tinyllama-1.1b-sum-simpo_beta2.0_gamma1.6_LR5e-8_3epochs

This model is a fine-tuned version of martimfasantos/tinyllama-1.1b-sum-sft-full_old on the openai/summarize_from_feedback dataset. It achieves the following results on the evaluation set:

  • Loss: 1.5991
  • Rewards/chosen: -4.1926
  • Rewards/rejected: -4.6137
  • Rewards/accuracies: 0.6231
  • Rewards/margins: 0.4211
  • Logps/rejected: -2.3069
  • Logps/chosen: -2.0963
  • Logits/rejected: -3.3338
  • Logits/chosen: -3.3372

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-08
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • distributed_type: multi-GPU
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss Rewards/chosen Rewards/rejected Rewards/accuracies Rewards/margins Logps/rejected Logps/chosen Logits/rejected Logits/chosen
1.6609 0.0172 100 1.6731 -2.9616 -3.2164 0.5960 0.2548 -1.6082 -1.4808 -3.6346 -3.6387
1.6971 0.0345 200 1.6731 -2.9618 -3.2164 0.5953 0.2546 -1.6082 -1.4809 -3.6367 -3.6408
1.6134 0.0517 300 1.6730 -2.9617 -3.2164 0.5950 0.2548 -1.6082 -1.4808 -3.6205 -3.6247
1.6601 0.0689 400 1.6731 -2.9617 -3.2165 0.5953 0.2548 -1.6083 -1.4809 -3.6372 -3.6413
1.7377 0.0861 500 1.6731 -2.9620 -3.2165 0.5941 0.2545 -1.6082 -1.4810 -3.6401 -3.6442
1.5377 0.1034 600 1.6726 -2.9613 -3.2165 0.5962 0.2552 -1.6083 -1.4807 -3.6397 -3.6438
1.8023 0.1206 700 1.6730 -2.9612 -3.2160 0.5955 0.2548 -1.6080 -1.4806 -3.6470 -3.6510
1.6595 0.1378 800 1.6727 -2.9617 -3.2169 0.5957 0.2552 -1.6085 -1.4809 -3.6409 -3.6450
1.8292 0.1551 900 1.6727 -2.9616 -3.2167 0.5953 0.2552 -1.6084 -1.4808 -3.6317 -3.6358
1.8326 0.1723 1000 1.6722 -2.9608 -3.2165 0.5943 0.2556 -1.6082 -1.4804 -3.6329 -3.6370
1.6767 0.1895 1100 1.6724 -2.9613 -3.2168 0.5948 0.2555 -1.6084 -1.4806 -3.6335 -3.6376
1.711 0.2068 1200 1.6725 -2.9615 -3.2169 0.5950 0.2554 -1.6084 -1.4807 -3.6391 -3.6431
1.6366 0.2240 1300 1.6723 -2.9612 -3.2167 0.5955 0.2555 -1.6083 -1.4806 -3.6354 -3.6394
1.7495 0.2412 1400 1.6719 -2.9613 -3.2174 0.5946 0.2561 -1.6087 -1.4806 -3.6341 -3.6381
1.7423 0.2584 1500 1.6714 -2.9610 -3.2178 0.5950 0.2569 -1.6089 -1.4805 -3.6286 -3.6326
1.6612 0.2757 1600 1.6713 -2.9614 -3.2185 0.5943 0.2571 -1.6093 -1.4807 -3.6393 -3.6433
1.6808 0.2929 1700 1.6704 -2.9613 -3.2194 0.5960 0.2581 -1.6097 -1.4806 -3.6301 -3.6342
1.6208 0.3101 1800 1.6702 -2.9615 -3.2196 0.5946 0.2581 -1.6098 -1.4808 -3.6222 -3.6263
1.82 0.3274 1900 1.6692 -2.9619 -3.2219 0.5946 0.2600 -1.6110 -1.4810 -3.6237 -3.6277
1.6569 0.3446 2000 1.6686 -2.9633 -3.2240 0.5941 0.2607 -1.6120 -1.4816 -3.6285 -3.6325
1.8076 0.3618 2100 1.6682 -2.9644 -3.2256 0.5946 0.2612 -1.6128 -1.4822 -3.6268 -3.6308
1.6012 0.3790 2200 1.6676 -2.9655 -3.2275 0.5941 0.2620 -1.6137 -1.4827 -3.6245 -3.6285
1.6718 0.3963 2300 1.6663 -2.9674 -3.2314 0.5964 0.2640 -1.6157 -1.4837 -3.6148 -3.6189
1.5252 0.4135 2400 1.6658 -2.9706 -3.2353 0.6018 0.2647 -1.6176 -1.4853 -3.6208 -3.6248
1.7441 0.4307 2500 1.6648 -2.9732 -3.2391 0.6022 0.2659 -1.6195 -1.4866 -3.6082 -3.6122
1.7247 0.4480 2600 1.6640 -2.9758 -3.2426 0.6043 0.2669 -1.6213 -1.4879 -3.6028 -3.6068
1.5961 0.4652 2700 1.6629 -2.9795 -3.2484 0.6055 0.2689 -1.6242 -1.4898 -3.6017 -3.6057
1.8025 0.4824 2800 1.6617 -2.9837 -3.2540 0.6059 0.2703 -1.6270 -1.4918 -3.6035 -3.6075
1.8171 0.4997 2900 1.6608 -2.9893 -3.2608 0.6057 0.2715 -1.6304 -1.4947 -3.6063 -3.6102
1.7331 0.5169 3000 1.6599 -2.9948 -3.2675 0.6059 0.2727 -1.6337 -1.4974 -3.6027 -3.6066
1.6335 0.5341 3100 1.6588 -2.9974 -3.2719 0.6073 0.2745 -1.6359 -1.4987 -3.6030 -3.6069
1.8053 0.5513 3200 1.6578 -3.0040 -3.2800 0.6132 0.2760 -1.6400 -1.5020 -3.5859 -3.5898
1.7128 0.5686 3300 1.6569 -3.0123 -3.2894 0.6143 0.2771 -1.6447 -1.5061 -3.5933 -3.5971
1.5567 0.5858 3400 1.6554 -3.0215 -3.3012 0.6141 0.2797 -1.6506 -1.5108 -3.5884 -3.5923
1.6557 0.6030 3500 1.6545 -3.0307 -3.3121 0.6150 0.2814 -1.6561 -1.5153 -3.5780 -3.5820
1.7443 0.6203 3600 1.6533 -3.0435 -3.3271 0.6143 0.2835 -1.6635 -1.5218 -3.5697 -3.5737
1.4954 0.6375 3700 1.6515 -3.0535 -3.3399 0.6132 0.2863 -1.6699 -1.5268 -3.5817 -3.5856
1.7495 0.6547 3800 1.6500 -3.0685 -3.3571 0.6150 0.2886 -1.6785 -1.5342 -3.5748 -3.5787
1.5238 0.6720 3900 1.6493 -3.0836 -3.3737 0.6120 0.2901 -1.6869 -1.5418 -3.5617 -3.5656
1.7001 0.6892 4000 1.6481 -3.1041 -3.3965 0.6053 0.2924 -1.6983 -1.5521 -3.5621 -3.5659
1.5842 0.7064 4100 1.6466 -3.1233 -3.4188 0.6085 0.2954 -1.7094 -1.5617 -3.5465 -3.5504
1.7415 0.7236 4200 1.6453 -3.1419 -3.4399 0.6090 0.2980 -1.7200 -1.5709 -3.5474 -3.5512
1.6328 0.7409 4300 1.6435 -3.1587 -3.4597 0.6092 0.3010 -1.7299 -1.5793 -3.5451 -3.5489
1.6841 0.7581 4400 1.6433 -3.1831 -3.4855 0.6122 0.3024 -1.7427 -1.5915 -3.5446 -3.5485
1.7753 0.7753 4500 1.6420 -3.2122 -3.5175 0.6138 0.3053 -1.7588 -1.6061 -3.5300 -3.5339
1.5101 0.7926 4600 1.6403 -3.2378 -3.5467 0.6150 0.3089 -1.7734 -1.6189 -3.5383 -3.5421
1.5603 0.8098 4700 1.6389 -3.2699 -3.5819 0.6169 0.3120 -1.7910 -1.6350 -3.5395 -3.5432
1.6925 0.8270 4800 1.6371 -3.3015 -3.6169 0.6197 0.3154 -1.8085 -1.6508 -3.5388 -3.5425
1.55 0.8442 4900 1.6352 -3.3371 -3.6557 0.6204 0.3187 -1.8279 -1.6685 -3.5296 -3.5334
1.7547 0.8615 5000 1.6344 -3.3516 -3.6717 0.6215 0.3200 -1.8358 -1.6758 -3.5221 -3.5259
1.5639 0.8787 5100 1.6325 -3.3917 -3.7152 0.6215 0.3234 -1.8576 -1.6959 -3.5201 -3.5238
1.5202 0.8959 5200 1.6311 -3.4276 -3.7534 0.6211 0.3258 -1.8767 -1.7138 -3.5244 -3.5282
1.5903 0.9132 5300 1.6297 -3.4486 -3.7771 0.6215 0.3286 -1.8886 -1.7243 -3.5069 -3.5107
1.4759 0.9304 5400 1.6283 -3.4696 -3.8007 0.6229 0.3311 -1.9004 -1.7348 -3.5057 -3.5095
1.5141 0.9476 5500 1.6276 -3.4762 -3.8091 0.6222 0.3328 -1.9045 -1.7381 -3.5203 -3.5240
1.6434 0.9649 5600 1.6268 -3.4817 -3.8157 0.6234 0.3340 -1.9079 -1.7408 -3.5045 -3.5082
1.6866 0.9821 5700 1.6257 -3.4974 -3.8333 0.6241 0.3359 -1.9167 -1.7487 -3.4916 -3.4954
1.4625 0.9993 5800 1.6247 -3.5213 -3.8599 0.6241 0.3386 -1.9300 -1.7606 -3.4941 -3.4979
1.5559 1.0165 5900 1.6238 -3.5308 -3.8712 0.6243 0.3404 -1.9356 -1.7654 -3.4910 -3.4947
1.5296 1.0338 6000 1.6234 -3.5406 -3.8820 0.6241 0.3414 -1.9410 -1.7703 -3.4929 -3.4966
1.7383 1.0510 6100 1.6228 -3.5586 -3.9015 0.6208 0.3429 -1.9508 -1.7793 -3.4874 -3.4910
1.5491 1.0682 6200 1.6215 -3.5797 -3.9249 0.6208 0.3452 -1.9624 -1.7899 -3.4773 -3.4810
1.5498 1.0855 6300 1.6214 -3.5931 -3.9393 0.6204 0.3462 -1.9696 -1.7965 -3.4747 -3.4784
1.613 1.1027 6400 1.6210 -3.6015 -3.9492 0.6206 0.3478 -1.9746 -1.8007 -3.4789 -3.4826
1.7929 1.1199 6500 1.6200 -3.6169 -3.9669 0.6220 0.3500 -1.9835 -1.8085 -3.4716 -3.4753
1.7372 1.1371 6600 1.6199 -3.6260 -3.9776 0.6215 0.3516 -1.9888 -1.8130 -3.4685 -3.4721
1.5748 1.1544 6700 1.6198 -3.6291 -3.9815 0.6227 0.3524 -1.9908 -1.8145 -3.4609 -3.4646
1.5268 1.1716 6800 1.6184 -3.6529 -4.0082 0.6217 0.3553 -2.0041 -1.8264 -3.4472 -3.4509
1.552 1.1888 6900 1.6182 -3.6682 -4.0248 0.6215 0.3565 -2.0124 -1.8341 -3.4597 -3.4633
1.5713 1.2061 7000 1.6170 -3.6855 -4.0446 0.6231 0.3591 -2.0223 -1.8427 -3.4683 -3.4718
1.6189 1.2233 7100 1.6174 -3.6787 -4.0380 0.6215 0.3592 -2.0190 -1.8394 -3.4553 -3.4589
1.488 1.2405 7200 1.6166 -3.7005 -4.0616 0.6217 0.3612 -2.0308 -1.8502 -3.4373 -3.4410
1.5506 1.2578 7300 1.6159 -3.7223 -4.0849 0.6224 0.3626 -2.0425 -1.8612 -3.4467 -3.4503
1.5274 1.2750 7400 1.6148 -3.7367 -4.1018 0.6241 0.3652 -2.0509 -1.8683 -3.4483 -3.4519
1.547 1.2922 7500 1.6138 -3.7467 -4.1131 0.6217 0.3664 -2.0565 -1.8734 -3.4345 -3.4381
1.4958 1.3094 7600 1.6142 -3.7568 -4.1234 0.6231 0.3666 -2.0617 -1.8784 -3.4377 -3.4412
1.4875 1.3267 7700 1.6137 -3.7720 -4.1409 0.6224 0.3689 -2.0704 -1.8860 -3.4242 -3.4279
1.5489 1.3439 7800 1.6132 -3.7819 -4.1526 0.6229 0.3707 -2.0763 -1.8909 -3.4258 -3.4294
1.6241 1.3611 7900 1.6129 -3.7899 -4.1616 0.6238 0.3717 -2.0808 -1.8949 -3.4339 -3.4374
1.6697 1.3784 8000 1.6118 -3.8009 -4.1750 0.6236 0.3741 -2.0875 -1.9004 -3.4167 -3.4203
1.5586 1.3956 8100 1.6123 -3.8096 -4.1836 0.6241 0.3740 -2.0918 -1.9048 -3.4215 -3.4250
1.3943 1.4128 8200 1.6110 -3.8221 -4.1990 0.6245 0.3769 -2.0995 -1.9110 -3.4127 -3.4163
1.6019 1.4300 8300 1.6098 -3.8372 -4.2158 0.6245 0.3786 -2.1079 -1.9186 -3.4157 -3.4193
1.475 1.4473 8400 1.6101 -3.8498 -4.2288 0.625 0.3789 -2.1144 -1.9249 -3.4113 -3.4149
1.5141 1.4645 8500 1.6099 -3.8489 -4.2288 0.6248 0.3799 -2.1144 -1.9245 -3.4117 -3.4152
1.5064 1.4817 8600 1.6103 -3.8593 -4.2395 0.6238 0.3802 -2.1198 -1.9297 -3.4115 -3.4151
1.5121 1.4990 8700 1.6100 -3.8718 -4.2527 0.6241 0.3810 -2.1264 -1.9359 -3.4052 -3.4087
1.4344 1.5162 8800 1.6097 -3.8842 -4.2665 0.625 0.3824 -2.1333 -1.9421 -3.4016 -3.4051
1.4826 1.5334 8900 1.6085 -3.9086 -4.2937 0.6259 0.3851 -2.1468 -1.9543 -3.4018 -3.4053
1.5369 1.5507 9000 1.6084 -3.9188 -4.3046 0.6257 0.3857 -2.1523 -1.9594 -3.3997 -3.4032
1.6204 1.5679 9100 1.6072 -3.9245 -4.3122 0.6255 0.3877 -2.1561 -1.9623 -3.3930 -3.3965
1.5032 1.5851 9200 1.6078 -3.9320 -4.3202 0.6273 0.3882 -2.1601 -1.9660 -3.3925 -3.3961
1.5816 1.6023 9300 1.6080 -3.9339 -4.3227 0.6273 0.3887 -2.1613 -1.9670 -3.3853 -3.3889
1.5464 1.6196 9400 1.6076 -3.9388 -4.3286 0.6266 0.3898 -2.1643 -1.9694 -3.3809 -3.3845
1.4955 1.6368 9500 1.6068 -3.9471 -4.3382 0.6271 0.3912 -2.1691 -1.9735 -3.3932 -3.3967
1.5395 1.6540 9600 1.6069 -3.9479 -4.3397 0.6259 0.3918 -2.1699 -1.9739 -3.3839 -3.3875
1.4387 1.6713 9700 1.6068 -3.9502 -4.3421 0.6276 0.3919 -2.1711 -1.9751 -3.3783 -3.3818
1.3438 1.6885 9800 1.6067 -3.9572 -4.3502 0.6276 0.3929 -2.1751 -1.9786 -3.3939 -3.3974
1.3561 1.7057 9900 1.6061 -3.9663 -4.3606 0.6285 0.3943 -2.1803 -1.9832 -3.3849 -3.3884
1.3892 1.7229 10000 1.6060 -3.9747 -4.3696 0.6257 0.3949 -2.1848 -1.9874 -3.3726 -3.3761
1.5131 1.7402 10100 1.6058 -3.9802 -4.3758 0.6266 0.3955 -2.1879 -1.9901 -3.3776 -3.3811
1.5061 1.7574 10200 1.6050 -3.9996 -4.3969 0.6266 0.3974 -2.1985 -1.9998 -3.3678 -3.3713
1.6132 1.7746 10300 1.6050 -4.0049 -4.4028 0.6231 0.3979 -2.2014 -2.0024 -3.3743 -3.3778
1.3357 1.7919 10400 1.6050 -4.0040 -4.4026 0.6213 0.3986 -2.2013 -2.0020 -3.3710 -3.3744
1.4868 1.8091 10500 1.6045 -4.0107 -4.4100 0.6187 0.3993 -2.2050 -2.0054 -3.3805 -3.3839
1.5879 1.8263 10600 1.6052 -4.0153 -4.4144 0.6197 0.3991 -2.2072 -2.0077 -3.3635 -3.3670
1.4603 1.8436 10700 1.6047 -4.0231 -4.4232 0.6220 0.4001 -2.2116 -2.0115 -3.3620 -3.3655
1.3798 1.8608 10800 1.6042 -4.0306 -4.4320 0.6227 0.4013 -2.2160 -2.0153 -3.3691 -3.3725
1.4895 1.8780 10900 1.6039 -4.0340 -4.4358 0.6208 0.4018 -2.2179 -2.0170 -3.3651 -3.3685
1.6103 1.8952 11000 1.6041 -4.0366 -4.4389 0.6220 0.4022 -2.2194 -2.0183 -3.3628 -3.3663
1.5105 1.9125 11100 1.6033 -4.0506 -4.4549 0.6220 0.4044 -2.2275 -2.0253 -3.3548 -3.3583
1.3955 1.9297 11200 1.6034 -4.0569 -4.4612 0.6213 0.4043 -2.2306 -2.0284 -3.3499 -3.3534
1.6675 1.9469 11300 1.6030 -4.0634 -4.4689 0.6248 0.4055 -2.2345 -2.0317 -3.3599 -3.3633
1.467 1.9642 11400 1.6030 -4.0655 -4.4710 0.6220 0.4055 -2.2355 -2.0327 -3.3408 -3.3444
1.6141 1.9814 11500 1.6028 -4.0747 -4.4814 0.6245 0.4067 -2.2407 -2.0373 -3.3533 -3.3568
1.4188 1.9986 11600 1.6029 -4.0793 -4.4862 0.6241 0.4070 -2.2431 -2.0396 -3.3551 -3.3585
1.3363 2.0159 11700 1.6028 -4.0903 -4.4982 0.6204 0.4079 -2.2491 -2.0452 -3.3525 -3.3559
1.5703 2.0331 11800 1.6029 -4.0960 -4.5040 0.625 0.4080 -2.2520 -2.0480 -3.3542 -3.3576
1.4007 2.0503 11900 1.6025 -4.1070 -4.5164 0.6264 0.4094 -2.2582 -2.0535 -3.3537 -3.3571
1.3923 2.0675 12000 1.6020 -4.1075 -4.5177 0.6220 0.4102 -2.2588 -2.0538 -3.3445 -3.3480
1.606 2.0848 12100 1.6018 -4.1121 -4.5235 0.6236 0.4113 -2.2617 -2.0561 -3.3442 -3.3476
1.5084 2.1020 12200 1.6017 -4.1195 -4.5307 0.6257 0.4112 -2.2654 -2.0597 -3.3465 -3.3499
1.4099 2.1192 12300 1.6014 -4.1198 -4.5311 0.6229 0.4112 -2.2655 -2.0599 -3.3430 -3.3464
1.5056 2.1365 12400 1.6009 -4.1230 -4.5360 0.6213 0.4129 -2.2680 -2.0615 -3.3393 -3.3427
1.3618 2.1537 12500 1.6016 -4.1280 -4.5395 0.6227 0.4115 -2.2698 -2.0640 -3.3424 -3.3459
1.3944 2.1709 12600 1.6015 -4.1305 -4.5426 0.6248 0.4121 -2.2713 -2.0652 -3.3480 -3.3514
1.5202 2.1881 12700 1.6014 -4.1351 -4.5482 0.6213 0.4131 -2.2741 -2.0675 -3.3383 -3.3418
1.5605 2.2054 12800 1.6009 -4.1366 -4.5507 0.6234 0.4141 -2.2754 -2.0683 -3.3370 -3.3404
1.3645 2.2226 12900 1.6009 -4.1383 -4.5525 0.6224 0.4142 -2.2762 -2.0691 -3.3402 -3.3436
1.5051 2.2398 13000 1.6006 -4.1434 -4.5586 0.6229 0.4151 -2.2793 -2.0717 -3.3364 -3.3398
1.4171 2.2571 13100 1.6011 -4.1443 -4.5592 0.6224 0.4149 -2.2796 -2.0721 -3.3394 -3.3428
1.4166 2.2743 13200 1.6005 -4.1497 -4.5654 0.6227 0.4158 -2.2827 -2.0748 -3.3398 -3.3432
1.5389 2.2915 13300 1.6007 -4.1508 -4.5665 0.6234 0.4157 -2.2832 -2.0754 -3.3449 -3.3483
1.4618 2.3088 13400 1.6007 -4.1553 -4.5710 0.6227 0.4157 -2.2855 -2.0776 -3.3437 -3.3471
1.3821 2.3260 13500 1.6001 -4.1574 -4.5743 0.6229 0.4170 -2.2872 -2.0787 -3.3213 -3.3248
1.4958 2.3432 13600 1.5997 -4.1605 -4.5782 0.6241 0.4177 -2.2891 -2.0802 -3.3339 -3.3374
1.5225 2.3604 13700 1.6000 -4.1639 -4.5813 0.6227 0.4174 -2.2906 -2.0820 -3.3315 -3.3349
1.5279 2.3777 13800 1.5999 -4.1666 -4.5843 0.6234 0.4177 -2.2921 -2.0833 -3.3375 -3.3409
1.5492 2.3949 13900 1.5997 -4.1676 -4.5857 0.6227 0.4182 -2.2929 -2.0838 -3.3367 -3.3401
1.4219 2.4121 14000 1.5998 -4.1724 -4.5908 0.6231 0.4184 -2.2954 -2.0862 -3.3231 -3.3265
1.4625 2.4294 14100 1.5994 -4.1764 -4.5952 0.6238 0.4188 -2.2976 -2.0882 -3.3154 -3.3189
1.3039 2.4466 14200 1.5993 -4.1746 -4.5941 0.6231 0.4195 -2.2971 -2.0873 -3.3285 -3.3319
1.4333 2.4638 14300 1.5993 -4.1779 -4.5973 0.6238 0.4194 -2.2987 -2.0889 -3.3319 -3.3353
1.4677 2.4810 14400 1.5992 -4.1805 -4.6002 0.6229 0.4197 -2.3001 -2.0902 -3.3219 -3.3253
1.3125 2.4983 14500 1.5994 -4.1824 -4.6024 0.6229 0.4200 -2.3012 -2.0912 -3.3350 -3.3384
1.4611 2.5155 14600 1.5989 -4.1839 -4.6043 0.6248 0.4204 -2.3021 -2.0920 -3.3344 -3.3378
1.4287 2.5327 14700 1.5989 -4.1868 -4.6073 0.6231 0.4205 -2.3037 -2.0934 -3.3421 -3.3455
1.5098 2.5500 14800 1.5989 -4.1855 -4.6061 0.6234 0.4206 -2.3031 -2.0928 -3.3370 -3.3403
1.3432 2.5672 14900 1.5995 -4.1878 -4.6080 0.6231 0.4202 -2.3040 -2.0939 -3.3237 -3.3271
1.6495 2.5844 15000 1.5992 -4.1893 -4.6094 0.6231 0.4201 -2.3047 -2.0947 -3.3315 -3.3349
1.4971 2.6017 15100 1.5992 -4.1890 -4.6095 0.6234 0.4205 -2.3048 -2.0945 -3.3235 -3.3270
1.3488 2.6189 15200 1.5990 -4.1909 -4.6118 0.6231 0.4210 -2.3059 -2.0954 -3.3239 -3.3273
1.3814 2.6361 15300 1.5994 -4.1911 -4.6115 0.6229 0.4204 -2.3058 -2.0955 -3.3206 -3.3240
1.4437 2.6533 15400 1.5993 -4.1905 -4.6109 0.6222 0.4204 -2.3054 -2.0952 -3.3217 -3.3252
1.5573 2.6706 15500 1.5995 -4.1915 -4.6116 0.6222 0.4201 -2.3058 -2.0958 -3.3258 -3.3293
1.4515 2.6878 15600 1.5986 -4.1902 -4.6120 0.6229 0.4219 -2.3060 -2.0951 -3.3170 -3.3205
1.3256 2.7050 15700 1.5993 -4.1914 -4.6118 0.6227 0.4204 -2.3059 -2.0957 -3.3388 -3.3421
1.4458 2.7223 15800 1.6001 -4.1918 -4.6113 0.6220 0.4195 -2.3057 -2.0959 -3.3286 -3.3321
1.3734 2.7395 15900 1.5991 -4.1906 -4.6111 0.6227 0.4206 -2.3056 -2.0953 -3.3224 -3.3258
1.4477 2.7567 16000 1.5998 -4.1924 -4.6122 0.6224 0.4198 -2.3061 -2.0962 -3.3408 -3.3441
1.401 2.7739 16100 1.5992 -4.1917 -4.6125 0.6234 0.4208 -2.3063 -2.0959 -3.3096 -3.3131
1.422 2.7912 16200 1.5998 -4.1927 -4.6123 0.6236 0.4196 -2.3062 -2.0964 -3.3248 -3.3282
1.4691 2.8084 16300 1.5994 -4.1918 -4.6125 0.6236 0.4207 -2.3062 -2.0959 -3.3187 -3.3222
1.4821 2.8256 16400 1.5993 -4.1923 -4.6129 0.6241 0.4206 -2.3064 -2.0962 -3.3167 -3.3202
1.539 2.8429 16500 1.6001 -4.1929 -4.6126 0.6234 0.4197 -2.3063 -2.0964 -3.3192 -3.3227
1.7983 2.8601 16600 1.5994 -4.1926 -4.6132 0.6224 0.4206 -2.3066 -2.0963 -3.3258 -3.3293
1.4889 2.8773 16700 1.5994 -4.1917 -4.6125 0.6229 0.4208 -2.3062 -2.0958 -3.3144 -3.3179
1.5191 2.8946 16800 1.5994 -4.1924 -4.6128 0.6222 0.4204 -2.3064 -2.0962 -3.3194 -3.3229
1.6401 2.9118 16900 1.5999 -4.1929 -4.6129 0.6224 0.4199 -2.3064 -2.0965 -3.3256 -3.3291
1.5593 2.9290 17000 1.5989 -4.1926 -4.6138 0.6227 0.4212 -2.3069 -2.0963 -3.3279 -3.3313
1.5395 2.9462 17100 1.5989 -4.1923 -4.6135 0.6234 0.4212 -2.3068 -2.0962 -3.3291 -3.3325
1.7984 2.9635 17200 1.5992 -4.1921 -4.6128 0.6227 0.4207 -2.3064 -2.0960 -3.3195 -3.3230
1.6222 2.9807 17300 1.5992 -4.1931 -4.6141 0.6238 0.4210 -2.3070 -2.0965 -3.3339 -3.3372
1.4575 2.9979 17400 1.5991 -4.1926 -4.6137 0.6231 0.4211 -2.3069 -2.0963 -3.3338 -3.3372

Framework versions

  • Transformers 4.41.2
  • Pytorch 2.1.2
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
36
Safetensors
Model size
1.1B params
Tensor type
BF16
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.

Model tree for martimfasantos/tinyllama-1.1b-sum-simpo_beta2.0_gamma1.6_LR5e-8_3epochs

Dataset used to train martimfasantos/tinyllama-1.1b-sum-simpo_beta2.0_gamma1.6_LR5e-8_3epochs