lapp0 commited on
Commit
8eac14b
1 Parent(s): c762f94

End of training

Browse files
README.md CHANGED
@@ -16,13 +16,13 @@ This student model is distilled from the teacher model [gpt2](https://huggingfac
16
  The [Distily](https://github.com/lapp0/distily) library was used for this distillation.
17
 
18
  It achieves the following results on the evaluation set:
19
- - eval_enwikippl: 215.4055
20
- - eval_frwikippl: 1190.7479
21
- - eval_zhwikippl: 547.2146
22
- - eval_loss: 1.2012
23
- - eval_runtime: 86.3928
24
- - eval_samples_per_second: 57.875
25
- - eval_steps_per_second: 7.234
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment.
@@ -45,7 +45,7 @@ More information needed
45
  ### Training hyperparameters
46
 
47
  The following hyperparameters were used during training:
48
- - distillation_objective: DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl, layer_mapper=None, projector=None), hs_loss_component=LossComponent(label=hs, weight=0, loss_fn=None, layer_mapper=None, projector=None), attn_loss_component=LossComponent(label=attn, weight=2.0, loss_fn=mse, layer_mapper=None, projector=None))
49
  - train_embeddings: True
50
  - learning_rate: 4e-05
51
  - train_batch_size: 8
@@ -62,69 +62,69 @@ Peak GPU Memory: 8.2206 GB
62
  | step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | zhwikippl |
63
  | --- | --- | --- | --- | --- | --- | --- | --- | --- |
64
  | **teacher eval** | | 30.2086 | 57.2728 | | | | | 18.1784 |
65
- | 0 | 0 | 56314.7695 | 59887.2773 | 5.8256 | 86.2439 | 57.975 | 7.247 | 59033.8086 |
66
- | 1000 | 0.0162 | 707.4770 | 4242.8809 | 1.8516 | 86.1491 | 58.039 | 7.255 | 11038.7695 |
67
- | 2000 | 0.0323 | 507.7405 | 3239.7178 | 1.6796 | 86.186 | 58.014 | 7.252 | 1887.9902 |
68
- | 3000 | 0.0485 | 425.0894 | 2858.4150 | 1.5756 | 86.0159 | 58.129 | 7.266 | 841.8765 |
69
- | 4000 | 0.0646 | 361.4349 | 2351.2927 | 1.4943 | 86.0626 | 58.097 | 7.262 | 1237.3851 |
70
- | 5000 | 0.0808 | 320.2736 | 1811.6420 | 1.4160 | 86.3077 | 57.932 | 7.242 | 941.8109 |
71
- | 6000 | 0.0970 | 279.4263 | 1586.2935 | 1.3478 | 86.3392 | 57.911 | 7.239 | 744.3502 |
72
- | 7000 | 0.1131 | 252.5366 | 1452.6782 | 1.2903 | 86.3844 | 57.881 | 7.235 | 651.1284 |
73
- | 8000 | 0.1293 | 229.7639 | 1333.1338 | 1.2422 | 86.4019 | 57.869 | 7.234 | 586.1718 |
74
- | 9000 | 0.1455 | 215.4055 | 1190.7479 | 1.2012 | 86.3928 | 57.875 | 7.234 | 547.2146 |
75
- | 10000 | 0.1616 | 195.7073 | 1147.2347 | 1.1512 | 86.3689 | 57.891 | 7.236 | 673.5028 |
76
- | 11000 | 0.1778 | 181.4088 | 1060.8735 | 1.1073 | 86.4921 | 57.809 | 7.226 | 521.8091 |
77
- | 12000 | 0.1939 | 164.0534 | 896.9886 | 1.0636 | 86.3399 | 57.911 | 7.239 | 488.8237 |
78
- | 13000 | 0.2101 | 157.4142 | 890.0587 | 1.0357 | 86.4286 | 57.851 | 7.231 | 510.7101 |
79
- | 14000 | 0.2263 | 148.5198 | 793.2602 | 1.0069 | 86.4451 | 57.84 | 7.23 | 415.8904 |
80
- | 15000 | 0.2424 | 143.5310 | 728.5455 | 0.9844 | 86.5014 | 57.803 | 7.225 | 414.5595 |
81
- | 16000 | 0.2586 | 139.7042 | 766.6470 | 0.9726 | 86.5584 | 57.764 | 7.221 | 539.9557 |
82
- | 17000 | 0.2747 | 136.4025 | 723.4780 | 0.9594 | 86.3816 | 57.883 | 7.235 | 877.2245 |
83
- | 18000 | 0.2909 | 133.8320 | 733.1834 | 0.9461 | 86.4657 | 57.826 | 7.228 | 582.4266 |
84
- | 19000 | 0.3071 | 130.4055 | 720.7795 | 0.9391 | 86.5854 | 57.746 | 7.218 | 564.7347 |
85
- | 20000 | 0.3232 | 128.2763 | 679.3307 | 0.9259 | 86.469 | 57.824 | 7.228 | 364.2420 |
86
- | 21000 | 0.3394 | 126.0545 | 666.4741 | 0.9208 | 86.3084 | 57.932 | 7.241 | 392.6297 |
87
- | 22000 | 0.3556 | 126.3289 | 618.9599 | 0.9146 | 86.2819 | 57.95 | 7.244 | 383.1512 |
88
- | 23000 | 0.3717 | 125.7710 | 652.6170 | 0.9106 | 86.3709 | 57.89 | 7.236 | 382.0272 |
89
- | 24000 | 0.3879 | 121.7352 | 649.1292 | 0.9010 | 86.4132 | 57.862 | 7.233 | 407.5338 |
90
- | 25000 | 0.4040 | 121.2164 | 677.1313 | 0.8985 | 86.5605 | 57.763 | 7.22 | 378.4727 |
91
- | 26000 | 0.4202 | 121.4331 | 604.5543 | 0.8920 | 86.6149 | 57.727 | 7.216 | 400.5201 |
92
- | 27000 | 0.4364 | 121.4896 | 636.5748 | 0.8898 | 86.977 | 57.486 | 7.186 | 344.3297 |
93
- | 28000 | 0.4525 | 120.0641 | 614.8710 | 0.8867 | 86.9971 | 57.473 | 7.184 | 385.8209 |
94
- | 29000 | 0.4687 | 121.5085 | 662.3517 | 0.8855 | 86.6921 | 57.675 | 7.209 | 386.8527 |
95
- | 30000 | 0.4848 | 121.3954 | 620.4891 | 0.8915 | 86.9396 | 57.511 | 7.189 | 805.0448 |
96
- | 31000 | 0.5010 | 119.1724 | 604.0428 | 0.8831 | 87.0473 | 57.44 | 7.18 | 382.2313 |
97
- | 32000 | 0.5172 | 118.1496 | 632.1021 | 0.8800 | 87.0169 | 57.46 | 7.183 | 377.2617 |
98
- | 33000 | 0.5333 | 116.5277 | 597.8567 | 0.8738 | 86.7512 | 57.636 | 7.205 | 322.2620 |
99
- | 34000 | 0.5495 | 116.1844 | 591.6924 | 0.8734 | 87.2311 | 57.319 | 7.165 | 431.3317 |
100
- | 35000 | 0.5657 | 115.5994 | 565.9454 | 0.8686 | 86.8167 | 57.593 | 7.199 | 336.3313 |
101
- | 36000 | 0.5818 | 115.9320 | 609.9918 | 0.8674 | 87.1488 | 57.373 | 7.172 | 253.6102 |
102
- | 37000 | 0.5980 | 115.0621 | 595.2911 | 0.8660 | 87.1004 | 57.405 | 7.176 | 323.4260 |
103
- | 38000 | 0.6141 | 115.5635 | 590.6086 | 0.8654 | 86.9067 | 57.533 | 7.192 | 282.2412 |
104
- | 39000 | 0.6303 | 113.5796 | 546.1489 | 0.8586 | 86.5012 | 57.803 | 7.225 | 306.1125 |
105
- | 40000 | 0.6465 | 113.4385 | 558.4144 | 0.8583 | 86.6261 | 57.719 | 7.215 | 246.7947 |
106
- | 41000 | 0.6626 | 112.7097 | 563.5562 | 0.8558 | 86.9289 | 57.518 | 7.19 | 263.4834 |
107
- | 42000 | 0.6788 | 112.6048 | 556.9202 | 0.8573 | 86.8975 | 57.539 | 7.192 | 287.7979 |
108
- | 43000 | 0.6949 | 112.9025 | 569.7087 | 0.8534 | 86.3213 | 57.923 | 7.24 | 295.2722 |
109
- | 44000 | 0.7111 | 111.3180 | 584.7252 | 0.8534 | 86.7833 | 57.615 | 7.202 | 311.5563 |
110
- | 45000 | 0.7273 | 112.7623 | 589.8597 | 0.8520 | 85.8832 | 58.219 | 7.277 | 452.9366 |
111
- | 46000 | 0.7434 | 111.0763 | 583.6953 | 0.8497 | 86.9028 | 57.536 | 7.192 | 323.7285 |
112
- | 47000 | 0.7596 | 110.0631 | 570.5529 | 0.8481 | 86.1396 | 58.045 | 7.256 | 278.4229 |
113
- | 48000 | 0.7758 | 112.4039 | 498.8431 | 0.8470 | 86.0091 | 58.133 | 7.267 | 315.6181 |
114
- | 49000 | 0.7919 | 111.2748 | 564.9885 | 0.8465 | 86.4014 | 57.869 | 7.234 | 261.0319 |
115
- | 50000 | 0.8081 | 111.5950 | 594.9554 | 0.8454 | 87.4501 | 57.175 | 7.147 | 240.7725 |
116
- | 51000 | 0.8242 | 110.0546 | 563.8345 | 0.8446 | 85.9134 | 58.198 | 7.275 | 320.1174 |
117
- | 52000 | 0.8404 | 109.2966 | 548.4256 | 0.8428 | 86.5788 | 57.751 | 7.219 | 318.7099 |
118
- | 53000 | 0.8566 | 109.3136 | 539.9846 | 0.8395 | 86.3394 | 57.911 | 7.239 | 340.8982 |
119
- | 54000 | 0.8727 | 110.7834 | 561.4149 | 0.8436 | 86.2011 | 58.004 | 7.25 | 361.5285 |
120
- | 55000 | 0.8889 | 110.2941 | 576.0907 | 0.8421 | 86.733 | 57.648 | 7.206 | 297.2107 |
121
- | 56000 | 0.9051 | 109.5600 | 571.4385 | 0.8433 | 86.2508 | 57.97 | 7.246 | 370.3730 |
122
- | 57000 | 0.9212 | 109.7474 | 566.3444 | 0.8457 | 86.5407 | 57.776 | 7.222 | 900.0065 |
123
- | 58000 | 0.9374 | 109.4155 | 621.2332 | 0.8426 | 86.3669 | 57.893 | 7.237 | 493.3487 |
124
- | 59000 | 0.9535 | 110.1230 | 581.3542 | 0.8391 | 86.4324 | 57.849 | 7.231 | 272.2826 |
125
- | 60000 | 0.9697 | 108.2997 | 582.5030 | 0.8340 | 86.046 | 58.108 | 7.264 | 323.5555 |
126
- | 61000 | 0.9859 | 109.2711 | 566.8240 | 0.8381 | 86.749 | 57.638 | 7.205 | 312.4312 |
127
- | 61875 | 1.0 | 109.1439 | 575.3599 | 0.8346 | 86.8825 | 57.549 | 7.194 | 265.7449 |
128
 
129
  ### Framework versions
130
  - Distily 0.2.0
 
16
  The [Distily](https://github.com/lapp0/distily) library was used for this distillation.
17
 
18
  It achieves the following results on the evaluation set:
19
+ - eval_enwikippl: 234.3043
20
+ - eval_frwikippl: 1329.5667
21
+ - eval_zhwikippl: 575.8531
22
+ - eval_loss: 2.4344
23
+ - eval_runtime: 87.576
24
+ - eval_samples_per_second: 57.093
25
+ - eval_steps_per_second: 7.137
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment.
 
45
  ### Training hyperparameters
46
 
47
  The following hyperparameters were used during training:
48
+ - distillation_objective: DistillationObjective(logits_loss_component=LossComponent(label=logits, weight=1, loss_fn=kl, layer_mapper=None, projector=None), hs_loss_component=LossComponent(label=hs, weight=0, loss_fn=None, layer_mapper=None, projector=None), attn_loss_component=LossComponent(label=attn, weight=2.0, loss_fn=cos, layer_mapper=None, projector=None))
49
  - train_embeddings: True
50
  - learning_rate: 4e-05
51
  - train_batch_size: 8
 
62
  | step | epoch | enwikippl | frwikippl | loss | runtime | samples_per_second | steps_per_second | zhwikippl |
63
  | --- | --- | --- | --- | --- | --- | --- | --- | --- |
64
  | **teacher eval** | | 30.2086 | 57.2728 | | | | | 18.1784 |
65
+ | 0 | 0 | 56314.7695 | 59887.2773 | 7.8201 | 86.1883 | 58.013 | 7.252 | 59033.8086 |
66
+ | 1000 | 0.0162 | 792.2872 | 4804.1196 | 3.2178 | 86.0729 | 58.09 | 7.261 | 14971.4619 |
67
+ | 2000 | 0.0323 | 563.5076 | 3594.0178 | 3.0017 | 86.8465 | 57.573 | 7.197 | 2436.5176 |
68
+ | 3000 | 0.0485 | 462.8162 | 3038.3840 | 2.8886 | 86.4026 | 57.869 | 7.234 | 990.3070 |
69
+ | 4000 | 0.0646 | 394.8591 | 2549.8823 | 2.7877 | 86.3461 | 57.907 | 7.238 | 1110.5270 |
70
+ | 5000 | 0.0808 | 353.7159 | 2113.5315 | 2.6983 | 86.5685 | 57.758 | 7.22 | 845.9332 |
71
+ | 6000 | 0.0970 | 305.3852 | 1907.5966 | 2.6161 | 86.5704 | 57.756 | 7.22 | 792.3533 |
72
+ | 7000 | 0.1131 | 275.6120 | 1710.8368 | 2.5482 | 89.5827 | 55.814 | 6.977 | 704.9745 |
73
+ | 8000 | 0.1293 | 248.5870 | 1491.7041 | 2.4852 | 87.8996 | 56.883 | 7.11 | 655.4906 |
74
+ | 9000 | 0.1455 | 234.3043 | 1329.5667 | 2.4344 | 87.576 | 57.093 | 7.137 | 575.8531 |
75
+ | 10000 | 0.1616 | 210.1025 | 1200.8650 | 2.3778 | 88.2437 | 56.661 | 7.083 | 680.8273 |
76
+ | 11000 | 0.1778 | 195.9659 | 1137.0879 | 2.3352 | 87.2617 | 57.299 | 7.162 | 574.3172 |
77
+ | 12000 | 0.1939 | 177.5484 | 986.7017 | 2.2840 | 88.3708 | 56.58 | 7.072 | 511.2562 |
78
+ | 13000 | 0.2101 | 168.3902 | 992.2828 | 2.2496 | 86.6049 | 57.733 | 7.217 | 493.3487 |
79
+ | 14000 | 0.2263 | 159.1225 | 889.1183 | 2.2152 | 86.6651 | 57.693 | 7.212 | 434.3372 |
80
+ | 15000 | 0.2424 | 153.0509 | 800.1130 | 2.1876 | 86.8087 | 57.598 | 7.2 | 389.5484 |
81
+ | 16000 | 0.2586 | 146.4697 | 801.1292 | 2.1678 | 86.6505 | 57.703 | 7.213 | 490.2618 |
82
+ | 17000 | 0.2747 | 143.1525 | 782.1525 | 2.1519 | 86.9013 | 57.537 | 7.192 | 536.9359 |
83
+ | 18000 | 0.2909 | 139.4116 | 832.3362 | 2.1366 | 86.6158 | 57.726 | 7.216 | 568.2902 |
84
+ | 19000 | 0.3071 | 134.7601 | 733.2869 | 2.1223 | 86.3964 | 57.873 | 7.234 | 516.8853 |
85
+ | 20000 | 0.3232 | 132.8793 | 726.0842 | 2.1108 | 86.4694 | 57.824 | 7.228 | 376.9092 |
86
+ | 21000 | 0.3394 | 130.5677 | 658.1619 | 2.0982 | 86.9266 | 57.52 | 7.19 | 386.2850 |
87
+ | 22000 | 0.3556 | 130.3043 | 657.9764 | 2.0894 | 87.5122 | 57.135 | 7.142 | 418.9560 |
88
+ | 23000 | 0.3717 | 128.9555 | 687.2317 | 2.0831 | 87.0863 | 57.414 | 7.177 | 419.0120 |
89
+ | 24000 | 0.3879 | 125.7514 | 657.8835 | 2.0732 | 86.5642 | 57.761 | 7.22 | 391.6348 |
90
+ | 25000 | 0.4040 | 124.6431 | 676.3678 | 2.0703 | 86.6818 | 57.682 | 7.21 | 384.3811 |
91
+ | 26000 | 0.4202 | 124.4787 | 653.3537 | 2.0597 | 87.1964 | 57.342 | 7.168 | 403.9578 |
92
+ | 27000 | 0.4364 | 122.7223 | 659.5090 | 2.0546 | 86.5911 | 57.743 | 7.218 | 316.7159 |
93
+ | 28000 | 0.4525 | 123.1997 | 631.8794 | 2.0501 | 86.9662 | 57.494 | 7.187 | 295.1540 |
94
+ | 29000 | 0.4687 | 122.9990 | 659.6486 | 2.0423 | 87.0187 | 57.459 | 7.182 | 310.8498 |
95
+ | 30000 | 0.4848 | 123.1041 | 617.8698 | 2.0464 | 87.3158 | 57.263 | 7.158 | 315.3653 |
96
+ | 31000 | 0.5010 | 120.1201 | 613.0962 | 2.0351 | 87.311 | 57.267 | 7.158 | 308.4515 |
97
+ | 32000 | 0.5172 | 119.4781 | 630.6332 | 2.0323 | 87.0327 | 57.45 | 7.181 | 274.0701 |
98
+ | 33000 | 0.5333 | 118.0396 | 624.0867 | 2.0277 | 86.8174 | 57.592 | 7.199 | 285.6919 |
99
+ | 34000 | 0.5495 | 119.1354 | 588.2814 | 2.0249 | 86.7837 | 57.614 | 7.202 | 316.9274 |
100
+ | 35000 | 0.5657 | 117.2539 | 567.9440 | 2.0212 | 87.7979 | 56.949 | 7.119 | 303.2645 |
101
+ | 36000 | 0.5818 | 117.5548 | 608.1882 | 2.0241 | 86.4044 | 57.867 | 7.233 | 256.7451 |
102
+ | 37000 | 0.5980 | 116.6455 | 600.1797 | 2.0164 | 87.1429 | 57.377 | 7.172 | 345.4349 |
103
+ | 38000 | 0.6141 | 116.6998 | 564.5506 | 2.0140 | 87.5854 | 57.087 | 7.136 | 283.1472 |
104
+ | 39000 | 0.6303 | 113.8975 | 538.0084 | 2.0085 | 86.8118 | 57.596 | 7.199 | 285.9208 |
105
+ | 40000 | 0.6465 | 115.6533 | 579.6762 | 2.0130 | 87.6272 | 57.06 | 7.132 | 268.8502 |
106
+ | 41000 | 0.6626 | 114.0037 | 569.7087 | 2.0107 | 86.8951 | 57.541 | 7.193 | 298.6428 |
107
+ | 42000 | 0.6788 | 114.4206 | 558.2177 | 2.0114 | 86.4056 | 57.867 | 7.233 | 325.7667 |
108
+ | 43000 | 0.6949 | 114.1277 | 570.9554 | 2.0067 | 86.7633 | 57.628 | 7.204 | 297.6475 |
109
+ | 44000 | 0.7111 | 112.6310 | 603.2343 | 2.0041 | 87.2036 | 57.337 | 7.167 | 265.9578 |
110
+ | 45000 | 0.7273 | 112.3951 | 582.9551 | 1.9978 | 86.1934 | 58.009 | 7.251 | 276.3855 |
111
+ | 46000 | 0.7434 | 112.9463 | 591.3171 | 1.9976 | 86.5934 | 57.741 | 7.218 | 270.1097 |
112
+ | 47000 | 0.7596 | 112.1510 | 564.6300 | 1.9943 | 86.827 | 57.586 | 7.198 | 323.6853 |
113
+ | 48000 | 0.7758 | 112.8236 | 513.6188 | 1.9968 | 86.3974 | 57.872 | 7.234 | 305.2553 |
114
+ | 49000 | 0.7919 | 112.7886 | 565.1480 | 1.9948 | 86.5571 | 57.765 | 7.221 | 276.0167 |
115
+ | 50000 | 0.8081 | 111.8900 | 592.2350 | 1.9932 | 86.9533 | 57.502 | 7.188 | 247.9840 |
116
+ | 51000 | 0.8242 | 111.4391 | 588.3229 | 1.9920 | 86.4566 | 57.832 | 7.229 | 298.2842 |
117
+ | 52000 | 0.8404 | 109.9350 | 549.1997 | 1.9904 | 86.5867 | 57.746 | 7.218 | 318.3695 |
118
+ | 53000 | 0.8566 | 110.8264 | 544.7263 | 1.9856 | 87.1758 | 57.355 | 7.169 | 311.5147 |
119
+ | 54000 | 0.8727 | 111.0849 | 544.9952 | 1.9857 | 87.388 | 57.216 | 7.152 | 334.9867 |
120
+ | 55000 | 0.8889 | 111.1799 | 602.7242 | 1.9865 | 86.8376 | 57.579 | 7.197 | 265.1776 |
121
+ | 56000 | 0.9051 | 111.2490 | 553.5536 | 1.9821 | 87.5367 | 57.119 | 7.14 | 326.5943 |
122
+ | 57000 | 0.9212 | 110.1914 | 582.8317 | 1.9870 | 87.0656 | 57.428 | 7.178 | 1162.1104 |
123
+ | 58000 | 0.9374 | 109.1016 | 657.6982 | 1.9860 | 86.6253 | 57.72 | 7.215 | 322.6926 |
124
+ | 59000 | 0.9535 | 111.8119 | 596.5937 | 1.9831 | 86.3097 | 57.931 | 7.241 | 408.6782 |
125
+ | 60000 | 0.9697 | 108.9746 | 586.0871 | 1.9748 | 86.6963 | 57.673 | 7.209 | 268.2405 |
126
+ | 61000 | 0.9859 | 109.9862 | 560.1890 | 1.9777 | 86.8734 | 57.555 | 7.194 | 294.4846 |
127
+ | 61875 | 1.0 | 110.1401 | 573.6587 | 1.9760 | 86.766 | 57.626 | 7.203 | 273.4851 |
128
 
129
  ### Framework versions
130
  - Distily 0.2.0
logs/attn_loss_fn=cos, attn_weight=2.0/events.out.tfevents.1723746666.93d6cbb3ad53 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7940ec36db9fed3dedc0789df447a26c9ae3729895d19b712f4a4daf40f825ed
3
+ size 529