 | 7350 | 0.3791 | | 0.3833 | 2.37 | 7400 | 0.3792 | | 0.3808 | 2.39 | 7450 | 0.3785 | | 0.3803 | 2.41 | 7500 | 0.3786 | | 0.3828 | 2.42 | 7550 | 0.3778 | | 0.3697 | 2.44 | 7600 | 0.3780 | | 0.3692 | 2.45 | 7650 | 0.3763 | | 0.3808 | 2.47 | 7700 | 0.3769 | | 0.3764 | 2.49 | 7750 | 0.3763 | | 0.3865 | 2.5 | 7800 | 0.3763 | | 0.375 | 2.52 | 7850 | 0.3760 | | 0.368 | 2.53 | 7900 | 0.3755 | | 0.3632 | 2.55 | 7950 | 0.3757 | | 0.3792 | 2.57 | 8000 | 0.3758 | | 0.374 | 2.58 | 8050 | 0.3748 | | 0.3689 | 2.6 | 8100 | 0.3741 | | 0.3843 | 2.62 | 8150 | 0.3741 | | 0.3669 | 2.63 | 8200 | 0.3739 | | 0.368 | 2.65 | 8250 | 0.3732 | | 0.3726 | 2.66 | 8300 | 0.3734 | | 0.3653 | 2.68 | 8350 | 0.3728 | | 0.3777 | 2.7 | 8400 | 0.3732 | | 0.3625 | 2.71 | 8450 | 0.3724 | | 0.3749 | 2.73 | 8500 | 0.3716 | | 0.3708 | 2.74 | 8550 | 0.3725 | | 0.3618 | 2.76 | 8600 | 0.3711 | | 0.3659 | 2.78 | 8650 | 0.3714 | | 0.3661 | 2.79 | 8700 | 0.3711 | | 0.3771 | 2.81 | 8750 | 0.3714 | | 0.3637 | 2.82 | 8800 | 0.3704 | | 0.3768 | 2.84 | 8850 | 0.3700 | | 0.3722 | 2.86 | 8900 | 0.3701 | | 0.366 | 2.87 | 8950 | 0.3693 | | 0.3716 | 2.89 | 9000 | 0.3690 | | 0.3622 | 2.9 | 9050 | 0.3688 | | 0.3594 | 2.92 | 9100 | 0.3682 | | 0.368 | 2.94 | 9150 | 0.3680 | | 0.3538 | 2.95 | 9200 | 0.3678 | | 0.3578 | 2.97 | 9250 | 0.3676 | | 0.3685 | 2.98 | 9300 | 0.3679 | | 0.3631 | 3.0 | 9350 | 0.3674 | | 0.3645 | 3.02 | 9400 | 0.3665 | | 0.3654 | 3.03 | 9450 | 0.3671 | | 0.3502 | 3.05 | 9500 | 0.3662 | | 0.356 | 3.06 | 9550 | 0.3665 | | 0.3642 | 3.08 | 9600 | 0.3662 | | 0.3688 | 3.1 | 9650 | 0.3659 | | 0.3514 | 3.11 | 9700 | 0.3655 | | 0.3463 | 3.13 | 9750 | 0.3656 | | 0.3517 | 3.14 | 9800 | 0.3651 | | 0.3666 | 3.16 | 9850 | 0.3650 | | 0.3617 | 3.18 | 9900 | 0.3660 | | 0.3452 | 3.19 | 9950 | 0.3649 | | 0.3591 | 3.21 | 10000 | 0.3647 | | 0.3509 | 3.22 | 10050 | 0.3643 | | 0.3618 | 3.24 | 10100 | 0.3641 | | 0.3571 | 3.26 | 10150 | 0.3640 | | 0.3587 | 3.27 | 10200 | 0.3633 | | 0.3664 | 3.29 | 10250 | 0.3637 | | 0.3502 | 3.3 | 10300 | 0.3633 | | 0.373 | 3.32 | 10350 | 0.3626 | | 0.3623 | 3.34 | 10400 | 0.3622 | | 0.3554 | 3.35 | 10450 | 0.3624 | | 0.3511 | 3.37 | 10500 | 0.3622 | | 0.3534 | 3.39 | 10550 | 0.3626 | | 0.3473 | 3.4 | 10600 | 0.3620 | | 0.3563 | 3.42 | 10650 | 0.3618 | | 0.3612 | 3.43 | 10700 | 0.3614 | | 0.3587 | 3.45 | 10750 | 0.3610 | | 0.3521 | 3.47 | 10800 | 0.3609 | | 0.3443 | 3.48 | 10850 | 0.3610 | | 0.3615 | 3.5 | 10900 | 0.3608 | | 0.3589 | 3.51 | 10950 | 0.3609 | | 0.364 | 3.53 | 11000 | 0.3598 | | 0.3498 | 3.55 | 11050 | 0.3600 | | 0.3541 | 3.56 | 11100 | 0.3597 | | 0.3555 | 3.58 | 11150 | 0.3594 | | 0.3491 | 3.59 | 11200 | 0.3596 | | 0.3498 | 3.61 | 11250 | 0.3589 | | 0.3484 | 3.63 | 11300 | 0.3590 | | 0.3483 | 3.64 | 11350 | 0.3586 | | 0.3533 | 3.66 | 11400 | 0.3580 | | 0.3479 | 3.67 | 11450 | 0.3589 | | 0.3539 | 3.69 | 11500 | 0.3580 | | 0.3507 | 3.71 | 11550 | 0.3582 | | 0.3534 | 3.72 | 11600 | 0.3579 | | 0.3559 | 3.74 | 11650 | 0.3575 | | 0.3477 | 3.75 | 11700 | 0.3577 | | 0.3501 | 3.77 | 11750 | 0.3574 | | 0.3491 | 3.79 | 11800 | 0.3569 | | 0.3661 | 3.8 | 11850 | 0.3569 | | 0.3455 | 3.82 | 11900 | 0.3568 | | 0.3522 | 3.83 | 11950 | 0.3564 | | 0.3532 | 3.85 | 12000 | 0.3562 | | 0.3513 | 3.87 | 12050 | 0.3559 | | 0.3527 | 3.88 | 12100 | 0.3561 | | 0.3575 | 3.9 | 12150 | 0.3556 | | 0.3403 | 3.92 | 12200 | 0.3550 | | 0.3495 | 3.93 | 12250 | 0.3554 | | 0.3514 | 3.95 | 12300 | 0.3548 | | 0.3556 | 3.96 | 12350 | 0.3547 | | 0.3549 | 3.98 | 12400 | 0.3545 | | 0.3454 | 2.0 | 12450 | 0.3545 | ### Framework versions - PEFT 0.7.1 - Transformers 4.36.2 - Pytorch 2.1.2+cu121 - Datasets 2.16.1 - Tokenizers 0.15.0