diff --git a/attention_kindselective_n_heads4_seed1339/args.json b/attention_kindselective_n_heads4_seed1339/args.json index 53efd56f3d3eb8c2b7965f35954a107c1064f19a..5bb6836cab970f8ce35eff1d86aacc714283dcde 100644 --- a/attention_kindselective_n_heads4_seed1339/args.json +++ b/attention_kindselective_n_heads4_seed1339/args.json @@ -1 +1 @@ -{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads4_seed1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 4, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 50000, "warmup_steps": 200, "group": "wider_is_better_7", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1339, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "none", "selection_head_linear_combo_scale": 1.0, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": true, "use_mini_model": false, "upload_to_hf": true, "seq_len": 256, "batch_size": 40, "total_batch_size": 10240, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": null, "mask_layernorm": false, "residual_attention_masks": false, "compute_base_shapes": false, "base_shapes_savefile": null, "mup": true, "disable_selection": false, "mup_enable_coord_check_logging": false, "max_lr": 0.0001, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "10e-5_10240_4_1339", "n_embd": 256} \ No newline at end of file +{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads4_seed1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 4, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 50000, "warmup_steps": 200, "group": "wider_is_better_7", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1339, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "none", "selection_head_linear_combo_scale": 1.0, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": true, "use_mini_model": false, "upload_to_hf": true, "seq_len": 256, "batch_size": 40, "total_batch_size": 10240, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": null, "mask_layernorm": false, "residual_attention_masks": false, "compute_base_shapes": false, "base_shapes_savefile": null, "mup": true, "disable_selection": false, "mup_enable_coord_check_logging": false, "max_lr": 2e-05, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "2e-5_10240_4_1339", "n_embd": 256} \ No newline at end of file diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_12500.pt b/attention_kindselective_n_heads4_seed1339/dataloader_12500.pt new file mode 100644 index 0000000000000000000000000000000000000000..f63806424a0177a7f2d678c2c63138219ed021f3 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_12500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab3779d33c2e0a7873fcd8c39402e44260740665950323ad1445480ec339965a +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_15000.pt b/attention_kindselective_n_heads4_seed1339/dataloader_15000.pt new file mode 100644 index 0000000000000000000000000000000000000000..0ea0f5a1bfab75667c4ebb0ca01b358cdc836a54 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_15000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:450020c7c306c73e5c07c463518ab937102d657515ea5a38da6f2e7291f20324 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_17500.pt b/attention_kindselective_n_heads4_seed1339/dataloader_17500.pt new file mode 100644 index 0000000000000000000000000000000000000000..eb9392348b1209c827e3e376b05eeda80e779aa8 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_17500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0696a655d7c8a9d7d275c7489f74a5a948ee029ac3941b045d6abaf12544a5b1 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_20000.pt b/attention_kindselective_n_heads4_seed1339/dataloader_20000.pt new file mode 100644 index 0000000000000000000000000000000000000000..ba32fca059aec40e4f758de96bdac6df23b9d9f5 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_20000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4eb226bdcee777fd1ca493533704dae226c077ef79c842fc9dc59a534d5381c1 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_22500.pt b/attention_kindselective_n_heads4_seed1339/dataloader_22500.pt new file mode 100644 index 0000000000000000000000000000000000000000..498dc444f528d893090328a5bd1e2f37da46dc12 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_22500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88833bfd821adac4edd1dd0772083ae007c7b8d33041f66e53a679e1fa8993e0 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_25000.pt b/attention_kindselective_n_heads4_seed1339/dataloader_25000.pt new file mode 100644 index 0000000000000000000000000000000000000000..b657bc134192d0ea956f984c289d0c682979a1f4 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_25000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:554555a425bac43d626c36f1c81c2b0aba51eda3281dab27a9cb56b61f413354 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_27500.pt b/attention_kindselective_n_heads4_seed1339/dataloader_27500.pt new file mode 100644 index 0000000000000000000000000000000000000000..d92d43d89390714f43db4f0782e49af0145b4a90 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_27500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a52940b3b45414e6cdbec0dbaeba848f52d681c2daf78c269027057332d7fbd +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_30000.pt b/attention_kindselective_n_heads4_seed1339/dataloader_30000.pt new file mode 100644 index 0000000000000000000000000000000000000000..06856f53253b6c8cbefd7d595d9b9b7266b22621 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_30000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:775519ea26122cb70d533c496bcdbbc19f759f3d096e83e98cca1dc10275fe8e +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_32500.pt b/attention_kindselective_n_heads4_seed1339/dataloader_32500.pt new file mode 100644 index 0000000000000000000000000000000000000000..dc0b4b40e57f41aa1046b3bd2697256635160c09 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_32500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab3422c8205fe45210246fed3dd6c317b75df02228cd8b75fba669574ce3b2d9 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_35000.pt b/attention_kindselective_n_heads4_seed1339/dataloader_35000.pt new file mode 100644 index 0000000000000000000000000000000000000000..62647b540fa7925361626b9f8dfa3959eebb7608 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_35000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:199ed3be67b88981f686112c97a2261729a37e0be3d4b0f4a289985a95d3cdf1 +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/dataloader_37500.pt b/attention_kindselective_n_heads4_seed1339/dataloader_37500.pt new file mode 100644 index 0000000000000000000000000000000000000000..475d8b538138a8e39b76a4cf04c8eaeac074d295 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/dataloader_37500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f77178b3be9dd3f8cd03c935236251f73fde6da7948ba9feda0c888fb8912dfe +size 964 diff --git a/attention_kindselective_n_heads4_seed1339/log2.txt b/attention_kindselective_n_heads4_seed1339/log2.txt index 00ac94bf8639caa8ea4b2dfd295ceccc271d6408..36d683e0dd5894f482d9c427980cb5c9062b376a 100644 --- a/attention_kindselective_n_heads4_seed1339/log2.txt +++ b/attention_kindselective_n_heads4_seed1339/log2.txt @@ -1,699 +1,3231 @@ max_steps: 50000 -38300 val loss 5.6784 -38300 val perplexity 292.4702 -38300 train 5.395936 (lr=1.0855e-05) (hash(x)=49878964) +33600 val loss 5.8398 +33600 val perplexity 343.7228 +33600 train 5.902711 (lr=4.8015e-06) (hash(x)=53759672) +200 val loss 9.2151 +200 val perplexity 10047.5400 +200 train 9.297607 (lr=3.0000e-05) (hash(x)=51276369) 0 val loss 11.2766 0 val perplexity 78951.9062 -38400 val loss 5.6700 -38400 val perplexity 290.0423 -38400 train 5.604312 (lr=1.0760e-05) (hash(x)=52855313) -38500 val loss 5.6699 -38500 val perplexity 289.9914 -38500 train 5.250736 (lr=1.0666e-05) (hash(x)=47213252) -0 train 11.297240 (lr=5.0000e-07) (hash(x)=55241167) -38600 val loss 5.6694 -38600 val perplexity 289.8528 -38600 train 5.830087 (lr=1.0572e-05) (hash(x)=50001498) -100 val loss 9.4564 -100 val perplexity 12789.4766 -100 train 9.441436 (lr=5.0500e-05) (hash(x)=46387919) -38700 val loss 5.6726 -38700 val perplexity 290.7901 -38700 train 5.401573 (lr=1.0479e-05) (hash(x)=49541736) -200 val loss 7.8121 -200 val perplexity 2470.3337 -200 train 7.900826 (lr=1.0000e-04) (hash(x)=51276369) -38800 val loss 5.6685 -38800 val perplexity 289.6135 -38800 train 5.601834 (lr=1.0386e-05) (hash(x)=49712820) -300 val loss 7.6376 -300 val perplexity 2074.7930 -300 train 7.534567 (lr=9.9999e-05) (hash(x)=48572728) -38900 val loss 5.6613 -38900 val perplexity 287.5113 -38900 train 5.612003 (lr=1.0294e-05) (hash(x)=50913051) -400 val loss 7.5966 -400 val perplexity 1991.4926 -400 train 7.331833 (lr=9.9996e-05) (hash(x)=44943400) -39000 val loss 5.6624 -39000 val perplexity 287.8408 -39000 train 5.621553 (lr=1.0203e-05) (hash(x)=46110703) -500 val loss 7.5399 -500 val perplexity 1881.5571 -500 train 7.449239 (lr=9.9992e-05) (hash(x)=48434590) -39100 val loss 5.6597 -39100 val perplexity 287.0684 -39100 train 5.493328 (lr=1.0113e-05) (hash(x)=43887045) -600 val loss 7.4966 -600 val perplexity 1801.9823 -600 train 7.344361 (lr=9.9986e-05) (hash(x)=47442266) -39200 val loss 5.6595 -39200 val perplexity 287.0032 -39200 train 5.897973 (lr=1.0023e-05) (hash(x)=53102477) -700 val loss 7.4837 -700 val perplexity 1778.8489 -700 train 7.543903 (lr=9.9978e-05) (hash(x)=56495712) -39300 val loss 5.6569 -39300 val perplexity 286.2463 -39300 train 5.512180 (lr=9.9341e-06) (hash(x)=44607212) -800 val loss 7.4533 -800 val perplexity 1725.5752 -800 train 7.440802 (lr=9.9968e-05) (hash(x)=50093738) -39400 val loss 5.6587 -39400 val perplexity 286.7726 -39400 train 5.844805 (lr=9.8458e-06) (hash(x)=55054109) -900 val loss 7.4095 -900 val perplexity 1651.6809 -900 train 7.374321 (lr=9.9956e-05) (hash(x)=49436061) -39500 val loss 5.6560 -39500 val perplexity 286.0097 -39500 train 5.519652 (lr=9.7581e-06) (hash(x)=47529853) -1000 val loss 7.3944 -1000 val perplexity 1626.8689 -1000 train 7.331343 (lr=9.9943e-05) (hash(x)=48818282) -39600 val loss 5.6537 -39600 val perplexity 285.3348 -39600 train 5.425322 (lr=9.6712e-06) (hash(x)=50783698) -1100 val loss 7.3509 -1100 val perplexity 1557.6047 -1100 train 7.407344 (lr=9.9927e-05) (hash(x)=55387425) -39700 val loss 5.6570 -39700 val perplexity 286.2832 -39700 train 5.444355 (lr=9.5849e-06) (hash(x)=47104160) -1200 val loss 7.5253 -1200 val perplexity 1854.3712 -1200 train 7.421935 (lr=9.9910e-05) (hash(x)=48328378) -39800 val loss 5.6608 -39800 val perplexity 287.3673 -39800 train 5.592721 (lr=9.4994e-06) (hash(x)=50941478) -1300 val loss 7.3479 -1300 val perplexity 1553.0095 -1300 train 7.264929 (lr=9.9892e-05) (hash(x)=48956717) -39900 val loss 5.6565 -39900 val perplexity 286.1535 -39900 train 5.388857 (lr=9.4146e-06) (hash(x)=47374830) -1400 val loss 7.3611 -1400 val perplexity 1573.6216 -1400 train 7.274777 (lr=9.9871e-05) (hash(x)=49782571) -40000 val loss 5.6581 -40000 val perplexity 286.6104 -40000 train 5.779648 (lr=9.3305e-06) (hash(x)=56749414) -1500 val loss 7.3527 -1500 val perplexity 1560.3909 -1500 train 7.145610 (lr=9.9849e-05) (hash(x)=51081994) -40100 val loss 5.6620 -40100 val perplexity 287.7258 -40100 train 5.607458 (lr=9.2472e-06) (hash(x)=49256359) -1600 val loss 7.3263 -1600 val perplexity 1519.7659 -1600 train 7.390820 (lr=9.9825e-05) (hash(x)=48791204) -40200 val loss 5.6555 -40200 val perplexity 285.8730 -40200 train 5.576644 (lr=9.1646e-06) (hash(x)=47932192) -1700 val loss 7.2937 -1700 val perplexity 1471.0563 -1700 train 7.009460 (lr=9.9799e-05) (hash(x)=46425011) -40300 val loss 5.6568 -40300 val perplexity 286.2384 -40300 train 5.476855 (lr=9.0827e-06) (hash(x)=42473499) -1800 val loss 7.3241 -1800 val perplexity 1516.4382 -1800 train 6.985591 (lr=9.9771e-05) (hash(x)=44919141) -40400 val loss 5.6550 -40400 val perplexity 285.7186 -40400 train 5.456766 (lr=9.0015e-06) (hash(x)=50469946) -1900 val loss 7.2486 -1900 val perplexity 1406.1027 -1900 train 7.351365 (lr=9.9741e-05) (hash(x)=48299675) -40500 val loss 5.6501 -40500 val perplexity 284.3304 -40500 train 5.743055 (lr=8.9211e-06) (hash(x)=48818656) -2000 val loss 7.2173 -2000 val perplexity 1362.7544 -2000 train 8.040308 (lr=9.9710e-05) (hash(x)=61804797) -40600 val loss 5.6457 -40600 val perplexity 283.0702 -40600 train 5.488149 (lr=8.8414e-06) (hash(x)=45215748) -2100 val loss 7.2014 -2100 val perplexity 1341.2904 -2100 train 7.292386 (lr=9.9677e-05) (hash(x)=53638030) -40700 val loss 5.6458 -40700 val perplexity 283.0904 -40700 train 5.773078 (lr=8.7624e-06) (hash(x)=48057577) -2200 val loss 7.1685 -2200 val perplexity 1297.8553 -2200 train 7.388167 (lr=9.9642e-05) (hash(x)=56364593) -40800 val loss 5.6444 -40800 val perplexity 282.7088 -40800 train 5.605831 (lr=8.6842e-06) (hash(x)=50103136) -2300 val loss 7.1717 -2300 val perplexity 1302.0543 -2300 train 7.251419 (lr=9.9606e-05) (hash(x)=52945580) -40900 val loss 5.6433 -40900 val perplexity 282.3985 -40900 train 5.499018 (lr=8.6068e-06) (hash(x)=45764202) -2400 val loss 7.1734 -2400 val perplexity 1304.2727 -2400 train 7.251097 (lr=9.9567e-05) (hash(x)=54595211) -41000 val loss 5.6416 -41000 val perplexity 281.9000 -41000 train 5.758846 (lr=8.5301e-06) (hash(x)=50724130) -2500 val loss 7.1699 -2500 val perplexity 1299.7263 -2500 train 7.189177 (lr=9.9527e-05) (hash(x)=52323363) -41100 val loss 5.6407 -41100 val perplexity 281.6723 -41100 train 5.572755 (lr=8.4541e-06) (hash(x)=48739125) -2600 val loss 7.1373 -2600 val perplexity 1257.9912 -2600 train 6.959945 (lr=9.9485e-05) (hash(x)=42652707) -41200 val loss 5.6379 -41200 val perplexity 280.8844 -41200 train 5.594127 (lr=8.3789e-06) (hash(x)=47473129) -2700 val loss 7.1502 -2700 val perplexity 1274.4105 -2700 train 6.867459 (lr=9.9442e-05) (hash(x)=45687855) -41300 val loss 5.6379 -41300 val perplexity 280.8649 -41300 train 5.563573 (lr=8.3045e-06) (hash(x)=51032823) -2800 val loss 7.1220 -2800 val perplexity 1238.9615 -2800 train 7.126947 (lr=9.9396e-05) (hash(x)=49962830) -41400 val loss 5.6520 -41400 val perplexity 284.8671 -41400 train 5.698008 (lr=8.2308e-06) (hash(x)=44411781) -2900 val loss 7.1669 -2900 val perplexity 1295.7585 -2900 train 7.199690 (lr=9.9349e-05) (hash(x)=52958451) -41500 val loss 5.6380 -41500 val perplexity 280.8879 -41500 train 5.683953 (lr=8.1579e-06) (hash(x)=48630357) -3000 val loss 7.1216 -3000 val perplexity 1238.4033 -3000 train 7.069978 (lr=9.9300e-05) (hash(x)=48896990) -41600 val loss 5.6353 -41600 val perplexity 280.1527 -41600 train 5.446326 (lr=8.0858e-06) (hash(x)=47836869) -3100 val loss 7.1046 -3100 val perplexity 1217.5770 -3100 train 7.054947 (lr=9.9249e-05) (hash(x)=47064930) -41700 val loss 5.6361 -41700 val perplexity 280.3648 -41700 train 5.515561 (lr=8.0144e-06) (hash(x)=49143658) -3200 val loss 7.0866 -3200 val perplexity 1195.8274 -3200 train 6.860336 (lr=9.9197e-05) (hash(x)=45145921) -41800 val loss 5.6389 -41800 val perplexity 281.1574 -41800 train 5.532392 (lr=7.9438e-06) (hash(x)=50426446) -3300 val loss 7.0858 -3300 val perplexity 1194.9000 -3300 train 6.935767 (lr=9.9142e-05) (hash(x)=48085018) -41900 val loss 5.6362 -41900 val perplexity 280.3896 -41900 train 5.521365 (lr=7.8740e-06) (hash(x)=48134303) -3400 val loss 7.0704 -3400 val perplexity 1176.6060 -3400 train 7.080620 (lr=9.9086e-05) (hash(x)=49596053) -42000 val loss 5.6366 -42000 val perplexity 280.5153 -42000 train 5.442455 (lr=7.8050e-06) (hash(x)=48151525) -3500 val loss 7.0368 -3500 val perplexity 1137.7137 -3500 train 6.621675 (lr=9.9028e-05) (hash(x)=41135634) -42100 val loss 5.6404 -42100 val perplexity 281.5832 -42100 train 5.503111 (lr=7.7368e-06) (hash(x)=49984437) -3600 val loss 7.0387 -3600 val perplexity 1139.8794 -3600 train 7.045468 (lr=9.8969e-05) (hash(x)=54286330) -42200 val loss 5.6393 -42200 val perplexity 281.2754 -42200 train 5.540201 (lr=7.6693e-06) (hash(x)=49103965) -3700 val loss 7.0348 -3700 val perplexity 1135.4819 -3700 train 6.938444 (lr=9.8908e-05) (hash(x)=57061993) -42300 val loss 5.6361 -42300 val perplexity 280.3770 -42300 train 5.343730 (lr=7.6027e-06) (hash(x)=52134795) -3800 val loss 7.0378 -3800 val perplexity 1138.8237 -3800 train 6.616735 (lr=9.8845e-05) (hash(x)=46544884) -42400 val loss 5.6382 -42400 val perplexity 280.9585 -42400 train 5.463866 (lr=7.5368e-06) (hash(x)=47289074) -3900 val loss 7.0345 -3900 val perplexity 1135.1008 -3900 train 6.881027 (lr=9.8780e-05) (hash(x)=48937435) -42500 val loss 5.6371 -42500 val perplexity 280.6571 -42500 train 6.041627 (lr=7.4717e-06) (hash(x)=54213272) -4000 val loss 7.0398 -4000 val perplexity 1141.1144 -4000 train 6.977884 (lr=9.8713e-05) (hash(x)=54466186) -42600 val loss 5.6364 -42600 val perplexity 280.4386 -42600 train 5.344518 (lr=7.4074e-06) (hash(x)=49347326) -4100 val loss 7.0053 -4100 val perplexity 1102.4150 -4100 train 6.836554 (lr=9.8645e-05) (hash(x)=51079995) -42700 val loss 5.6367 -42700 val perplexity 280.5423 -42700 train 5.406415 (lr=7.3440e-06) (hash(x)=45762383) -4200 val loss 6.9973 -4200 val perplexity 1093.6879 -4200 train 7.157095 (lr=9.8575e-05) (hash(x)=56495581) -42800 val loss 5.6380 -42800 val perplexity 280.8899 -42800 train 5.719070 (lr=7.2813e-06) (hash(x)=50377506) -4300 val loss 6.9613 -4300 val perplexity 1055.0035 -4300 train 6.850510 (lr=9.8503e-05) (hash(x)=45095478) -42900 val loss 5.6290 -42900 val perplexity 278.3965 -42900 train 5.998972 (lr=7.2194e-06) (hash(x)=57332431) -4400 val loss 6.9297 -4400 val perplexity 1022.1531 -4400 train 6.888327 (lr=9.8430e-05) (hash(x)=49954028) -43000 val loss 5.6294 -43000 val perplexity 278.5089 -43000 train 5.674274 (lr=7.1583e-06) (hash(x)=50599951) -4500 val loss 6.9208 -4500 val perplexity 1013.0889 -4500 train 6.849075 (lr=9.8355e-05) (hash(x)=48975821) -43100 val loss 5.6275 -43100 val perplexity 277.9758 -43100 train 5.593292 (lr=7.0981e-06) (hash(x)=51139362) -4600 val loss 6.8843 -4600 val perplexity 976.8253 -4600 train 6.973327 (lr=9.8278e-05) (hash(x)=49293028) -43200 val loss 5.6260 -43200 val perplexity 277.5468 -43200 train 5.478598 (lr=7.0386e-06) (hash(x)=45759314) -4700 val loss 6.8725 -4700 val perplexity 965.3833 -4700 train 6.789083 (lr=9.8199e-05) (hash(x)=48077753) -43300 val loss 5.6245 -43300 val perplexity 277.1455 -43300 train 5.677057 (lr=6.9800e-06) (hash(x)=48251106) -4800 val loss 6.8344 -4800 val perplexity 929.3071 -4800 train 6.655324 (lr=9.8119e-05) (hash(x)=45037879) -43400 val loss 5.6273 -43400 val perplexity 277.9023 -43400 train 5.983092 (lr=6.9222e-06) (hash(x)=49941825) -4900 val loss 6.7952 -4900 val perplexity 893.5521 -4900 train 6.717526 (lr=9.8036e-05) (hash(x)=50320154) -5000 val loss 6.8113 -5000 val perplexity 908.0694 -43500 val loss 5.6248 -43500 val perplexity 277.2100 -43500 train 6.025346 (lr=6.8652e-06) (hash(x)=47282597) -5000 train 7.030609 (lr=9.7953e-05) (hash(x)=55024523) -5100 val loss 6.7562 -5100 val perplexity 859.3580 -5100 train 6.689296 (lr=9.7867e-05) (hash(x)=47570607) -43600 val loss 5.6236 -43600 val perplexity 276.8875 -43600 train 5.733875 (lr=6.8090e-06) (hash(x)=47311813) -5200 val loss 6.7385 -5200 val perplexity 844.3349 -5200 train 6.866223 (lr=9.7780e-05) (hash(x)=54311349) -43700 val loss 5.6253 -43700 val perplexity 277.3623 -43700 train 5.692357 (lr=6.7537e-06) (hash(x)=51739445) -5300 val loss 6.7112 -5300 val perplexity 821.5640 -5300 train 6.617697 (lr=9.7691e-05) (hash(x)=47178524) -43800 val loss 5.6239 -43800 val perplexity 276.9691 -43800 train 5.663078 (lr=6.6992e-06) (hash(x)=58597156) -5400 val loss 6.6875 -5400 val perplexity 802.3312 -5400 train 6.683461 (lr=9.7600e-05) (hash(x)=49425088) -43900 val loss 5.6227 -43900 val perplexity 276.6289 -43900 train 5.352713 (lr=6.6455e-06) (hash(x)=45351304) -5500 val loss 6.6716 -5500 val perplexity 789.6293 -5500 train 6.500800 (lr=9.7508e-05) (hash(x)=46383189) -44000 val loss 5.6236 -44000 val perplexity 276.8810 -44000 train 5.404914 (lr=6.5926e-06) (hash(x)=46603503) -5600 val loss 6.6700 -5600 val perplexity 788.3792 -5600 train 6.856142 (lr=9.7414e-05) (hash(x)=54084990) -44100 val loss 5.6214 -44100 val perplexity 276.2868 -44100 train 5.590413 (lr=6.5406e-06) (hash(x)=52189096) -5700 val loss 6.6305 -5700 val perplexity 757.8840 -5700 train 6.303418 (lr=9.7318e-05) (hash(x)=47384182) -44200 val loss 5.6225 -44200 val perplexity 276.5793 -44200 train 5.789032 (lr=6.4894e-06) (hash(x)=51527792) -5800 val loss 6.6221 -5800 val perplexity 751.5416 -5800 train 6.522151 (lr=9.7221e-05) (hash(x)=51683744) -44300 val loss 5.6208 -44300 val perplexity 276.1172 -44300 train 5.475013 (lr=6.4390e-06) (hash(x)=47982628) -5900 val loss 6.6169 -5900 val perplexity 747.5950 -5900 train 6.624623 (lr=9.7122e-05) (hash(x)=52798836) -44400 val loss 5.6230 -44400 val perplexity 276.7267 -44400 train 5.611701 (lr=6.3895e-06) (hash(x)=54902968) -6000 val loss 6.6217 -6000 val perplexity 751.2009 -6000 train 6.511095 (lr=9.7021e-05) (hash(x)=50945000) -44500 val loss 5.6205 -44500 val perplexity 276.0341 -44500 train 5.537436 (lr=6.3408e-06) (hash(x)=49008493) -6100 val loss 6.6186 -6100 val perplexity 748.9037 -6100 train 6.304089 (lr=9.6919e-05) (hash(x)=48964427) -44600 val loss 5.6222 -44600 val perplexity 276.4967 -44600 train 5.353870 (lr=6.2929e-06) (hash(x)=43643824) -6200 val loss 6.6284 -6200 val perplexity 756.2642 -6200 train 6.390724 (lr=9.6815e-05) (hash(x)=47148610) -44700 val loss 5.6218 -44700 val perplexity 276.3996 -44700 train 5.470081 (lr=6.2459e-06) (hash(x)=50465872) -6300 val loss 6.5986 -6300 val perplexity 734.0524 -6300 train 6.511916 (lr=9.6709e-05) (hash(x)=50860553) -44800 val loss 5.6218 -44800 val perplexity 276.3798 -44800 train 5.378158 (lr=6.1998e-06) (hash(x)=51962160) -6400 val loss 6.5951 -6400 val perplexity 731.4692 -6400 train 6.364479 (lr=9.6602e-05) (hash(x)=50553992) -44900 val loss 5.6235 -44900 val perplexity 276.8581 -44900 train 5.763238 (lr=6.1545e-06) (hash(x)=57370039) -6500 val loss 6.5914 -6500 val perplexity 728.8076 -6500 train 6.389678 (lr=9.6493e-05) (hash(x)=52521332) -45000 val loss 5.6243 -45000 val perplexity 277.0884 -45000 train 5.469142 (lr=6.1100e-06) (hash(x)=52648130) -6600 val loss 6.5748 -6600 val perplexity 716.8139 -6600 train 6.929626 (lr=9.6382e-05) (hash(x)=52609843) -45100 val loss 5.6554 -45100 val perplexity 285.8250 -45100 train 5.327322 (lr=6.0664e-06) (hash(x)=45155459) -6700 val loss 6.5304 -6700 val perplexity 685.6540 -6700 train 6.558313 (lr=9.6270e-05) (hash(x)=53632957) -45200 val loss 5.6170 -45200 val perplexity 275.0746 -45200 train 5.611628 (lr=6.0237e-06) (hash(x)=52304992) -6800 val loss 6.5309 -6800 val perplexity 685.9938 -6800 train 6.539653 (lr=9.6156e-05) (hash(x)=46927608) -45300 val loss 5.6149 -45300 val perplexity 274.4989 -45300 train 5.468201 (lr=5.9818e-06) (hash(x)=46456407) -6900 val loss 6.5383 -6900 val perplexity 691.1232 -6900 train 6.552355 (lr=9.6040e-05) (hash(x)=54483057) -45400 val loss 5.6139 -45400 val perplexity 274.2122 -45400 train 5.607281 (lr=5.9407e-06) (hash(x)=48325493) -7000 val loss 6.5042 -7000 val perplexity 667.9302 -7000 train 6.676917 (lr=9.5923e-05) (hash(x)=51493884) -45500 val loss 5.6116 -45500 val perplexity 273.5927 -45500 train 5.583129 (lr=5.9005e-06) (hash(x)=48869064) -7100 val loss 6.4774 -7100 val perplexity 650.3052 -7100 train 6.625372 (lr=9.5804e-05) (hash(x)=53303341) -7200 val loss 6.4821 -7200 val perplexity 653.3736 -7200 train 6.203289 (lr=9.5683e-05) (hash(x)=45272178) -45600 val loss 5.6123 -45600 val perplexity 273.7640 -45600 train 5.690610 (lr=5.8612e-06) (hash(x)=46077805) -7300 val loss 6.4693 -7300 val perplexity 645.0018 -7300 train 6.416029 (lr=9.5561e-05) (hash(x)=50389872) -45700 val loss 5.6112 -45700 val perplexity 273.4644 -45700 train 5.849724 (lr=5.8227e-06) (hash(x)=56831165) -7400 val loss 6.4705 -7400 val perplexity 645.8201 -7400 train 6.127152 (lr=9.5437e-05) (hash(x)=43796301) -45800 val loss 5.6116 -45800 val perplexity 273.5795 -45800 train 5.459206 (lr=5.7851e-06) (hash(x)=50124921) -7500 val loss 6.4612 -7500 val perplexity 639.8350 -7500 train 6.404037 (lr=9.5312e-05) (hash(x)=47808686) -45900 val loss 5.6117 -45900 val perplexity 273.6022 -45900 train 5.222868 (lr=5.7484e-06) (hash(x)=43227400) -7600 val loss 6.4687 -7600 val perplexity 644.6377 -7600 train 6.201755 (lr=9.5185e-05) (hash(x)=41936898) -46000 val loss 5.6104 -46000 val perplexity 273.2413 -46000 train 5.602647 (lr=5.7125e-06) (hash(x)=58308887) -7700 val loss 6.5138 -7700 val perplexity 674.3773 -7700 train 6.732550 (lr=9.5057e-05) (hash(x)=57550318) -46100 val loss 5.6094 -46100 val perplexity 272.9798 -46100 train 5.677024 (lr=5.6775e-06) (hash(x)=48571632) -7800 val loss 6.4561 -7800 val perplexity 636.5423 -7800 train 6.328799 (lr=9.4926e-05) (hash(x)=47485210) -46200 val loss 5.6103 -46200 val perplexity 273.2153 -46200 train 5.874604 (lr=5.6434e-06) (hash(x)=46368862) -7900 val loss 6.4544 -7900 val perplexity 635.5160 -7900 train 6.812044 (lr=9.4795e-05) (hash(x)=53228688) -46300 val loss 5.6123 -46300 val perplexity 273.7760 -46300 train 5.376794 (lr=5.6101e-06) (hash(x)=52820317) -8000 val loss 6.4555 -8000 val perplexity 636.1951 -8000 train 6.513896 (lr=9.4661e-05) (hash(x)=52018673) -46400 val loss 5.6107 -46400 val perplexity 273.3253 -46400 train 5.410422 (lr=5.5777e-06) (hash(x)=51871487) -8100 val loss 6.4554 -8100 val perplexity 636.1124 -8100 train 6.303450 (lr=9.4526e-05) (hash(x)=47079349) -46500 val loss 5.6121 -46500 val perplexity 273.7071 -46500 train 5.565622 (lr=5.5462e-06) (hash(x)=47867973) -8200 val loss 6.4537 -8200 val perplexity 635.0416 -8200 train 6.513351 (lr=9.4390e-05) (hash(x)=57921563) -46600 val loss 5.6095 -46600 val perplexity 272.9956 -46600 train 5.271016 (lr=5.5156e-06) (hash(x)=50685713) -8300 val loss 6.4452 -8300 val perplexity 629.6801 -8300 train 5.972620 (lr=9.4252e-05) (hash(x)=45038933) -46700 val loss 5.6092 -46700 val perplexity 272.9158 -46700 train 5.732937 (lr=5.4858e-06) (hash(x)=50975526) -8400 val loss 6.4808 -8400 val perplexity 652.5134 -8400 train 6.312977 (lr=9.4112e-05) (hash(x)=47763246) -46800 val loss 5.6107 -46800 val perplexity 273.3414 -46800 train 5.750267 (lr=5.4569e-06) (hash(x)=51012129) -8500 val loss 6.4612 -8500 val perplexity 639.8493 -8500 train 6.493088 (lr=9.3971e-05) (hash(x)=56176595) -46900 val loss 5.6143 -46900 val perplexity 274.3083 -46900 train 5.333468 (lr=5.4289e-06) (hash(x)=51205353) -8600 val loss 6.4510 -8600 val perplexity 633.3370 -8600 train 6.529971 (lr=9.3828e-05) (hash(x)=55184249) -47000 val loss 5.6131 -47000 val perplexity 274.0013 -47000 train 5.759924 (lr=5.4017e-06) (hash(x)=48939661) -8700 val loss 6.4552 -8700 val perplexity 636.0110 -8700 train 6.379639 (lr=9.3684e-05) (hash(x)=46471646) -47100 val loss 5.6096 -47100 val perplexity 273.0318 -47100 train 5.531120 (lr=5.3755e-06) (hash(x)=51002557) -8800 val loss 6.4945 -8800 val perplexity 661.4907 -8800 train 6.198891 (lr=9.3538e-05) (hash(x)=46233162) -47200 val loss 5.6103 -47200 val perplexity 273.2244 -47200 train 5.444582 (lr=5.3501e-06) (hash(x)=43422209) -8900 val loss 6.5133 -8900 val perplexity 674.0529 -8900 train 6.432613 (lr=9.3391e-05) (hash(x)=47233684) -47300 val loss 5.6093 -47300 val perplexity 272.9432 -47300 train 5.365450 (lr=5.3256e-06) (hash(x)=45870552) -9000 val loss 6.4790 -9000 val perplexity 651.3448 -9000 train 6.380252 (lr=9.3242e-05) (hash(x)=48374529) -47400 val loss 5.6096 -47400 val perplexity 273.0215 -47400 train 5.453257 (lr=5.3020e-06) (hash(x)=47001810) -9100 val loss 6.4612 -9100 val perplexity 639.8325 -9100 train 6.426319 (lr=9.3092e-05) (hash(x)=48065371) -47500 val loss 5.6062 -47500 val perplexity 272.1136 -47500 train 5.676814 (lr=5.2792e-06) (hash(x)=49146619) -47600 val loss 5.6047 -47600 val perplexity 271.6933 -47600 train 5.688634 (lr=5.2574e-06) (hash(x)=53742853) -9200 val loss 6.4337 -9200 val perplexity 622.4777 -9200 train 6.365408 (lr=9.2940e-05) (hash(x)=47408078) -47700 val loss 5.6023 -47700 val perplexity 271.0357 -47700 train 5.858260 (lr=5.2364e-06) (hash(x)=53019458) -9300 val loss 6.4346 -9300 val perplexity 623.0532 -9300 train 6.420694 (lr=9.2786e-05) (hash(x)=50749781) -47800 val loss 5.6015 -47800 val perplexity 270.8320 -47800 train 5.688690 (lr=5.2163e-06) (hash(x)=54497914) -47900 val loss 5.6040 -47900 val perplexity 271.5063 -47900 train 5.530357 (lr=5.1972e-06) (hash(x)=47109012) -9400 val loss 6.4372 -9400 val perplexity 624.6458 -9400 train 6.634754 (lr=9.2632e-05) (hash(x)=48560169) -48000 val loss 5.6005 -48000 val perplexity 270.5672 -48000 train 5.590864 (lr=5.1788e-06) (hash(x)=48429816) -9500 val loss 6.4296 -9500 val perplexity 619.9107 -9500 train 6.521097 (lr=9.2475e-05) (hash(x)=50936392) -48100 val loss 5.6008 -48100 val perplexity 270.6499 -48100 train 5.544402 (lr=5.1614e-06) (hash(x)=46500465) -48200 val loss 5.5982 -48200 val perplexity 269.9372 -48200 train 5.762817 (lr=5.1449e-06) (hash(x)=50888195) -9600 val loss 6.4141 -9600 val perplexity 610.3770 -9600 train 6.430016 (lr=9.2317e-05) (hash(x)=50651714) -48300 val loss 5.5988 -48300 val perplexity 270.1088 -48300 train 5.666016 (lr=5.1293e-06) (hash(x)=56406586) -9700 val loss 6.4110 -9700 val perplexity 608.4774 -9700 train 6.487106 (lr=9.2158e-05) (hash(x)=47311384) -48400 val loss 5.5989 -48400 val perplexity 270.1341 -48400 train 5.751607 (lr=5.1145e-06) (hash(x)=48713969) -48500 val loss 5.5980 -48500 val perplexity 269.8947 -48500 train 5.026878 (lr=5.1007e-06) (hash(x)=37691696) -9800 val loss 6.4147 -9800 val perplexity 610.7438 -9800 train 6.526601 (lr=9.1997e-05) (hash(x)=50921139) -48600 val loss 5.5988 -48600 val perplexity 270.1123 -48600 train 5.892519 (lr=5.0877e-06) (hash(x)=48712321) -48700 val loss 5.5987 -48700 val perplexity 270.0880 -9900 val loss 6.4657 -9900 val perplexity 642.6958 -48700 train 6.032611 (lr=5.0756e-06) (hash(x)=63513096) -9900 train 6.607420 (lr=9.1835e-05) (hash(x)=48142455) -48800 val loss 5.6003 -48800 val perplexity 270.4991 -48800 train 5.218384 (lr=5.0644e-06) (hash(x)=44917721) -10000 val loss 6.4405 -10000 val perplexity 626.7271 -10000 train 6.596187 (lr=9.1671e-05) (hash(x)=50757167) -48900 val loss 5.5981 -48900 val perplexity 269.9094 -48900 train 5.816957 (lr=5.0542e-06) (hash(x)=51426559) -49000 val loss 5.5992 -49000 val perplexity 270.2160 -49000 train 5.508775 (lr=5.0448e-06) (hash(x)=47909690) -10100 val loss 6.4674 -10100 val perplexity 643.8083 -10100 train 6.442811 (lr=9.1506e-05) (hash(x)=46128585) -49100 val loss 5.5971 -49100 val perplexity 269.6422 -49100 train 5.500440 (lr=5.0363e-06) (hash(x)=48329226) -10200 val loss 6.4787 -10200 val perplexity 651.1212 -10200 train 6.703011 (lr=9.1339e-05) (hash(x)=57016296) -49200 val loss 5.5991 -49200 val perplexity 270.1703 -49200 train 5.600343 (lr=5.0286e-06) (hash(x)=49512472) -49300 val loss 5.5989 -49300 val perplexity 270.1354 -49300 train 5.452957 (lr=5.0219e-06) (hash(x)=46541146) -10300 val loss 6.5369 -10300 val perplexity 690.1228 -10300 train 6.600325 (lr=9.1171e-05) (hash(x)=49839464) -49400 val loss 5.6009 -49400 val perplexity 270.6745 -49400 train 5.543798 (lr=5.0161e-06) (hash(x)=52192792) -10400 val loss 6.5206 -10400 val perplexity 678.9719 -10400 train 6.516890 (lr=9.1001e-05) (hash(x)=50237852) -49500 val loss 5.5995 -49500 val perplexity 270.2928 -49500 train 5.386661 (lr=5.0112e-06) (hash(x)=48954331) -49600 val loss 5.5985 -49600 val perplexity 270.0147 -49600 train 5.711846 (lr=5.0072e-06) (hash(x)=50364098) -10500 val loss 6.5185 -10500 val perplexity 677.5903 -10500 train 6.626068 (lr=9.0830e-05) (hash(x)=48003282) -49700 val loss 5.5968 -49700 val perplexity 269.5677 -49700 train 5.554048 (lr=5.0040e-06) (hash(x)=52016774) -49800 val loss 5.5979 -49800 val perplexity 269.8479 -49800 train 5.830290 (lr=5.0018e-06) (hash(x)=54182957) -10600 val loss 6.5319 -10600 val perplexity 686.7331 -10600 train 6.466997 (lr=9.0658e-05) (hash(x)=53088427) -49900 val loss 5.5940 -49900 val perplexity 268.8109 -49900 train 5.332826 (lr=5.0004e-06) (hash(x)=48188126) -10700 val loss 6.5500 -10700 val perplexity 699.2339 -10700 train 6.378216 (lr=9.0484e-05) (hash(x)=47687508) -49999 val loss 5.6180 -49999 val perplexity 275.3266 -10800 val loss 6.5394 -10800 val perplexity 691.8988 -10800 train 6.397013 (lr=9.0308e-05) (hash(x)=47658698) -10900 val loss 6.4993 -10900 val perplexity 664.6454 -10900 train 6.503762 (lr=9.0132e-05) (hash(x)=52957012) -11000 val loss 6.4866 -11000 val perplexity 656.2906 -11000 train 6.625940 (lr=8.9954e-05) (hash(x)=54632498) -11100 val loss 6.5099 -11100 val perplexity 671.7436 -11100 train 6.986665 (lr=8.9774e-05) (hash(x)=54028593) -11200 val loss 6.4837 -11200 val perplexity 654.3682 -11200 train 6.616082 (lr=8.9593e-05) (hash(x)=54084412) -11300 val loss 6.4871 -11300 val perplexity 656.6074 -11300 train 6.309471 (lr=8.9411e-05) (hash(x)=48170961) -11400 val loss 6.4986 -11400 val perplexity 664.2282 -11400 train 6.540001 (lr=8.9227e-05) (hash(x)=52124243) +33700 val loss 5.8359 +33700 val perplexity 342.3787 +33700 train 6.057962 (lr=4.7650e-06) (hash(x)=50177499) +300 val loss 8.1089 +300 val perplexity 3323.7583 +300 train 8.052432 (lr=3.0000e-05) (hash(x)=48572728) +33800 val loss 5.8378 +33800 val perplexity 343.0172 +33800 train 5.699886 (lr=4.7286e-06) (hash(x)=49208351) +400 val loss 7.7868 +400 val perplexity 2408.5815 +400 train 7.547433 (lr=2.9999e-05) (hash(x)=44943400) +0 train 11.297240 (lr=1.0000e-07) (hash(x)=55241167) +33900 val loss 5.8364 +33900 val perplexity 342.5365 +33900 train 5.650459 (lr=4.6923e-06) (hash(x)=51136014) +500 val loss 7.6491 +500 val perplexity 2098.8386 +100 val loss 10.0906 +100 val perplexity 24114.7969 +500 train 7.568335 (lr=2.9998e-05) (hash(x)=48434590) +100 train 10.095575 (lr=1.0100e-05) (hash(x)=46387919) +34000 val loss 5.8330 +34000 val perplexity 341.3656 +34000 train 5.645186 (lr=4.6562e-06) (hash(x)=55323465) +200 val loss 9.5906 +200 val perplexity 14626.1689 +200 train 9.670161 (lr=2.0000e-05) (hash(x)=51276369) +600 val loss 7.6004 +600 val perplexity 1999.0105 +600 train 7.467304 (lr=2.9996e-05) (hash(x)=47442266) +34100 val loss 5.8293 +34100 val perplexity 340.1234 +34100 train 5.920761 (lr=4.6202e-06) (hash(x)=46512090) +300 val loss 8.8083 +300 val perplexity 6689.8057 +300 train 8.772751 (lr=2.0000e-05) (hash(x)=48572728) +700 val loss 7.5740 +700 val perplexity 1946.9456 +700 train 7.634149 (lr=2.9993e-05) (hash(x)=56495712) +34200 val loss 5.8297 +34200 val perplexity 340.2686 +34200 train 5.896380 (lr=4.5844e-06) (hash(x)=52513966) +400 val loss 8.1677 +400 val perplexity 3525.3738 +400 train 7.954683 (lr=1.9999e-05) (hash(x)=44943400) +800 val loss 7.5566 +800 val perplexity 1913.4150 +800 train 7.555591 (lr=2.9990e-05) (hash(x)=50093738) +34300 val loss 5.8241 +34300 val perplexity 338.3555 +34300 train 5.833245 (lr=4.5487e-06) (hash(x)=51181125) +500 val loss 7.9014 +500 val perplexity 2700.9519 +500 train 7.835149 (lr=1.9998e-05) (hash(x)=48434590) +900 val loss 7.5339 +900 val perplexity 1870.3042 +900 train 7.498625 (lr=2.9987e-05) (hash(x)=49436061) +600 val loss 7.7636 +600 val perplexity 2353.4207 +600 train 7.659897 (lr=1.9997e-05) (hash(x)=47442266) +34400 val loss 5.8230 +34400 val perplexity 337.9906 +34400 train 5.968364 (lr=4.5132e-06) (hash(x)=50979117) +1000 val loss 7.5171 +1000 val perplexity 1839.2045 +1000 train 7.444939 (lr=2.9983e-05) (hash(x)=48818282) +700 val loss 7.6947 +700 val perplexity 2196.6680 +700 train 7.735430 (lr=1.9996e-05) (hash(x)=56495712) +34500 val loss 5.8348 +34500 val perplexity 341.9822 +34500 train 5.924613 (lr=4.4778e-06) (hash(x)=50394150) +800 val loss 7.6471 +800 val perplexity 2094.5276 +800 train 7.650714 (lr=1.9994e-05) (hash(x)=50093738) +1100 val loss 7.4899 +1100 val perplexity 1789.7965 +1100 train 7.576001 (lr=2.9978e-05) (hash(x)=55387425) +34600 val loss 5.8228 +34600 val perplexity 337.9297 +34600 train 5.740018 (lr=4.4426e-06) (hash(x)=52666346) +900 val loss 7.6148 +900 val perplexity 2027.9049 +900 train 7.589414 (lr=1.9991e-05) (hash(x)=49436061) +1200 val loss 7.4727 +1200 val perplexity 1759.4178 +1200 train 7.390953 (lr=2.9973e-05) (hash(x)=48328378) +34700 val loss 5.8207 +34700 val perplexity 337.2155 +34700 train 6.059371 (lr=4.4075e-06) (hash(x)=48275110) +1000 val loss 7.5952 +1000 val perplexity 1988.5869 +1000 train 7.534963 (lr=1.9989e-05) (hash(x)=48818282) +1300 val loss 7.4452 +1300 val perplexity 1711.5704 +1300 train 7.370911 (lr=2.9968e-05) (hash(x)=48956717) +34800 val loss 5.8212 +34800 val perplexity 337.3668 +34800 train 5.616254 (lr=4.3725e-06) (hash(x)=43282226) +1100 val loss 7.5792 +1100 val perplexity 1957.0305 +1100 train 7.667435 (lr=1.9985e-05) (hash(x)=55387425) +1400 val loss 7.4414 +1400 val perplexity 1705.1912 +1400 train 7.337984 (lr=2.9961e-05) (hash(x)=49782571) +34900 val loss 5.8189 +34900 val perplexity 336.5891 +34900 train 5.993647 (lr=4.3378e-06) (hash(x)=55252392) +1200 val loss 7.5610 +1200 val perplexity 1921.7184 +1200 train 7.517821 (lr=1.9982e-05) (hash(x)=48328378) +1500 val loss 7.4276 +1500 val perplexity 1681.6942 +1500 train 7.234368 (lr=2.9955e-05) (hash(x)=51081994) +35000 val loss 5.8184 +35000 val perplexity 336.4492 +35000 train 5.822789 (lr=4.3031e-06) (hash(x)=51882979) +1300 val loss 7.5452 +1300 val perplexity 1891.6453 +1300 train 7.486642 (lr=1.9978e-05) (hash(x)=48956717) +1600 val loss 7.4041 +1600 val perplexity 1642.6790 +1600 train 7.488369 (lr=2.9947e-05) (hash(x)=48791204) +35100 val loss 5.8176 +35100 val perplexity 336.1750 +35100 train 5.607735 (lr=4.2687e-06) (hash(x)=56053572) +1400 val loss 7.5371 +1400 val perplexity 1876.4707 +1400 train 7.447511 (lr=1.9974e-05) (hash(x)=49782571) +1500 val loss 7.5300 +1500 val perplexity 1863.0446 +1500 train 7.357488 (lr=1.9970e-05) (hash(x)=51081994) +1700 val loss 7.3872 +1700 val perplexity 1615.1201 +1700 train 7.097632 (lr=2.9940e-05) (hash(x)=46425011) +35200 val loss 5.8154 +35200 val perplexity 335.4419 +35200 train 5.900864 (lr=4.2344e-06) (hash(x)=51979525) +1600 val loss 7.5170 +1600 val perplexity 1839.1036 +1600 train 7.606800 (lr=1.9965e-05) (hash(x)=48791204) +1800 val loss 7.3742 +1800 val perplexity 1594.3560 +1800 train 7.004421 (lr=2.9931e-05) (hash(x)=44919141) +35300 val loss 5.8159 +35300 val perplexity 335.6072 +35300 train 5.821369 (lr=4.2002e-06) (hash(x)=50873666) +1700 val loss 7.5092 +1700 val perplexity 1824.7133 +1700 train 7.239395 (lr=1.9960e-05) (hash(x)=46425011) +35400 val loss 5.8190 +35400 val perplexity 336.6331 +35400 train 6.430342 (lr=4.1662e-06) (hash(x)=69087970) +1900 val loss 7.3463 +1900 val perplexity 1550.4406 +1900 train 7.457504 (lr=2.9922e-05) (hash(x)=48299675) +1800 val loss 7.4960 +1800 val perplexity 1800.7626 +1800 train 7.168148 (lr=1.9954e-05) (hash(x)=44919141) +35500 val loss 5.8163 +35500 val perplexity 335.7176 +35500 train 5.542082 (lr=4.1324e-06) (hash(x)=48654918) +2000 val loss 7.3123 +2000 val perplexity 1498.5682 +2000 train 8.124590 (lr=2.9913e-05) (hash(x)=61804797) +1900 val loss 7.4810 +1900 val perplexity 1774.0857 +1900 train 7.582687 (lr=1.9948e-05) (hash(x)=48299675) +35600 val loss 5.8122 +35600 val perplexity 334.3397 +35600 train 5.544744 (lr=4.0988e-06) (hash(x)=49987504) +2100 val loss 7.2817 +2100 val perplexity 1453.4210 +2100 train 7.371889 (lr=2.9903e-05) (hash(x)=53638030) +2000 val loss 7.4572 +2000 val perplexity 1732.2786 +2000 train 8.312654 (lr=1.9942e-05) (hash(x)=61804797) +35700 val loss 5.8128 +35700 val perplexity 334.5605 +35700 train 5.531723 (lr=4.0653e-06) (hash(x)=46372898) +2100 val loss 7.4426 +2100 val perplexity 1707.1665 +2100 train 7.532056 (lr=1.9935e-05) (hash(x)=53638030) +2200 val loss 7.2525 +2200 val perplexity 1411.6283 +2200 train 7.515120 (lr=2.9893e-05) (hash(x)=56364593) +35800 val loss 5.8134 +35800 val perplexity 334.7676 +35800 train 5.844424 (lr=4.0319e-06) (hash(x)=49996578) +2200 val loss 7.4275 +2200 val perplexity 1681.6797 +2200 train 7.687351 (lr=1.9928e-05) (hash(x)=56364593) +2300 val loss 7.2451 +2300 val perplexity 1401.1873 +2300 train 7.386978 (lr=2.9882e-05) (hash(x)=52945580) +2300 val loss 7.4150 +2300 val perplexity 1660.7234 +2300 train 7.550599 (lr=1.9921e-05) (hash(x)=52945580) +35900 val loss 5.8160 +35900 val perplexity 335.6321 +35900 train 5.864781 (lr=3.9988e-06) (hash(x)=49932960) +2400 val loss 7.2141 +2400 val perplexity 1358.4562 +2400 train 7.272038 (lr=2.9870e-05) (hash(x)=54595211) +2400 val loss 7.3976 +2400 val perplexity 1632.1005 +2400 train 7.465549 (lr=1.9913e-05) (hash(x)=54595211) +36000 val loss 5.8170 +36000 val perplexity 335.9565 +36000 train 5.560926 (lr=3.9658e-06) (hash(x)=41502642) +2500 val loss 7.2012 +2500 val perplexity 1341.0461 +2500 train 7.218399 (lr=2.9858e-05) (hash(x)=52323363) +2500 val loss 7.3902 +2500 val perplexity 1620.0651 +2500 train 7.420754 (lr=1.9905e-05) (hash(x)=52323363) +36100 val loss 5.8145 +36100 val perplexity 335.1194 +36100 train 5.684670 (lr=3.9330e-06) (hash(x)=55380312) +2600 val loss 7.1796 +2600 val perplexity 1312.3779 +2600 train 6.924984 (lr=2.9846e-05) (hash(x)=42652707) +2600 val loss 7.3798 +2600 val perplexity 1603.3353 +2600 train 7.155263 (lr=1.9897e-05) (hash(x)=42652707) +36200 val loss 5.8142 +36200 val perplexity 335.0327 +36200 train 5.532636 (lr=3.9003e-06) (hash(x)=51323559) +2700 val loss 7.1601 +2700 val perplexity 1287.0636 +2700 train 6.833584 (lr=2.9832e-05) (hash(x)=45687855) +2700 val loss 7.3715 +2700 val perplexity 1589.9473 +2700 train 7.065453 (lr=1.9888e-05) (hash(x)=45687855) +36300 val loss 5.8167 +36300 val perplexity 335.8526 +36300 train 5.738066 (lr=3.8678e-06) (hash(x)=47058693) +2800 val loss 7.3589 +2800 val perplexity 1570.1243 +2800 train 7.352308 (lr=1.9879e-05) (hash(x)=49962830) +2800 val loss 7.1396 +2800 val perplexity 1260.9327 +2800 train 7.136701 (lr=2.9819e-05) (hash(x)=49962830) +36400 val loss 5.8143 +36400 val perplexity 335.0504 +36400 train 5.401338 (lr=3.8355e-06) (hash(x)=45928242) +2900 val loss 7.3447 +2900 val perplexity 1548.0123 +2900 train 7.423705 (lr=1.9870e-05) (hash(x)=52958451) +2900 val loss 7.1169 +2900 val perplexity 1232.5786 +2900 train 7.170207 (lr=2.9805e-05) (hash(x)=52958451) +36500 val loss 5.8067 +36500 val perplexity 332.5169 +36500 train 5.894205 (lr=3.8034e-06) (hash(x)=52834624) +3000 val loss 7.3295 +3000 val perplexity 1524.6101 +3000 train 7.270491 (lr=1.9860e-05) (hash(x)=48896990) +3000 val loss 7.0877 +3000 val perplexity 1197.1145 +3000 train 7.031383 (lr=2.9790e-05) (hash(x)=48896990) +36600 val loss 5.8065 +36600 val perplexity 332.4544 +36600 train 5.957966 (lr=3.7715e-06) (hash(x)=47195270) +3100 val loss 7.3149 +3100 val perplexity 1502.5178 +3100 train 7.264447 (lr=1.9850e-05) (hash(x)=47064930) +3100 val loss 7.0670 +3100 val perplexity 1172.6707 +3100 train 7.028967 (lr=2.9775e-05) (hash(x)=47064930) +3200 val loss 7.2988 +3200 val perplexity 1478.4625 +3200 train 7.046788 (lr=1.9839e-05) (hash(x)=45145921) +36700 val loss 5.8061 +36700 val perplexity 332.3221 +36700 train 5.879849 (lr=3.7397e-06) (hash(x)=52501734) +3200 val loss 7.0466 +3200 val perplexity 1148.9368 +3200 train 6.750253 (lr=2.9759e-05) (hash(x)=45145921) +3300 val loss 7.2850 +3300 val perplexity 1458.2157 +3300 train 7.155090 (lr=1.9828e-05) (hash(x)=48085018) +36800 val loss 5.8022 +36800 val perplexity 331.0301 +36800 train 5.870245 (lr=3.7081e-06) (hash(x)=50333721) +3400 val loss 7.2729 +3400 val perplexity 1440.7467 +3400 train 7.300655 (lr=1.9817e-05) (hash(x)=49596053) +3300 val loss 7.0280 +3300 val perplexity 1127.7700 +3300 train 6.882772 (lr=2.9743e-05) (hash(x)=48085018) +36900 val loss 5.8026 +36900 val perplexity 331.1578 +36900 train 5.697065 (lr=3.6767e-06) (hash(x)=46841481) +3500 val loss 7.2600 +3500 val perplexity 1422.2412 +3500 train 6.824724 (lr=1.9806e-05) (hash(x)=41135634) +3400 val loss 7.0033 +3400 val perplexity 1100.3044 +3400 train 7.017133 (lr=2.9726e-05) (hash(x)=49596053) +37000 val loss 5.7998 +37000 val perplexity 330.2393 +37000 train 5.810065 (lr=3.6455e-06) (hash(x)=61838433) +3600 val loss 7.2484 +3600 val perplexity 1405.9210 +3600 train 7.262741 (lr=1.9794e-05) (hash(x)=54286330) +3500 val loss 6.9882 +3500 val perplexity 1083.7484 +3500 train 6.565709 (lr=2.9709e-05) (hash(x)=41135634) +37100 val loss 5.7992 +37100 val perplexity 330.0240 +37100 train 5.751407 (lr=3.6144e-06) (hash(x)=47537252) +3700 val loss 7.2370 +3700 val perplexity 1389.9325 +3700 train 7.162038 (lr=1.9782e-05) (hash(x)=57061993) +3600 val loss 6.9675 +3600 val perplexity 1061.5562 +3600 train 6.973201 (lr=2.9691e-05) (hash(x)=54286330) +37200 val loss 5.7976 +37200 val perplexity 329.5115 +37200 train 5.930354 (lr=3.5836e-06) (hash(x)=51840280) +3800 val loss 7.2328 +3800 val perplexity 1384.1268 +3800 train 6.832160 (lr=1.9769e-05) (hash(x)=46544884) +3700 val loss 6.9464 +3700 val perplexity 1039.3790 +3700 train 6.857440 (lr=2.9672e-05) (hash(x)=57061993) +37300 val loss 5.7975 +37300 val perplexity 329.4584 +37300 train 5.625167 (lr=3.5529e-06) (hash(x)=48870226) +3900 val loss 7.2190 +3900 val perplexity 1365.1732 +3900 train 7.089311 (lr=1.9756e-05) (hash(x)=48937435) +4000 val loss 7.2168 +4000 val perplexity 1362.0677 +4000 train 7.269147 (lr=1.9743e-05) (hash(x)=54466186) +37400 val loss 5.7968 +37400 val perplexity 329.2310 +3800 val loss 6.9304 +3800 val perplexity 1022.8522 +37400 train 5.740648 (lr=3.5224e-06) (hash(x)=45398282) +3800 train 6.501587 (lr=2.9653e-05) (hash(x)=46544884) +4100 val loss 7.1981 +4100 val perplexity 1336.8820 +4100 train 7.055575 (lr=1.9729e-05) (hash(x)=51079995) +37500 val loss 5.7964 +37500 val perplexity 329.1070 +3900 val loss 6.9080 +3900 val perplexity 1000.2896 +3900 train 6.761992 (lr=2.9634e-05) (hash(x)=48937435) +37500 train 5.828522 (lr=3.4921e-06) (hash(x)=46278516) +4200 val loss 7.1890 +4200 val perplexity 1324.7817 +4200 train 7.372856 (lr=1.9715e-05) (hash(x)=56495581) +37600 val loss 5.7965 +37600 val perplexity 329.1561 +37600 train 5.740655 (lr=3.4620e-06) (hash(x)=49553278) +4000 val loss 6.8934 +4000 val perplexity 985.7018 +4000 train 6.972066 (lr=2.9614e-05) (hash(x)=54466186) +4300 val loss 7.1735 +4300 val perplexity 1304.4320 +4300 train 7.035427 (lr=1.9701e-05) (hash(x)=45095478) +37700 val loss 5.7978 +37700 val perplexity 329.5857 +37700 train 6.067697 (lr=3.4321e-06) (hash(x)=61653685) +4100 val loss 6.8611 +4100 val perplexity 954.4407 +4100 train 6.686050 (lr=2.9593e-05) (hash(x)=51079995) +4400 val loss 7.1649 +4400 val perplexity 1293.2438 +4400 train 7.096426 (lr=1.9686e-05) (hash(x)=49954028) +37800 val loss 5.7950 +37800 val perplexity 328.6425 +37800 train 5.944009 (lr=3.4024e-06) (hash(x)=58942903) +4500 val loss 7.1583 +4500 val perplexity 1284.7354 +4500 train 7.085289 (lr=1.9671e-05) (hash(x)=48975821) +4200 val loss 6.8441 +4200 val perplexity 938.3320 +4200 train 7.024286 (lr=2.9572e-05) (hash(x)=56495581) +4600 val loss 7.1407 +4600 val perplexity 1262.2863 +4600 train 7.201608 (lr=1.9656e-05) (hash(x)=49293028) +37900 val loss 5.7956 +37900 val perplexity 328.8648 +37900 train 5.790917 (lr=3.3728e-06) (hash(x)=51529837) +4300 val loss 6.7990 +4300 val perplexity 896.9570 +4300 train 6.645703 (lr=2.9551e-05) (hash(x)=45095478) +4700 val loss 7.1330 +4700 val perplexity 1252.6864 +4700 train 7.070783 (lr=1.9640e-05) (hash(x)=48077753) +38000 val loss 5.7981 +38000 val perplexity 329.6606 +38000 train 5.539510 (lr=3.3435e-06) (hash(x)=51047484) +4400 val loss 6.7719 +4400 val perplexity 872.9525 +4400 train 6.745496 (lr=2.9529e-05) (hash(x)=49954028) +4800 val loss 7.1196 +4800 val perplexity 1235.9380 +4800 train 6.941299 (lr=1.9624e-05) (hash(x)=45037879) +38100 val loss 5.7935 +38100 val perplexity 328.1628 +38100 train 5.710298 (lr=3.3143e-06) (hash(x)=49117671) +4500 val loss 6.7479 +4500 val perplexity 852.2253 +4500 train 6.657944 (lr=2.9506e-05) (hash(x)=48975821) +4900 val loss 7.1146 +4900 val perplexity 1229.7860 +4900 train 7.020064 (lr=1.9607e-05) (hash(x)=50320154) +38200 val loss 5.7948 +38200 val perplexity 328.6006 +38200 train 5.764774 (lr=3.2854e-06) (hash(x)=49501130) +5000 val loss 7.1089 +5000 val perplexity 1222.8322 +5000 train 7.310444 (lr=1.9591e-05) (hash(x)=55024523) +4600 val loss 6.7148 +4600 val perplexity 824.5152 +4600 train 6.792094 (lr=2.9483e-05) (hash(x)=49293028) +38300 val loss 5.7972 +38300 val perplexity 329.3913 +38300 train 5.543228 (lr=3.2566e-06) (hash(x)=49878964) +5100 val loss 7.0949 +5100 val perplexity 1205.7651 +5100 train 7.014016 (lr=1.9573e-05) (hash(x)=47570607) +4700 val loss 6.7072 +4700 val perplexity 818.2806 +4700 train 6.633926 (lr=2.9460e-05) (hash(x)=48077753) +38400 val loss 5.7951 +38400 val perplexity 328.7002 +38400 train 5.766051 (lr=3.2281e-06) (hash(x)=52855313) +5200 val loss 7.0852 +5200 val perplexity 1194.1630 +5200 train 7.183423 (lr=1.9556e-05) (hash(x)=54311349) +4800 val loss 6.6812 +4800 val perplexity 797.2798 +4800 train 6.528817 (lr=2.9436e-05) (hash(x)=45037879) +5300 val loss 7.0796 +5300 val perplexity 1187.4440 +5300 train 6.981455 (lr=1.9538e-05) (hash(x)=47178524) +38500 val loss 5.7955 +38500 val perplexity 328.8162 +38500 train 5.388383 (lr=3.1997e-06) (hash(x)=47213252) +4900 val loss 6.6659 +4900 val perplexity 785.1442 +4900 train 6.586228 (lr=2.9411e-05) (hash(x)=50320154) +5400 val loss 7.0597 +5400 val perplexity 1164.0609 +5400 train 7.063051 (lr=1.9520e-05) (hash(x)=49425088) +38600 val loss 5.7935 +38600 val perplexity 328.1527 +38600 train 5.920391 (lr=3.1716e-06) (hash(x)=50001498) +5000 val loss 6.6530 +5000 val perplexity 775.1246 +5500 val loss 7.0488 +5500 val perplexity 1151.4272 +5500 train 6.889327 (lr=1.9502e-05) (hash(x)=46383189) +5000 train 6.892908 (lr=2.9386e-05) (hash(x)=55024523) +38700 val loss 5.7957 +38700 val perplexity 328.8871 +38700 train 5.534037 (lr=3.1436e-06) (hash(x)=49541736) +5600 val loss 7.0516 +5600 val perplexity 1154.6696 +5600 train 7.266815 (lr=1.9483e-05) (hash(x)=54084990) +5100 val loss 6.6195 +5100 val perplexity 749.6032 +5100 train 6.538449 (lr=2.9360e-05) (hash(x)=47570607) +38800 val loss 5.7923 +38800 val perplexity 327.7642 +38800 train 5.717169 (lr=3.1159e-06) (hash(x)=49712820) +5700 val loss 7.0366 +5700 val perplexity 1137.5190 +5700 train 6.739789 (lr=1.9464e-05) (hash(x)=47384182) +5200 val loss 6.6087 +5200 val perplexity 741.4822 +5200 train 6.744793 (lr=2.9334e-05) (hash(x)=54311349) +38900 val loss 5.7883 +38900 val perplexity 326.4608 +38900 train 5.729952 (lr=3.0883e-06) (hash(x)=50913051) +5800 val loss 7.0238 +5800 val perplexity 1123.0829 +5800 train 6.915489 (lr=1.9444e-05) (hash(x)=51683744) +5300 val loss 6.5907 +5300 val perplexity 728.2642 +5300 train 6.476836 (lr=2.9307e-05) (hash(x)=47178524) +39000 val loss 5.7872 +39000 val perplexity 326.1024 +39000 train 5.743138 (lr=3.0610e-06) (hash(x)=46110703) +5900 val loss 7.0134 +5900 val perplexity 1111.4440 +5900 train 7.015720 (lr=1.9424e-05) (hash(x)=52798836) +5400 val loss 6.5806 +5400 val perplexity 720.9445 +5400 train 6.593421 (lr=2.9280e-05) (hash(x)=49425088) +6000 val loss 7.0065 +6000 val perplexity 1103.8094 +6000 train 6.897607 (lr=1.9404e-05) (hash(x)=50945000) +39100 val loss 5.7852 +39100 val perplexity 325.4557 +39100 train 5.631864 (lr=3.0339e-06) (hash(x)=43887045) +6100 val loss 7.0006 +6100 val perplexity 1097.3037 +6100 train 6.742208 (lr=1.9384e-05) (hash(x)=48964427) +5500 val loss 6.5640 +5500 val perplexity 709.0792 +5500 train 6.384793 (lr=2.9252e-05) (hash(x)=46383189) +39200 val loss 5.7857 +39200 val perplexity 325.6196 +39200 train 6.004618 (lr=3.0069e-06) (hash(x)=53102477) +6200 val loss 6.9905 +6200 val perplexity 1086.2888 +6200 train 6.784522 (lr=1.9363e-05) (hash(x)=47148610) +5600 val loss 6.5647 +5600 val perplexity 709.6157 +5600 train 6.768056 (lr=2.9224e-05) (hash(x)=54084990) +39300 val loss 5.7845 +39300 val perplexity 325.2097 +39300 train 5.627706 (lr=2.9802e-06) (hash(x)=44607212) +6300 val loss 6.9838 +6300 val perplexity 1078.9619 +6300 train 6.915902 (lr=1.9342e-05) (hash(x)=50860553) +5700 val loss 6.5415 +5700 val perplexity 693.2911 +5700 train 6.221973 (lr=2.9196e-05) (hash(x)=47384182) +39400 val loss 5.7861 +39400 val perplexity 325.7243 +39400 train 5.963056 (lr=2.9537e-06) (hash(x)=55054109) +6400 val loss 6.9679 +6400 val perplexity 1061.9789 +6400 train 6.763657 (lr=1.9320e-05) (hash(x)=50553992) +5800 val loss 6.5270 +5800 val perplexity 683.3751 +5800 train 6.407080 (lr=2.9166e-05) (hash(x)=51683744) +39500 val loss 5.7840 +39500 val perplexity 325.0728 +39500 train 5.619909 (lr=2.9274e-06) (hash(x)=47529853) +6500 val loss 6.9552 +6500 val perplexity 1048.6045 +6500 train 6.804325 (lr=1.9299e-05) (hash(x)=52521332) +5900 val loss 6.5141 +5900 val perplexity 674.6008 +6600 val loss 6.9297 +6600 val perplexity 1022.1730 +5900 train 6.519782 (lr=2.9137e-05) (hash(x)=52798836) +6600 train 7.288078 (lr=1.9276e-05) (hash(x)=52609843) +39600 val loss 5.7836 +39600 val perplexity 324.9316 +39600 train 5.552695 (lr=2.9014e-06) (hash(x)=50783698) +6700 val loss 6.9078 +6700 val perplexity 1000.0559 +6700 train 6.977970 (lr=1.9254e-05) (hash(x)=53632957) +39700 val loss 5.7852 +39700 val perplexity 325.4482 +39700 train 5.562631 (lr=2.8755e-06) (hash(x)=47104160) +6000 val loss 6.5176 +6000 val perplexity 676.9708 +6000 train 6.384697 (lr=2.9106e-05) (hash(x)=50945000) +6800 val loss 6.8954 +6800 val perplexity 987.7189 +6800 train 6.930350 (lr=1.9231e-05) (hash(x)=46927608) +39800 val loss 5.7853 +39800 val perplexity 325.4706 +39800 train 5.739317 (lr=2.8498e-06) (hash(x)=50941478) +6100 val loss 6.5074 +6100 val perplexity 670.0612 +6100 train 6.192907 (lr=2.9076e-05) (hash(x)=48964427) +6900 val loss 6.8779 +6900 val perplexity 970.5731 +6900 train 6.932907 (lr=1.9208e-05) (hash(x)=54483057) +39900 val loss 5.7848 +39900 val perplexity 325.3125 +39900 train 5.545476 (lr=2.8244e-06) (hash(x)=47374830) +6200 val loss 6.5034 +6200 val perplexity 667.3871 +6200 train 6.262025 (lr=2.9044e-05) (hash(x)=47148610) +7000 val loss 6.8638 +7000 val perplexity 957.0306 +7000 train 7.053998 (lr=1.9185e-05) (hash(x)=51493884) +40000 val loss 5.7867 +40000 val perplexity 325.9514 +40000 train 5.935668 (lr=2.7992e-06) (hash(x)=56749414) +6300 val loss 6.4829 +6300 val perplexity 653.8392 +6300 train 6.405093 (lr=2.9013e-05) (hash(x)=50860553) +7100 val loss 6.8460 +7100 val perplexity 940.0978 +7100 train 7.001909 (lr=1.9161e-05) (hash(x)=53303341) +7200 val loss 6.8421 +7200 val perplexity 936.4824 +40100 val loss 5.7858 +40100 val perplexity 325.6400 +7200 train 6.517679 (lr=1.9137e-05) (hash(x)=45272178) +40100 train 5.720423 (lr=2.7742e-06) (hash(x)=49256359) +6400 val loss 6.4767 +6400 val perplexity 649.8254 +6400 train 6.259035 (lr=2.8981e-05) (hash(x)=50553992) +7300 val loss 6.8211 +7300 val perplexity 916.9527 +7300 train 6.767941 (lr=1.9112e-05) (hash(x)=50389872) +40200 val loss 5.7827 +40200 val perplexity 324.6291 +40200 train 5.715927 (lr=2.7494e-06) (hash(x)=47932192) +6500 val loss 6.4723 +6500 val perplexity 646.9954 +6500 train 6.288477 (lr=2.8948e-05) (hash(x)=52521332) +7400 val loss 6.8030 +7400 val perplexity 900.5517 +7400 train 6.427528 (lr=1.9087e-05) (hash(x)=43796301) +40300 val loss 5.7832 +40300 val perplexity 324.7965 +40300 train 5.581563 (lr=2.7248e-06) (hash(x)=42473499) +6600 val loss 6.4459 +6600 val perplexity 630.1279 +6600 train 6.795638 (lr=2.8915e-05) (hash(x)=52609843) +7500 val loss 6.7910 +7500 val perplexity 889.7838 +7500 train 6.738329 (lr=1.9062e-05) (hash(x)=47808686) +40400 val loss 5.7844 +40400 val perplexity 325.1807 +40400 train 5.580541 (lr=2.7004e-06) (hash(x)=50469946) +6700 val loss 6.4205 +6700 val perplexity 614.2811 +6700 train 6.469879 (lr=2.8881e-05) (hash(x)=53632957) +7600 val loss 6.7730 +7600 val perplexity 873.9321 +7600 train 6.473551 (lr=1.9037e-05) (hash(x)=41936898) +40500 val loss 5.7795 +40500 val perplexity 323.5975 +40500 train 5.882727 (lr=2.6763e-06) (hash(x)=48818656) +7700 val loss 6.7735 +7700 val perplexity 874.3864 +7700 train 7.002127 (lr=1.9011e-05) (hash(x)=57550318) +6800 val loss 6.4169 +6800 val perplexity 612.1063 +6800 train 6.440267 (lr=2.8847e-05) (hash(x)=46927608) +40600 val loss 5.7761 +40600 val perplexity 322.4850 +40600 train 5.621453 (lr=2.6524e-06) (hash(x)=45215748) +7800 val loss 6.7512 +7800 val perplexity 855.0763 +7800 train 6.634337 (lr=1.8985e-05) (hash(x)=47485210) +6900 val loss 6.3975 +6900 val perplexity 600.3157 +6900 train 6.420102 (lr=2.8812e-05) (hash(x)=54483057) +40700 val loss 5.7749 +40700 val perplexity 322.1227 +40700 train 5.890401 (lr=2.6287e-06) (hash(x)=48057577) +7900 val loss 6.7360 +7900 val perplexity 842.1660 +7900 train 7.099861 (lr=1.8959e-05) (hash(x)=53228688) +7000 val loss 6.3860 +7000 val perplexity 593.4590 +7000 train 6.540714 (lr=2.8777e-05) (hash(x)=51493884) +8000 val loss 6.7151 +8000 val perplexity 824.7649 +8000 train 6.780042 (lr=1.8932e-05) (hash(x)=52018673) +40800 val loss 5.7748 +40800 val perplexity 322.0649 +40800 train 5.730608 (lr=2.6053e-06) (hash(x)=50103136) +7100 val loss 6.3767 +7100 val perplexity 588.0135 +7100 train 6.511652 (lr=2.8741e-05) (hash(x)=53303341) +8100 val loss 6.7021 +8100 val perplexity 814.0945 +8100 train 6.506268 (lr=1.8905e-05) (hash(x)=47079349) +40900 val loss 5.7735 +40900 val perplexity 321.6584 +40900 train 5.624897 (lr=2.5820e-06) (hash(x)=45764202) +7200 val loss 6.3850 +7200 val perplexity 592.9035 +7200 train 6.072328 (lr=2.8705e-05) (hash(x)=45272178) +8200 val loss 6.6916 +8200 val perplexity 805.6308 +8200 train 6.746307 (lr=1.8878e-05) (hash(x)=57921563) +41000 val loss 5.7715 +41000 val perplexity 321.0153 +41000 train 5.896519 (lr=2.5590e-06) (hash(x)=50724130) +8300 val loss 6.6734 +8300 val perplexity 791.0953 +8300 train 6.213883 (lr=1.8850e-05) (hash(x)=45038933) +7300 val loss 6.3726 +7300 val perplexity 585.5887 +7300 train 6.310030 (lr=2.8668e-05) (hash(x)=50389872) +41100 val loss 5.7715 +41100 val perplexity 321.0335 +41100 train 5.706398 (lr=2.5362e-06) (hash(x)=48739125) +8400 val loss 6.6707 +8400 val perplexity 788.9225 +8400 train 6.543269 (lr=1.8822e-05) (hash(x)=47763246) +7400 val loss 6.3553 +7400 val perplexity 575.5070 +7400 train 5.994948 (lr=2.8631e-05) (hash(x)=43796301) +41200 val loss 5.7710 +41200 val perplexity 320.8729 +41200 train 5.710758 (lr=2.5137e-06) (hash(x)=47473129) +8500 val loss 6.6524 +8500 val perplexity 774.6360 +8500 train 6.687089 (lr=1.8794e-05) (hash(x)=56176595) +7500 val loss 6.3446 +7500 val perplexity 569.4368 +7500 train 6.272412 (lr=2.8594e-05) (hash(x)=47808686) +41300 val loss 5.7707 +41300 val perplexity 320.7626 +41300 train 5.709377 (lr=2.4913e-06) (hash(x)=51032823) +8600 val loss 6.6403 +8600 val perplexity 765.3503 +8600 train 6.707767 (lr=1.8766e-05) (hash(x)=55184249) +7600 val loss 6.3387 +7600 val perplexity 566.0871 +7600 train 6.058947 (lr=2.8556e-05) (hash(x)=41936898) +41400 val loss 5.7742 +41400 val perplexity 321.8810 +41400 train 5.778426 (lr=2.4692e-06) (hash(x)=44411781) +8700 val loss 6.6323 +8700 val perplexity 759.2610 +8700 train 6.604059 (lr=1.8737e-05) (hash(x)=46471646) +7700 val loss 6.3445 +7700 val perplexity 569.3470 +7700 train 6.587870 (lr=2.8517e-05) (hash(x)=57550318) +8800 val loss 6.6213 +8800 val perplexity 750.8896 +8800 train 6.343414 (lr=1.8708e-05) (hash(x)=46233162) +41500 val loss 5.7695 +41500 val perplexity 320.3860 +41500 train 5.783124 (lr=2.4474e-06) (hash(x)=48630357) +8900 val loss 6.6093 +8900 val perplexity 741.9526 +8900 train 6.505754 (lr=1.8678e-05) (hash(x)=47233684) +7800 val loss 6.3351 +7800 val perplexity 564.0251 +7800 train 6.191846 (lr=2.8478e-05) (hash(x)=47485210) +41600 val loss 5.7687 +41600 val perplexity 320.1079 +41600 train 5.588441 (lr=2.4257e-06) (hash(x)=47836869) +9000 val loss 6.5952 +9000 val perplexity 731.5745 +9000 train 6.508955 (lr=1.8648e-05) (hash(x)=48374529) +7900 val loss 6.3239 +7900 val perplexity 557.7404 +7900 train 6.699908 (lr=2.8438e-05) (hash(x)=53228688) +41700 val loss 5.7687 +41700 val perplexity 320.1276 +41700 train 5.665323 (lr=2.4043e-06) (hash(x)=49143658) +9100 val loss 6.5730 +9100 val perplexity 715.5148 +9100 train 6.541916 (lr=1.8618e-05) (hash(x)=48065371) +41800 val loss 5.7688 +41800 val perplexity 320.1568 +8000 val loss 6.3096 +8000 val perplexity 549.8417 +41800 train 5.673004 (lr=2.3832e-06) (hash(x)=50426446) +8000 train 6.366764 (lr=2.8398e-05) (hash(x)=52018673) +9200 val loss 6.5654 +9200 val perplexity 710.0919 +9200 train 6.549735 (lr=1.8588e-05) (hash(x)=47408078) +41900 val loss 5.7695 +41900 val perplexity 320.3637 +41900 train 5.658653 (lr=2.3622e-06) (hash(x)=48134303) +8100 val loss 6.3074 +8100 val perplexity 548.6074 +8100 train 6.122249 (lr=2.8358e-05) (hash(x)=47079349) +9300 val loss 6.5449 +9300 val perplexity 695.6824 +9300 train 6.548703 (lr=1.8557e-05) (hash(x)=50749781) +9400 val loss 6.5308 +9400 val perplexity 685.9705 +9400 train 6.719339 (lr=1.8526e-05) (hash(x)=48560169) +42000 val loss 5.7688 +42000 val perplexity 320.1577 +42000 train 5.573346 (lr=2.3415e-06) (hash(x)=48151525) +8200 val loss 6.2916 +8200 val perplexity 540.0425 +8200 train 6.349628 (lr=2.8317e-05) (hash(x)=57921563) +9500 val loss 6.5206 +9500 val perplexity 679.0118 +9500 train 6.603598 (lr=1.8495e-05) (hash(x)=50936392) +42100 val loss 5.7701 +42100 val perplexity 320.5750 +42100 train 5.636930 (lr=2.3210e-06) (hash(x)=49984437) +8300 val loss 6.2932 +8300 val perplexity 540.8981 +8300 train 5.816945 (lr=2.8276e-05) (hash(x)=45038933) +9600 val loss 6.5109 +9600 val perplexity 672.4102 +9600 train 6.522950 (lr=1.8463e-05) (hash(x)=50651714) +42200 val loss 5.7700 +42200 val perplexity 320.5257 +42200 train 5.678516 (lr=2.3008e-06) (hash(x)=49103965) +8400 val loss 6.3042 +8400 val perplexity 546.8452 +8400 train 6.171069 (lr=2.8234e-05) (hash(x)=47763246) +9700 val loss 6.5084 +9700 val perplexity 670.7731 +9700 train 6.576026 (lr=1.8432e-05) (hash(x)=47311384) +42300 val loss 5.7697 +42300 val perplexity 320.4535 +42300 train 5.515322 (lr=2.2808e-06) (hash(x)=52134795) +8500 val loss 6.2846 +8500 val perplexity 536.2499 +8500 train 6.372394 (lr=2.8191e-05) (hash(x)=56176595) +9800 val loss 6.4974 +9800 val perplexity 663.4011 +9800 train 6.597688 (lr=1.8399e-05) (hash(x)=50921139) +42400 val loss 5.7712 +42400 val perplexity 320.9329 +42400 train 5.581810 (lr=2.2610e-06) (hash(x)=47289074) +8600 val loss 6.2855 +8600 val perplexity 536.7562 +8600 train 6.381540 (lr=2.8149e-05) (hash(x)=55184249) +9900 val loss 6.5219 +9900 val perplexity 679.8723 +9900 train 6.616925 (lr=1.8367e-05) (hash(x)=48142455) +42500 val loss 5.7702 +42500 val perplexity 320.5929 +42500 train 6.167914 (lr=2.2415e-06) (hash(x)=54213272) +10000 val loss 6.4810 +10000 val perplexity 652.6176 +10000 train 6.606616 (lr=1.8334e-05) (hash(x)=50757167) +8700 val loss 6.2763 +8700 val perplexity 531.8396 +8700 train 6.214106 (lr=2.8105e-05) (hash(x)=46471646) +42600 val loss 5.7688 +42600 val perplexity 320.1549 +42600 train 5.480625 (lr=2.2222e-06) (hash(x)=49347326) +10100 val loss 6.4756 +10100 val perplexity 649.0920 +10100 train 6.451486 (lr=1.8301e-05) (hash(x)=46128585) +8800 val loss 6.2701 +8800 val perplexity 528.5235 +8800 train 5.967916 (lr=2.8062e-05) (hash(x)=46233162) +42700 val loss 5.7696 +42700 val perplexity 320.4168 +10200 val loss 6.4720 +10200 val perplexity 646.7810 +42700 train 5.547189 (lr=2.2032e-06) (hash(x)=45762383) +10200 train 6.728794 (lr=1.8268e-05) (hash(x)=57016296) +8900 val loss 6.2747 +8900 val perplexity 530.9890 +8900 train 6.181352 (lr=2.8017e-05) (hash(x)=47233684) +10300 val loss 6.4666 +10300 val perplexity 643.2651 +10300 train 6.542679 (lr=1.8234e-05) (hash(x)=49839464) +42800 val loss 5.7686 +42800 val perplexity 320.0826 +42800 train 5.833604 (lr=2.1844e-06) (hash(x)=50377506) +9000 val loss 6.2607 +9000 val perplexity 523.5752 +9000 train 6.137102 (lr=2.7973e-05) (hash(x)=48374529) +10400 val loss 6.4571 +10400 val perplexity 637.2253 +10400 train 6.469795 (lr=1.8200e-05) (hash(x)=50237852) +9100 val loss 6.2385 +9100 val perplexity 512.0798 +9100 train 6.242193 (lr=2.7928e-05) (hash(x)=48065371) +10500 val loss 6.4489 +10500 val perplexity 632.0123 +10500 train 6.597267 (lr=1.8166e-05) (hash(x)=48003282) +9200 val loss 6.2368 +9200 val perplexity 511.2420 +9200 train 6.242299 (lr=2.7882e-05) (hash(x)=47408078) +10600 val loss 6.4427 +10600 val perplexity 628.1267 +10600 train 6.398111 (lr=1.8132e-05) (hash(x)=53088427) +9300 val loss 6.2152 +9300 val perplexity 500.2896 +9300 train 6.192973 (lr=2.7836e-05) (hash(x)=50749781) +9400 val loss 6.2071 +9400 val perplexity 496.2725 +9400 train 6.426061 (lr=2.7789e-05) (hash(x)=48560169) +10700 val loss 6.4355 +10700 val perplexity 623.5706 +10700 train 6.243379 (lr=1.8097e-05) (hash(x)=47687508) +9500 val loss 6.2001 +9500 val perplexity 492.7877 +9500 train 6.341508 (lr=2.7743e-05) (hash(x)=50936392) +10800 val loss 6.4282 +10800 val perplexity 619.0659 +10800 train 6.280035 (lr=1.8062e-05) (hash(x)=47658698) +9600 val loss 6.2042 +9600 val perplexity 494.8195 +9600 train 6.216803 (lr=2.7695e-05) (hash(x)=50651714) +10900 val loss 6.4205 +10900 val perplexity 614.3039 +10900 train 6.407661 (lr=1.8026e-05) (hash(x)=52957012) +9700 val loss 6.1897 +9700 val perplexity 487.6891 +9700 train 6.289503 (lr=2.7647e-05) (hash(x)=47311384) +11000 val loss 6.4105 +11000 val perplexity 608.2001 +11000 train 6.551227 (lr=1.7991e-05) (hash(x)=54632498) +9800 val loss 6.1763 +9800 val perplexity 481.2321 +9800 train 6.309625 (lr=2.7599e-05) (hash(x)=50921139) +9900 val loss 6.2239 +9900 val perplexity 504.6846 +9900 train 6.350302 (lr=2.7550e-05) (hash(x)=48142455) +11100 val loss 6.3966 +11100 val perplexity 599.8021 +11100 train 6.895236 (lr=1.7955e-05) (hash(x)=54028593) +10000 val loss 6.1756 +10000 val perplexity 480.8712 +10000 train 6.340471 (lr=2.7501e-05) (hash(x)=50757167) +11200 val loss 6.3918 +11200 val perplexity 596.9574 +11200 train 6.542569 (lr=1.7919e-05) (hash(x)=54084412) +10100 val loss 6.1706 +10100 val perplexity 478.4886 +10100 train 6.169375 (lr=2.7452e-05) (hash(x)=46128585) +11300 val loss 6.3862 +11300 val perplexity 593.6195 +11300 train 6.175961 (lr=1.7882e-05) (hash(x)=48170961) +11400 val loss 6.3808 +11400 val perplexity 590.3752 +11400 train 6.366016 (lr=1.7845e-05) (hash(x)=52124243) +10200 val loss 6.1620 +10200 val perplexity 474.3784 +10200 train 6.415475 (lr=2.7402e-05) (hash(x)=57016296) +11500 val loss 6.3720 +11500 val perplexity 585.2554 +11500 train 6.305048 (lr=1.7809e-05) (hash(x)=46985517) +10300 val loss 6.1599 +10300 val perplexity 473.4041 +10300 train 6.253690 (lr=2.7351e-05) (hash(x)=49839464) +11600 val loss 6.3632 +11600 val perplexity 580.1172 +11600 train 6.110388 (lr=1.7771e-05) (hash(x)=46741028) +10400 val loss 6.1502 +10400 val perplexity 468.8151 +10400 train 6.197786 (lr=2.7300e-05) (hash(x)=50237852) +11700 val loss 6.3580 +11700 val perplexity 577.0950 +11700 train 6.245870 (lr=1.7734e-05) (hash(x)=45729773) +10500 val loss 6.1400 +10500 val perplexity 464.0623 +10500 train 6.272721 (lr=2.7249e-05) (hash(x)=48003282) +11800 val loss 6.3496 +11800 val perplexity 572.2514 +11800 train 6.316463 (lr=1.7696e-05) (hash(x)=50540350) +10600 val loss 6.1378 +10600 val perplexity 463.0401 +10600 train 6.085890 (lr=2.7197e-05) (hash(x)=53088427) +11900 val loss 6.3411 +11900 val perplexity 567.4111 +11900 train 6.274208 (lr=1.7658e-05) (hash(x)=50313941) +10700 val loss 6.1356 +10700 val perplexity 462.0251 +10700 train 5.950803 (lr=2.7145e-05) (hash(x)=47687508) +12000 val loss 6.3365 +12000 val perplexity 564.8201 +12000 train 6.264954 (lr=1.7619e-05) (hash(x)=49621720) +10800 val loss 6.1311 +10800 val perplexity 459.9437 +10800 train 6.005104 (lr=2.7093e-05) (hash(x)=47658698) +12100 val loss 6.3317 +12100 val perplexity 562.1218 +12100 train 6.222105 (lr=1.7581e-05) (hash(x)=48903010) +10900 val loss 6.1185 +10900 val perplexity 454.1880 +10900 train 6.146273 (lr=2.7040e-05) (hash(x)=52957012) +12200 val loss 6.3268 +12200 val perplexity 559.3723 +12200 train 6.281783 (lr=1.7542e-05) (hash(x)=51643030) +12300 val loss 6.3216 +12300 val perplexity 556.4411 +12300 train 6.273603 (lr=1.7503e-05) (hash(x)=48863143) +11000 val loss 6.1150 +11000 val perplexity 452.6028 +11000 train 6.223407 (lr=2.6986e-05) (hash(x)=54632498) +12400 val loss 6.3165 +12400 val perplexity 553.6301 +12400 train 6.194585 (lr=1.7464e-05) (hash(x)=53251144) +11100 val loss 6.1119 +11100 val perplexity 451.1766 +11100 train 6.439250 (lr=2.6932e-05) (hash(x)=54028593) +12500 val loss 6.3150 +12500 val perplexity 552.7852 +12500 train 6.088791 (lr=1.7424e-05) (hash(x)=45178705) +11200 val loss 6.1040 +11200 val perplexity 447.6534 +11200 train 6.284250 (lr=2.6878e-05) (hash(x)=54084412) +12600 val loss 6.3478 +12600 val perplexity 571.2446 +12600 train 6.336140 (lr=1.7384e-05) (hash(x)=52779738) +11300 val loss 6.1023 +11300 val perplexity 446.8621 +11300 train 5.915199 (lr=2.6823e-05) (hash(x)=48170961) +12700 val loss 6.3010 +12700 val perplexity 545.1324 +12700 train 6.153219 (lr=1.7344e-05) (hash(x)=50253395) +11400 val loss 6.0927 +11400 val perplexity 442.5963 +11400 train 6.083128 (lr=2.6768e-05) (hash(x)=52124243) +12800 val loss 6.2913 +12800 val perplexity 539.8543 +12800 train 6.223473 (lr=1.7303e-05) (hash(x)=50786101) +11500 val loss 6.0878 +11500 val perplexity 440.4308 +11500 train 6.052198 (lr=2.6713e-05) (hash(x)=46985517) +12900 val loss 6.2871 +12900 val perplexity 537.6138 +12900 train 6.369383 (lr=1.7263e-05) (hash(x)=51296645) +11600 val loss 6.0798 +11600 val perplexity 436.9231 +11600 train 5.825112 (lr=2.6657e-05) (hash(x)=46741028) +13000 val loss 6.2791 +13000 val perplexity 533.3074 +13000 train 5.997715 (lr=1.7222e-05) (hash(x)=44184494) +11700 val loss 6.0807 +11700 val perplexity 437.3475 +11700 train 5.982686 (lr=2.6601e-05) (hash(x)=45729773) +13100 val loss 6.2678 +13100 val perplexity 527.2916 +13100 train 6.203880 (lr=1.7181e-05) (hash(x)=48196980) +11800 val loss 6.0793 +11800 val perplexity 436.7238 +11800 train 6.027054 (lr=2.6544e-05) (hash(x)=50540350) +13200 val loss 6.2634 +13200 val perplexity 524.9874 +13200 train 6.360188 (lr=1.7139e-05) (hash(x)=52207779) +11900 val loss 6.0706 +11900 val perplexity 432.9360 +11900 train 6.012947 (lr=2.6487e-05) (hash(x)=50313941) +13300 val loss 6.2615 +13300 val perplexity 523.9798 +13300 train 6.272325 (lr=1.7098e-05) (hash(x)=51012914) +13400 val loss 6.2582 +13400 val perplexity 522.2868 +13400 train 6.269378 (lr=1.7056e-05) (hash(x)=48426131) +12000 val loss 6.0586 +12000 val perplexity 427.7959 +12000 train 5.994100 (lr=2.6429e-05) (hash(x)=49621720) +13500 val loss 6.2518 +13500 val perplexity 518.9600 +13500 train 6.150234 (lr=1.7014e-05) (hash(x)=50588044) +12100 val loss 6.0577 +12100 val perplexity 427.3901 +12100 train 5.938934 (lr=2.6371e-05) (hash(x)=48903010) +13600 val loss 6.2461 +13600 val perplexity 516.0189 +13600 train 6.403970 (lr=1.6971e-05) (hash(x)=57937534) +12200 val loss 6.0579 +12200 val perplexity 427.4853 +12200 train 6.045485 (lr=2.6313e-05) (hash(x)=51643030) +13700 val loss 6.2414 +13700 val perplexity 513.5801 +13700 train 6.327751 (lr=1.6929e-05) (hash(x)=51168202) +12300 val loss 6.0496 +12300 val perplexity 423.9645 +12300 train 6.004856 (lr=2.6254e-05) (hash(x)=48863143) +13800 val loss 6.2342 +13800 val perplexity 509.9113 +13800 train 5.953720 (lr=1.6886e-05) (hash(x)=44144400) +12400 val loss 6.0488 +12400 val perplexity 423.5872 +12400 train 5.930643 (lr=2.6195e-05) (hash(x)=53251144) +13900 val loss 6.2334 +13900 val perplexity 509.4706 +13900 train 6.366139 (lr=1.6843e-05) (hash(x)=54816018) +12500 val loss 6.0447 +12500 val perplexity 421.8845 +12500 train 5.827845 (lr=2.6136e-05) (hash(x)=45178705) +14000 val loss 6.2291 +14000 val perplexity 507.2911 +14000 train 6.100036 (lr=1.6800e-05) (hash(x)=46103642) +14100 val loss 6.2238 +14100 val perplexity 504.6290 +14100 train 6.146904 (lr=1.6756e-05) (hash(x)=49605955) +12600 val loss 6.1032 +12600 val perplexity 447.2795 +12600 train 6.109112 (lr=2.6076e-05) (hash(x)=52779738) +14200 val loss 6.2229 +14200 val perplexity 504.1569 +14200 train 6.083589 (lr=1.6712e-05) (hash(x)=48663080) +12700 val loss 6.0315 +12700 val perplexity 416.3470 +12700 train 5.901758 (lr=2.6016e-05) (hash(x)=50253395) +14300 val loss 6.2238 +14300 val perplexity 504.6240 +14300 train 6.223890 (lr=1.6668e-05) (hash(x)=52021000) +12800 val loss 6.0299 +12800 val perplexity 415.6594 +12800 train 5.968020 (lr=2.5955e-05) (hash(x)=50786101) +14400 val loss 6.2128 +14400 val perplexity 499.1189 +14400 train 6.616567 (lr=1.6624e-05) (hash(x)=55092055) +12900 val loss 6.0260 +12900 val perplexity 414.0648 +12900 train 6.137391 (lr=2.5894e-05) (hash(x)=51296645) +14500 val loss 6.2092 +14500 val perplexity 497.3018 +14500 train 6.175682 (lr=1.6580e-05) (hash(x)=51163106) +13000 val loss 6.0190 +13000 val perplexity 411.1553 +13000 train 5.728059 (lr=2.5833e-05) (hash(x)=44184494) +14600 val loss 6.2011 +14600 val perplexity 493.3033 +14600 train 6.021957 (lr=1.6535e-05) (hash(x)=43587182) +13100 val loss 6.0108 +13100 val perplexity 407.8032 +13100 train 5.944895 (lr=2.5771e-05) (hash(x)=48196980) +14700 val loss 6.2002 +14700 val perplexity 492.8578 +14700 train 6.244566 (lr=1.6490e-05) (hash(x)=52950516) +13200 val loss 6.0057 +13200 val perplexity 405.7547 +13200 train 6.138175 (lr=2.5709e-05) (hash(x)=52207779) +14800 val loss 6.2009 +14800 val perplexity 493.2000 +14800 train 6.385430 (lr=1.6445e-05) (hash(x)=55136517) +13300 val loss 6.0019 +13300 val perplexity 404.1921 +13300 train 6.026430 (lr=2.5647e-05) (hash(x)=51012914) +14900 val loss 6.1912 +14900 val perplexity 488.4457 +14900 train 6.151821 (lr=1.6400e-05) (hash(x)=49852275) +15000 val loss 6.1845 +15000 val perplexity 485.1857 +13400 val loss 6.0048 +13400 val perplexity 405.3834 +13400 train 6.038949 (lr=2.5584e-05) (hash(x)=48426131) +15000 train 6.226887 (lr=1.6354e-05) (hash(x)=49783972) +15100 val loss 6.1777 +15100 val perplexity 481.8759 +15100 train 6.194267 (lr=1.6308e-05) (hash(x)=52480637) +13500 val loss 5.9974 +13500 val perplexity 402.3917 +13500 train 5.843425 (lr=2.5521e-05) (hash(x)=50588044) +15200 val loss 6.1729 +15200 val perplexity 479.5715 +15200 train 6.115196 (lr=1.6262e-05) (hash(x)=52067533) +13600 val loss 5.9921 +13600 val perplexity 400.2439 +13600 train 6.133362 (lr=2.5457e-05) (hash(x)=57937534) +15300 val loss 6.1787 +15300 val perplexity 482.3497 +15300 train 6.046903 (lr=1.6216e-05) (hash(x)=49883403) +13700 val loss 5.9864 +13700 val perplexity 397.9738 +13700 train 6.125409 (lr=2.5393e-05) (hash(x)=51168202) +15400 val loss 6.1757 +15400 val perplexity 480.9192 +15400 train 6.010082 (lr=1.6170e-05) (hash(x)=46270791) +13800 val loss 5.9803 +13800 val perplexity 395.5565 +13800 train 5.707791 (lr=2.5329e-05) (hash(x)=44144400) +15500 val loss 6.1653 +15500 val perplexity 475.9405 +15500 train 6.250294 (lr=1.6123e-05) (hash(x)=51703939) +13900 val loss 5.9861 +13900 val perplexity 397.8416 +13900 train 6.106628 (lr=2.5264e-05) (hash(x)=54816018) +15600 val loss 6.1615 +15600 val perplexity 474.1274 +15600 train 7.043086 (lr=1.6077e-05) (hash(x)=62165678) +14000 val loss 5.9833 +14000 val perplexity 396.7303 +14000 train 5.867708 (lr=2.5199e-05) (hash(x)=46103642) +15700 val loss 6.1591 +15700 val perplexity 472.9898 +15700 train 6.043394 (lr=1.6030e-05) (hash(x)=47311753) +14100 val loss 5.9780 +14100 val perplexity 394.6567 +14100 train 5.910935 (lr=2.5134e-05) (hash(x)=49605955) +15800 val loss 6.1556 +15800 val perplexity 471.3327 +15800 train 6.131933 (lr=1.5982e-05) (hash(x)=52207302) +14200 val loss 5.9755 +14200 val perplexity 393.6552 +14200 train 5.845827 (lr=2.5068e-05) (hash(x)=48663080) +15900 val loss 6.1577 +15900 val perplexity 472.3560 +15900 train 5.995483 (lr=1.5935e-05) (hash(x)=51909196) +14300 val loss 5.9712 +14300 val perplexity 391.9720 +14300 train 5.956995 (lr=2.5002e-05) (hash(x)=52021000) +16000 val loss 6.1503 +16000 val perplexity 468.8770 +16000 train 5.913560 (lr=1.5887e-05) (hash(x)=41044250) +16100 val loss 6.1501 +16100 val perplexity 468.7865 +16100 train 6.110799 (lr=1.5840e-05) (hash(x)=51378388) +14400 val loss 5.9682 +14400 val perplexity 390.7853 +14400 train 6.435887 (lr=2.4936e-05) (hash(x)=55092055) +16200 val loss 6.1428 +16200 val perplexity 465.3760 +16200 train 5.894137 (lr=1.5792e-05) (hash(x)=46846254) +14500 val loss 5.9667 +14500 val perplexity 390.2196 +14500 train 5.912242 (lr=2.4869e-05) (hash(x)=51163106) +16300 val loss 6.1424 +16300 val perplexity 465.1870 +16300 train 6.205682 (lr=1.5744e-05) (hash(x)=49455544) +14600 val loss 5.9550 +14600 val perplexity 385.6945 +14600 train 5.733799 (lr=2.4802e-05) (hash(x)=43587182) +16400 val loss 6.1331 +16400 val perplexity 460.8530 +16400 train 6.119483 (lr=1.5695e-05) (hash(x)=49391900) +14700 val loss 5.9540 +14700 val perplexity 385.3067 +14700 train 6.025739 (lr=2.4735e-05) (hash(x)=52950516) +16500 val loss 6.1293 +16500 val perplexity 459.1338 +16500 train 6.366375 (lr=1.5647e-05) (hash(x)=51821231) +14800 val loss 5.9566 +14800 val perplexity 386.3012 +14800 train 6.162257 (lr=2.4668e-05) (hash(x)=55136517) +16600 val loss 6.1203 +16600 val perplexity 454.9787 +16600 train 6.387796 (lr=1.5598e-05) (hash(x)=59206850) +14900 val loss 5.9492 +14900 val perplexity 383.4615 +14900 train 5.915737 (lr=2.4600e-05) (hash(x)=49852275) +16700 val loss 6.1202 +16700 val perplexity 454.9436 +16700 train 6.077214 (lr=1.5549e-05) (hash(x)=48167967) +15000 val loss 5.9479 +15000 val perplexity 382.9325 +15000 train 5.994071 (lr=2.4531e-05) (hash(x)=49783972) +16800 val loss 6.1145 +16800 val perplexity 452.3644 +16800 train 6.114574 (lr=1.5500e-05) (hash(x)=54230562) +15100 val loss 5.9314 +15100 val perplexity 376.6728 +15100 train 5.936438 (lr=2.4463e-05) (hash(x)=52480637) +16900 val loss 6.1120 +16900 val perplexity 451.2605 +16900 train 5.845520 (lr=1.5451e-05) (hash(x)=40802028) +17000 val loss 6.1041 +17000 val perplexity 447.6990 +17000 train 6.067628 (lr=1.5401e-05) (hash(x)=48123845) +15200 val loss 5.9312 +15200 val perplexity 376.6062 +15200 train 5.877983 (lr=2.4394e-05) (hash(x)=52067533) +17100 val loss 6.1031 +17100 val perplexity 447.2200 +17100 train 5.994694 (lr=1.5352e-05) (hash(x)=52646709) +15300 val loss 5.9368 +15300 val perplexity 378.7236 +15300 train 5.826802 (lr=2.4324e-05) (hash(x)=49883403) +17200 val loss 6.1011 +17200 val perplexity 446.3661 +17200 train 6.152263 (lr=1.5302e-05) (hash(x)=49699031) +15400 val loss 5.9347 +15400 val perplexity 377.9084 +15400 train 5.814553 (lr=2.4255e-05) (hash(x)=46270791) +17300 val loss 6.0961 +17300 val perplexity 444.1320 +17300 train 6.249147 (lr=1.5252e-05) (hash(x)=54509180) +15500 val loss 5.9242 +15500 val perplexity 373.9697 +15500 train 6.042672 (lr=2.4185e-05) (hash(x)=51703939) +17400 val loss 6.0962 +17400 val perplexity 444.1722 +17400 train 5.904134 (lr=1.5202e-05) (hash(x)=48132725) +15600 val loss 5.9184 +15600 val perplexity 371.8223 +15600 train 6.721773 (lr=2.4115e-05) (hash(x)=62165678) +17500 val loss 6.0945 +17500 val perplexity 443.4203 +17500 train 5.554464 (lr=1.5152e-05) (hash(x)=42062605) +15700 val loss 5.9157 +15700 val perplexity 370.8311 +15700 train 5.816174 (lr=2.4044e-05) (hash(x)=47311753) +17600 val loss 6.0888 +17600 val perplexity 440.9120 +17600 train 6.016416 (lr=1.5101e-05) (hash(x)=51292000) +15800 val loss 5.9161 +15800 val perplexity 370.9722 +15800 train 5.892066 (lr=2.3974e-05) (hash(x)=52207302) +17700 val loss 6.0882 +17700 val perplexity 440.6306 +17700 train 5.940475 (lr=1.5050e-05) (hash(x)=51626643) +17800 val loss 6.0890 +17800 val perplexity 440.9809 +17800 train 5.936723 (lr=1.5000e-05) (hash(x)=53093658) +15900 val loss 5.9138 +15900 val perplexity 370.0938 +15900 train 5.759565 (lr=2.3903e-05) (hash(x)=51909196) +17900 val loss 6.0855 +17900 val perplexity 439.4597 +17900 train 6.175018 (lr=1.4949e-05) (hash(x)=51370179) +16000 val loss 5.9071 +16000 val perplexity 367.6250 +16000 train 5.663516 (lr=2.3831e-05) (hash(x)=41044250) +18000 val loss 6.0839 +18000 val perplexity 438.7534 +18000 train 5.932031 (lr=1.4898e-05) (hash(x)=49974890) +16100 val loss 5.9048 +16100 val perplexity 366.8050 +16100 train 5.869083 (lr=2.3760e-05) (hash(x)=51378388) +18100 val loss 6.0744 +18100 val perplexity 434.5851 +18100 train 6.050805 (lr=1.4846e-05) (hash(x)=55831634) +16200 val loss 5.9023 +16200 val perplexity 365.8700 +16200 train 5.674711 (lr=2.3688e-05) (hash(x)=46846254) +18200 val loss 6.0670 +18200 val perplexity 431.3642 +18200 train 6.032460 (lr=1.4795e-05) (hash(x)=49919405) +16300 val loss 5.9040 +16300 val perplexity 366.5148 +16300 train 5.984703 (lr=2.3615e-05) (hash(x)=49455544) +18300 val loss 6.0684 +18300 val perplexity 431.9704 +18300 train 6.192651 (lr=1.4743e-05) (hash(x)=53822543) +16400 val loss 5.8953 +16400 val perplexity 363.3400 +16400 train 5.897575 (lr=2.3543e-05) (hash(x)=49391900) +18400 val loss 6.0621 +18400 val perplexity 429.2573 +18400 train 6.411750 (lr=1.4692e-05) (hash(x)=61904736) +16500 val loss 5.8990 +16500 val perplexity 364.6721 +16500 train 6.106722 (lr=2.3470e-05) (hash(x)=51821231) +18500 val loss 6.0546 +18500 val perplexity 426.0665 +18500 train 5.896453 (lr=1.4640e-05) (hash(x)=47570884) +16600 val loss 5.9003 +16600 val perplexity 365.1296 +16600 train 6.136813 (lr=2.3397e-05) (hash(x)=59206850) +18600 val loss 6.0490 +18600 val perplexity 423.6753 +18600 train 5.960378 (lr=1.4588e-05) (hash(x)=49104519) +18700 val loss 6.0496 +18700 val perplexity 423.9500 +18700 train 5.939255 (lr=1.4536e-05) (hash(x)=54160283) +16700 val loss 5.8852 +16700 val perplexity 359.6775 +16700 train 5.861080 (lr=2.3324e-05) (hash(x)=48167967) +18800 val loss 6.0451 +18800 val perplexity 422.0511 +18800 train 6.026224 (lr=1.4483e-05) (hash(x)=47840786) +16800 val loss 5.8821 +16800 val perplexity 358.5589 +16800 train 5.869708 (lr=2.3250e-05) (hash(x)=54230562) +18900 val loss 6.0416 +18900 val perplexity 420.5448 +18900 train 6.142548 (lr=1.4431e-05) (hash(x)=48397694) +16900 val loss 5.8824 +16900 val perplexity 358.6738 +16900 train 5.661252 (lr=2.3176e-05) (hash(x)=40802028) +19000 val loss 6.0381 +19000 val perplexity 419.1133 +19000 train 5.952710 (lr=1.4378e-05) (hash(x)=50089065) +17000 val loss 5.8791 +17000 val perplexity 357.4866 +17000 train 5.878530 (lr=2.3102e-05) (hash(x)=48123845) +19100 val loss 6.0372 +19100 val perplexity 418.7146 +19100 train 5.821996 (lr=1.4326e-05) (hash(x)=46007516) +17100 val loss 5.8743 +17100 val perplexity 355.7910 +17100 train 5.778979 (lr=2.3028e-05) (hash(x)=52646709) +19200 val loss 6.0390 +19200 val perplexity 419.4774 +19200 train 6.021871 (lr=1.4273e-05) (hash(x)=42677001) +17200 val loss 5.8683 +17200 val perplexity 353.6585 +17200 train 5.937093 (lr=2.2953e-05) (hash(x)=49699031) +19300 val loss 6.0336 +19300 val perplexity 417.2236 +19300 train 5.941748 (lr=1.4220e-05) (hash(x)=46979320) +17300 val loss 5.8676 +17300 val perplexity 353.4057 +17300 train 6.004000 (lr=2.2878e-05) (hash(x)=54509180) +19400 val loss 6.0341 +19400 val perplexity 417.4403 +19400 train 5.714228 (lr=1.4167e-05) (hash(x)=47572038) +17400 val loss 5.8660 +17400 val perplexity 352.8398 +17400 train 5.654855 (lr=2.2803e-05) (hash(x)=48132725) +19500 val loss 6.0336 +19500 val perplexity 417.2037 +19500 train 6.014361 (lr=1.4114e-05) (hash(x)=52774451) +17500 val loss 5.8620 +17500 val perplexity 351.4308 +17500 train 5.350478 (lr=2.2727e-05) (hash(x)=42062605) +19600 val loss 6.0309 +19600 val perplexity 416.0959 +19600 train 6.133531 (lr=1.4060e-05) (hash(x)=53716950) +19700 val loss 6.0247 +19700 val perplexity 413.5050 +19700 train 5.745566 (lr=1.4007e-05) (hash(x)=45863098) +17600 val loss 5.8609 +17600 val perplexity 351.0490 +17600 train 5.795611 (lr=2.2652e-05) (hash(x)=51292000) +19800 val loss 6.0209 +19800 val perplexity 411.9560 +19800 train 6.141560 (lr=1.3953e-05) (hash(x)=51188683) +17700 val loss 5.8627 +17700 val perplexity 351.6779 +17700 train 5.707959 (lr=2.2576e-05) (hash(x)=51626643) +19900 val loss 6.0218 +19900 val perplexity 412.3146 +19900 train 5.933441 (lr=1.3900e-05) (hash(x)=49990053) +17800 val loss 5.8562 +17800 val perplexity 349.3922 +17800 train 5.693523 (lr=2.2500e-05) (hash(x)=53093658) +20000 val loss 6.0218 +20000 val perplexity 412.3254 +20000 train 5.761784 (lr=1.3846e-05) (hash(x)=44849004) +17900 val loss 5.8519 +17900 val perplexity 347.8867 +17900 train 5.952915 (lr=2.2423e-05) (hash(x)=51370179) +20100 val loss 6.0160 +20100 val perplexity 409.9290 +20100 train 5.869083 (lr=1.3792e-05) (hash(x)=50132035) +18000 val loss 5.8545 +18000 val perplexity 348.8140 +18000 train 5.719038 (lr=2.2346e-05) (hash(x)=49974890) +20200 val loss 6.0044 +20200 val perplexity 405.2035 +20200 train 6.187902 (lr=1.3738e-05) (hash(x)=54258143) +18100 val loss 5.8436 +18100 val perplexity 345.0060 +18100 train 5.807959 (lr=2.2270e-05) (hash(x)=55831634) +20300 val loss 6.0021 +20300 val perplexity 404.2697 +20300 train 6.529205 (lr=1.3684e-05) (hash(x)=54953105) +18200 val loss 5.8410 +18200 val perplexity 344.1384 +18200 train 5.861757 (lr=2.2192e-05) (hash(x)=49919405) +20400 val loss 5.9999 +20400 val perplexity 403.3897 +20400 train 6.054474 (lr=1.3630e-05) (hash(x)=51574049) +18300 val loss 5.8348 +18300 val perplexity 341.9892 +18300 train 5.957066 (lr=2.2115e-05) (hash(x)=53822543) +20500 val loss 5.9959 +20500 val perplexity 401.7834 +20500 train 6.178652 (lr=1.3575e-05) (hash(x)=51616142) +18400 val loss 5.8288 +18400 val perplexity 339.9663 +18400 train 6.025560 (lr=2.2038e-05) (hash(x)=61904736) +20600 val loss 5.9886 +20600 val perplexity 398.8451 +20600 train 6.286069 (lr=1.3521e-05) (hash(x)=52549976) +18500 val loss 5.8366 +18500 val perplexity 342.6103 +18500 train 5.679751 (lr=2.1960e-05) (hash(x)=47570884) +20700 val loss 5.9864 +20700 val perplexity 397.9928 +20700 train 5.951801 (lr=1.3466e-05) (hash(x)=60058155) +18600 val loss 5.8235 +18600 val perplexity 338.1550 +20800 val loss 5.9851 +20800 val perplexity 397.4624 +18600 train 5.759852 (lr=2.1882e-05) (hash(x)=49104519) +20800 train 5.951099 (lr=1.3412e-05) (hash(x)=51348881) +20900 val loss 5.9884 +20900 val perplexity 398.7848 +20900 train 5.973612 (lr=1.3357e-05) (hash(x)=56417746) +18700 val loss 5.8200 +18700 val perplexity 336.9660 +18700 train 5.679973 (lr=2.1804e-05) (hash(x)=54160283) +21000 val loss 5.9831 +21000 val perplexity 396.6817 +21000 train 5.910104 (lr=1.3302e-05) (hash(x)=50984540) +18800 val loss 5.8196 +18800 val perplexity 336.8245 +18800 train 5.819785 (lr=2.1725e-05) (hash(x)=47840786) +21100 val loss 5.9805 +21100 val perplexity 395.6345 +21100 train 5.985803 (lr=1.3247e-05) (hash(x)=47991688) +18900 val loss 5.8165 +18900 val perplexity 335.8023 +18900 train 5.984718 (lr=2.1647e-05) (hash(x)=48397694) +21200 val loss 5.9739 +21200 val perplexity 393.0222 +21200 train 5.421027 (lr=1.3192e-05) (hash(x)=38120795) +19000 val loss 5.8153 +19000 val perplexity 335.3795 +19000 train 5.732840 (lr=2.1568e-05) (hash(x)=50089065) +21300 val loss 5.9783 +21300 val perplexity 394.7702 +21300 train 5.895738 (lr=1.3137e-05) (hash(x)=43713373) +19100 val loss 5.8142 +19100 val perplexity 335.0178 +19100 train 5.590156 (lr=2.1489e-05) (hash(x)=46007516) +21400 val loss 5.9803 +21400 val perplexity 395.5607 +21400 train 5.938613 (lr=1.3082e-05) (hash(x)=50567003) +19200 val loss 5.8186 +19200 val perplexity 336.5158 +19200 train 5.819208 (lr=2.1409e-05) (hash(x)=42677001) +21500 val loss 5.9708 +21500 val perplexity 391.8250 +21500 train 5.866252 (lr=1.3026e-05) (hash(x)=49430102) +19300 val loss 5.8119 +19300 val perplexity 334.2517 +19300 train 5.738682 (lr=2.1330e-05) (hash(x)=46979320) +21600 val loss 5.9748 +21600 val perplexity 393.3805 +21600 train 5.894325 (lr=1.2971e-05) (hash(x)=49016562) +21700 val loss 5.9695 +21700 val perplexity 391.3068 +19400 val loss 5.8148 +19400 val perplexity 335.2123 +21700 train 5.949615 (lr=1.2916e-05) (hash(x)=48939247) +19400 train 5.502016 (lr=2.1250e-05) (hash(x)=47572038) +21800 val loss 5.9655 +21800 val perplexity 389.7344 +21800 train 6.056383 (lr=1.2860e-05) (hash(x)=55200591) +19500 val loss 5.8153 +19500 val perplexity 335.3952 +19500 train 5.778347 (lr=2.1171e-05) (hash(x)=52774451) +21900 val loss 5.9656 +21900 val perplexity 389.7818 +21900 train 5.741512 (lr=1.2805e-05) (hash(x)=45779836) +19600 val loss 5.8077 +19600 val perplexity 332.8546 +19600 train 5.896122 (lr=2.1091e-05) (hash(x)=53716950) +22000 val loss 5.9626 +22000 val perplexity 388.6237 +22000 train 5.774112 (lr=1.2749e-05) (hash(x)=54998738) +19700 val loss 5.8048 +19700 val perplexity 331.8799 +19700 train 5.539446 (lr=2.1010e-05) (hash(x)=45863098) +22100 val loss 5.9584 +22100 val perplexity 387.0090 +22100 train 6.048635 (lr=1.2693e-05) (hash(x)=52133006) +19800 val loss 5.8027 +19800 val perplexity 331.1885 +19800 train 5.938042 (lr=2.0930e-05) (hash(x)=51188683) +22200 val loss 5.9564 +22200 val perplexity 386.2010 +22200 train 5.964255 (lr=1.2637e-05) (hash(x)=51835734) +19900 val loss 5.8017 +19900 val perplexity 330.8692 +19900 train 5.713326 (lr=2.0850e-05) (hash(x)=49990053) +22300 val loss 5.9453 +22300 val perplexity 381.9538 +22300 train 5.763449 (lr=1.2581e-05) (hash(x)=47416361) +20000 val loss 5.8035 +20000 val perplexity 331.4460 +22400 val loss 5.9432 +22400 val perplexity 381.1614 +22400 train 6.051149 (lr=1.2526e-05) (hash(x)=53442231) +20000 train 5.523485 (lr=2.0769e-05) (hash(x)=44849004) +22500 val loss 5.9401 +22500 val perplexity 379.9714 +22500 train 6.009403 (lr=1.2470e-05) (hash(x)=51470764) +20100 val loss 5.7975 +20100 val perplexity 329.4620 +20100 train 5.645189 (lr=2.0688e-05) (hash(x)=50132035) +22600 val loss 5.9390 +22600 val perplexity 379.5370 +22600 train 6.174877 (lr=1.2414e-05) (hash(x)=57017795) +20200 val loss 5.7883 +20200 val perplexity 326.4685 +20200 train 5.988879 (lr=2.0607e-05) (hash(x)=54258143) +22700 val loss 5.9369 +22700 val perplexity 378.7643 +22700 train 5.906285 (lr=1.2357e-05) (hash(x)=51579140) +20300 val loss 5.7862 +20300 val perplexity 325.7596 +20300 train 6.329162 (lr=2.0526e-05) (hash(x)=54953105) +22800 val loss 5.9340 +22800 val perplexity 377.6465 +22800 train 5.953415 (lr=1.2301e-05) (hash(x)=54012893) +20400 val loss 5.7830 +20400 val perplexity 324.7239 +20400 train 5.816953 (lr=2.0444e-05) (hash(x)=51574049) +22900 val loss 5.9355 +22900 val perplexity 378.2212 +22900 train 5.940351 (lr=1.2245e-05) (hash(x)=47758345) +20500 val loss 5.7754 +20500 val perplexity 322.2583 +20500 train 5.945559 (lr=2.0363e-05) (hash(x)=51616142) +23000 val loss 5.9337 +23000 val perplexity 377.5433 +23000 train 6.334531 (lr=1.2189e-05) (hash(x)=57909888) +20600 val loss 5.7756 +20600 val perplexity 322.3298 +20600 train 6.115684 (lr=2.0281e-05) (hash(x)=52549976) +23100 val loss 5.9305 +23100 val perplexity 376.3360 +23100 train 6.088027 (lr=1.2133e-05) (hash(x)=53979657) +20700 val loss 5.7725 +20700 val perplexity 321.3473 +20700 train 5.657379 (lr=2.0199e-05) (hash(x)=60058155) +23200 val loss 5.9320 +23200 val perplexity 376.9218 +23200 train 5.751225 (lr=1.2076e-05) (hash(x)=52104721) +20800 val loss 5.7652 +20800 val perplexity 319.0172 +23300 val loss 5.9275 +23300 val perplexity 375.2220 +20800 train 5.760611 (lr=2.0117e-05) (hash(x)=51348881) +23300 train 5.792253 (lr=1.2020e-05) (hash(x)=51364855) +23400 val loss 5.9261 +23400 val perplexity 374.6865 +23400 train 5.849146 (lr=1.1963e-05) (hash(x)=49542635) +20900 val loss 5.7665 +20900 val perplexity 319.4292 +20900 train 5.730649 (lr=2.0035e-05) (hash(x)=56417746) +23500 val loss 5.9281 +23500 val perplexity 375.4353 +23500 train 5.701175 (lr=1.1907e-05) (hash(x)=46017969) +21000 val loss 5.7711 +21000 val perplexity 320.8982 +21000 train 5.708296 (lr=1.9953e-05) (hash(x)=50984540) +23600 val loss 5.9233 +23600 val perplexity 373.6278 +23600 train 5.985129 (lr=1.1850e-05) (hash(x)=51565375) +21100 val loss 5.7653 +21100 val perplexity 319.0336 +21100 train 5.777886 (lr=1.9871e-05) (hash(x)=47991688) +23700 val loss 5.9217 +23700 val perplexity 373.0592 +23700 train 5.810538 (lr=1.1794e-05) (hash(x)=48511656) +21200 val loss 5.7595 +21200 val perplexity 317.1782 +21200 train 5.210901 (lr=1.9788e-05) (hash(x)=38120795) +23800 val loss 5.9207 +23800 val perplexity 372.6757 +23800 train 5.888032 (lr=1.1737e-05) (hash(x)=51503144) +21300 val loss 5.7671 +21300 val perplexity 319.6230 +21300 train 5.701467 (lr=1.9705e-05) (hash(x)=43713373) +23900 val loss 5.9154 +23900 val perplexity 370.6891 +23900 train 5.994982 (lr=1.1681e-05) (hash(x)=53271580) +21400 val loss 5.7625 +21400 val perplexity 318.1310 +21400 train 5.717454 (lr=1.9623e-05) (hash(x)=50567003) +24000 val loss 5.9111 +24000 val perplexity 369.1242 +24000 train 5.913059 (lr=1.1624e-05) (hash(x)=45797109) +21500 val loss 5.7563 +21500 val perplexity 316.1808 +21500 train 5.601381 (lr=1.9540e-05) (hash(x)=49430102) +24100 val loss 5.9086 +24100 val perplexity 368.1822 +24100 train 6.196603 (lr=1.1567e-05) (hash(x)=54215536) +21600 val loss 5.7586 +21600 val perplexity 316.9136 +21600 train 5.641023 (lr=1.9457e-05) (hash(x)=49016562) +24200 val loss 5.9042 +24200 val perplexity 366.5886 +24200 train 5.879582 (lr=1.1511e-05) (hash(x)=50087698) +21700 val loss 5.7559 +21700 val perplexity 316.0577 +21700 train 5.748583 (lr=1.9373e-05) (hash(x)=48939247) +24300 val loss 5.9112 +24300 val perplexity 369.1459 +24300 train 6.129028 (lr=1.1454e-05) (hash(x)=52070789) +21800 val loss 5.7545 +21800 val perplexity 315.6178 +21800 train 5.810953 (lr=1.9290e-05) (hash(x)=55200591) +24400 val loss 5.9040 +24400 val perplexity 366.4828 +24400 train 6.093769 (lr=1.1397e-05) (hash(x)=51743588) +24500 val loss 5.8966 +24500 val perplexity 363.7904 +24500 train 5.750652 (lr=1.1341e-05) (hash(x)=48574502) +21900 val loss 5.7576 +21900 val perplexity 316.5737 +21900 train 5.534771 (lr=1.9207e-05) (hash(x)=45779836) +24600 val loss 5.8935 +24600 val perplexity 362.6777 +24600 train 5.616612 (lr=1.1284e-05) (hash(x)=42950372) +22000 val loss 5.7519 +22000 val perplexity 314.8027 +22000 train 5.532240 (lr=1.9123e-05) (hash(x)=54998738) +24700 val loss 5.8929 +24700 val perplexity 362.4545 +24700 train 5.881211 (lr=1.1227e-05) (hash(x)=48073103) +22100 val loss 5.7488 +22100 val perplexity 313.8141 +22100 train 5.866989 (lr=1.9040e-05) (hash(x)=52133006) +24800 val loss 5.8908 +24800 val perplexity 361.6855 +24800 train 5.827632 (lr=1.1170e-05) (hash(x)=55420175) +22200 val loss 5.7481 +22200 val perplexity 313.5800 +22200 train 5.745663 (lr=1.8956e-05) (hash(x)=51835734) +24900 val loss 5.8884 +24900 val perplexity 360.8165 +24900 train 5.926189 (lr=1.1114e-05) (hash(x)=50911918) +22300 val loss 5.7384 +22300 val perplexity 310.5713 +22300 train 5.516079 (lr=1.8872e-05) (hash(x)=47416361) +25000 val loss 5.8951 +25000 val perplexity 363.2461 +25000 train 5.847051 (lr=1.1057e-05) (hash(x)=56577519) +22400 val loss 5.7391 +22400 val perplexity 310.7941 +22400 train 5.858449 (lr=1.8788e-05) (hash(x)=53442231) +25100 val loss 5.8871 +25100 val perplexity 360.3444 +25100 train 5.949383 (lr=1.1000e-05) (hash(x)=54225130) +22500 val loss 5.7271 +22500 val perplexity 307.0699 +22500 train 5.813729 (lr=1.8704e-05) (hash(x)=51470764) +25200 val loss 5.8893 +25200 val perplexity 361.1499 +25200 train 5.990960 (lr=1.0943e-05) (hash(x)=50144293) +25300 val loss 5.8866 +25300 val perplexity 360.1645 +25300 train 5.666760 (lr=1.0886e-05) (hash(x)=43198841) +22600 val loss 5.7294 +22600 val perplexity 307.7846 +22600 train 5.985736 (lr=1.8620e-05) (hash(x)=57017795) +25400 val loss 5.8857 +25400 val perplexity 359.8430 +25400 train 5.611393 (lr=1.0830e-05) (hash(x)=48956461) +22700 val loss 5.7247 +22700 val perplexity 306.3399 +22700 train 5.633924 (lr=1.8536e-05) (hash(x)=51579140) +25500 val loss 5.8835 +25500 val perplexity 359.0599 +25500 train 6.041482 (lr=1.0773e-05) (hash(x)=57573293) +22800 val loss 5.7288 +22800 val perplexity 307.6085 +22800 train 5.764848 (lr=1.8452e-05) (hash(x)=54012893) +25600 val loss 5.8764 +25600 val perplexity 356.5126 +25600 train 5.823350 (lr=1.0716e-05) (hash(x)=47499056) +22900 val loss 5.7213 +22900 val perplexity 305.3115 +22900 train 5.698467 (lr=1.8368e-05) (hash(x)=47758345) +25700 val loss 5.8771 +25700 val perplexity 356.7663 +25700 train 7.027942 (lr=1.0659e-05) (hash(x)=42134899) +23000 val loss 5.7252 +23000 val perplexity 306.5012 +23000 train 6.113164 (lr=1.8283e-05) (hash(x)=57909888) +25800 val loss 5.8707 +25800 val perplexity 354.5051 +25800 train 5.863268 (lr=1.0603e-05) (hash(x)=52870010) +23100 val loss 5.7192 +23100 val perplexity 304.6544 +23100 train 5.909606 (lr=1.8199e-05) (hash(x)=53979657) +25900 val loss 5.8664 +25900 val perplexity 352.9640 +25900 train 5.805500 (lr=1.0546e-05) (hash(x)=51558271) +23200 val loss 5.7222 +23200 val perplexity 305.5668 +23200 train 5.548705 (lr=1.8114e-05) (hash(x)=52104721) +26000 val loss 5.8644 +26000 val perplexity 352.2628 +26000 train 5.882143 (lr=1.0489e-05) (hash(x)=49663062) +23300 val loss 5.7184 +23300 val perplexity 304.4150 +23300 train 5.571858 (lr=1.8030e-05) (hash(x)=51364855) +26100 val loss 5.8633 +26100 val perplexity 351.8946 +26100 train 6.115038 (lr=1.0433e-05) (hash(x)=48349329) +23400 val loss 5.7136 +23400 val perplexity 302.9636 +23400 train 5.647821 (lr=1.7945e-05) (hash(x)=49542635) +26200 val loss 5.8636 +26200 val perplexity 351.9834 +26200 train 5.802253 (lr=1.0376e-05) (hash(x)=48588410) +26300 val loss 5.8606 +26300 val perplexity 350.9521 +26300 train 5.891696 (lr=1.0319e-05) (hash(x)=52286930) +23500 val loss 5.7207 +23500 val perplexity 305.1249 +23500 train 5.473540 (lr=1.7860e-05) (hash(x)=46017969) +26400 val loss 5.8563 +26400 val perplexity 349.4251 +26400 train 5.690542 (lr=1.0263e-05) (hash(x)=43797796) +23600 val loss 5.7135 +23600 val perplexity 302.9402 +23600 train 5.797878 (lr=1.7776e-05) (hash(x)=51565375) +26500 val loss 5.8553 +26500 val perplexity 349.0819 +26500 train 5.826995 (lr=1.0206e-05) (hash(x)=50812654) +23700 val loss 5.7116 +23700 val perplexity 302.3485 +23700 train 5.598224 (lr=1.7691e-05) (hash(x)=48511656) +26600 val loss 5.8552 +26600 val perplexity 349.0456 +26600 train 5.613261 (lr=1.0150e-05) (hash(x)=45633368) +23800 val loss 5.7167 +23800 val perplexity 303.9111 +23800 train 5.672143 (lr=1.7606e-05) (hash(x)=51503144) +26700 val loss 5.8547 +26700 val perplexity 348.8693 +26700 train 6.176394 (lr=1.0093e-05) (hash(x)=58836955) +23900 val loss 5.7042 +23900 val perplexity 300.1384 +23900 train 5.792049 (lr=1.7521e-05) (hash(x)=53271580) +26800 val loss 5.8514 +26800 val perplexity 347.7104 +26800 train 5.999731 (lr=1.0037e-05) (hash(x)=53665925) +24000 val loss 5.6999 +24000 val perplexity 298.8437 +24000 train 5.698016 (lr=1.7436e-05) (hash(x)=45797109) +26900 val loss 5.8570 +26900 val perplexity 349.6895 +26900 train 6.004295 (lr=9.9802e-06) (hash(x)=49535575) +24100 val loss 5.6943 +24100 val perplexity 297.1815 +24100 train 6.031198 (lr=1.7351e-05) (hash(x)=54215536) +27000 val loss 5.8521 +27000 val perplexity 347.9539 +27000 train 5.467065 (lr=9.9238e-06) (hash(x)=46799308) +24200 val loss 5.6960 +24200 val perplexity 297.6638 +24200 train 5.697694 (lr=1.7266e-05) (hash(x)=50087698) +27100 val loss 5.8481 +27100 val perplexity 346.5844 +27100 train 5.500708 (lr=9.8675e-06) (hash(x)=42191940) +24300 val loss 5.6977 +24300 val perplexity 298.1792 +24300 train 5.962052 (lr=1.7181e-05) (hash(x)=52070789) +27200 val loss 5.8530 +27200 val perplexity 348.2829 +27200 train 5.592017 (lr=9.8112e-06) (hash(x)=42152270) +27300 val loss 5.8493 +27300 val perplexity 346.9922 +27300 train 5.723346 (lr=9.7549e-06) (hash(x)=44970208) +24400 val loss 5.6922 +24400 val perplexity 296.5355 +24400 train 5.827345 (lr=1.7096e-05) (hash(x)=51743588) +27400 val loss 5.8404 +27400 val perplexity 343.9197 +27400 train 5.933327 (lr=9.6987e-06) (hash(x)=52832889) +24500 val loss 5.6883 +24500 val perplexity 295.4049 +24500 train 5.551994 (lr=1.7011e-05) (hash(x)=48574502) +27500 val loss 5.8378 +27500 val perplexity 343.0135 +27500 train 5.768693 (lr=9.6426e-06) (hash(x)=49318466) +24600 val loss 5.6894 +24600 val perplexity 295.7131 +24600 train 5.419545 (lr=1.6926e-05) (hash(x)=42950372) +27600 val loss 5.8381 +27600 val perplexity 343.1184 +27600 train 5.896399 (lr=9.5865e-06) (hash(x)=50216713) +24700 val loss 5.6863 +24700 val perplexity 294.7985 +24700 train 5.698661 (lr=1.6841e-05) (hash(x)=48073103) +27700 val loss 5.8336 +27700 val perplexity 341.5766 +27700 train 5.812750 (lr=9.5304e-06) (hash(x)=44599482) +24800 val loss 5.6860 +24800 val perplexity 294.7076 +24800 train 5.582422 (lr=1.6755e-05) (hash(x)=55420175) +27800 val loss 5.8298 +27800 val perplexity 340.2769 +27800 train 5.952037 (lr=9.4745e-06) (hash(x)=52853588) +24900 val loss 5.6850 +24900 val perplexity 294.4222 +24900 train 5.709652 (lr=1.6670e-05) (hash(x)=50911918) +27900 val loss 5.8536 +27900 val perplexity 348.4835 +27900 train 5.772947 (lr=9.4185e-06) (hash(x)=45297678) +25000 val loss 5.6879 +25000 val perplexity 295.2711 +25000 train 5.630611 (lr=1.6585e-05) (hash(x)=56577519) +28000 val loss 5.8263 +28000 val perplexity 339.0886 +28000 train 5.655169 (lr=9.3627e-06) (hash(x)=45648934) +25100 val loss 5.6847 +25100 val perplexity 294.3364 +25100 train 5.724168 (lr=1.6500e-05) (hash(x)=54225130) +28100 val loss 5.8301 +28100 val perplexity 340.3922 +28100 train 5.812257 (lr=9.3069e-06) (hash(x)=46498270) +25200 val loss 5.6842 +25200 val perplexity 294.1722 +25200 train 5.795018 (lr=1.6415e-05) (hash(x)=50144293) +28200 val loss 5.8259 +28200 val perplexity 338.9700 +28200 train 5.671429 (lr=9.2511e-06) (hash(x)=45066849) +25300 val loss 5.6789 +25300 val perplexity 292.6372 +25300 train 5.487818 (lr=1.6330e-05) (hash(x)=43198841) +28300 val loss 5.8254 +28300 val perplexity 338.8084 +28300 train 5.754888 (lr=9.1955e-06) (hash(x)=46542887) +28400 val loss 5.8195 +28400 val perplexity 336.7893 +28400 train 5.843278 (lr=9.1399e-06) (hash(x)=50854598) +25400 val loss 5.6792 +25400 val perplexity 292.7068 +25400 train 5.415684 (lr=1.6245e-05) (hash(x)=48956461) +28500 val loss 5.8194 +28500 val perplexity 336.7724 +28500 train 5.931666 (lr=9.0844e-06) (hash(x)=50005909) +25500 val loss 5.6797 +25500 val perplexity 292.8660 +25500 train 5.833297 (lr=1.6159e-05) (hash(x)=57573293) +28600 val loss 5.8208 +28600 val perplexity 337.2304 +28600 train 5.869174 (lr=9.0290e-06) (hash(x)=50777665) +25600 val loss 5.6693 +25600 val perplexity 289.8215 +25600 train 5.498159 (lr=1.6074e-05) (hash(x)=47499056) +28700 val loss 5.8197 +28700 val perplexity 336.8558 +28700 train 6.245265 (lr=8.9736e-06) (hash(x)=60124649) +25700 val loss 5.6724 +25700 val perplexity 290.7197 +25700 train 6.751441 (lr=1.5989e-05) (hash(x)=42134899) +28800 val loss 5.8179 +28800 val perplexity 336.2687 +28800 train 5.713576 (lr=8.9183e-06) (hash(x)=49592362) +25800 val loss 5.6656 +25800 val perplexity 288.7690 +25800 train 5.650281 (lr=1.5904e-05) (hash(x)=52870010) +28900 val loss 5.8155 +28900 val perplexity 335.4672 +28900 train 5.763836 (lr=8.8631e-06) (hash(x)=48094168) +25900 val loss 5.6631 +25900 val perplexity 288.0313 +25900 train 5.635216 (lr=1.5819e-05) (hash(x)=51558271) +29000 val loss 5.8172 +29000 val perplexity 336.0302 +29000 train 5.642217 (lr=8.8080e-06) (hash(x)=49283091) +26000 val loss 5.6598 +26000 val perplexity 287.0926 +26000 train 5.691453 (lr=1.5734e-05) (hash(x)=49663062) +29100 val loss 5.8152 +29100 val perplexity 335.3426 +29100 train 5.587608 (lr=8.7530e-06) (hash(x)=49704892) +26100 val loss 5.6610 +26100 val perplexity 287.4292 +26100 train 5.939886 (lr=1.5649e-05) (hash(x)=48349329) +29200 val loss 5.8157 +29200 val perplexity 335.5421 +29200 train 5.849316 (lr=8.6981e-06) (hash(x)=53661448) +26200 val loss 5.6577 +26200 val perplexity 286.4882 +26200 train 5.590721 (lr=1.5564e-05) (hash(x)=48588410) +29300 val loss 5.8089 +29300 val perplexity 333.2541 +29300 train 5.891900 (lr=8.6432e-06) (hash(x)=52935872) +26300 val loss 5.6548 +26300 val perplexity 285.6681 +26300 train 5.675578 (lr=1.5479e-05) (hash(x)=52286930) +29400 val loss 5.8030 +29400 val perplexity 331.3065 +29400 train 5.540308 (lr=8.5885e-06) (hash(x)=47385579) +29500 val loss 5.8053 +29500 val perplexity 332.0479 +26400 val loss 5.6507 +26400 val perplexity 284.4903 +29500 train 5.741642 (lr=8.5338e-06) (hash(x)=48758515) +26400 train 5.531841 (lr=1.5394e-05) (hash(x)=43797796) +29600 val loss 5.7983 +29600 val perplexity 329.7498 +29600 train 5.941875 (lr=8.4793e-06) (hash(x)=53373488) +26500 val loss 5.6497 +26500 val perplexity 284.2129 +26500 train 5.632333 (lr=1.5309e-05) (hash(x)=50812654) +29700 val loss 5.8037 +29700 val perplexity 331.5317 +29700 train 5.643142 (lr=8.4248e-06) (hash(x)=46924542) +26600 val loss 5.6476 +26600 val perplexity 283.6052 +26600 train 5.406253 (lr=1.5224e-05) (hash(x)=45633368) +29800 val loss 5.7988 +29800 val perplexity 329.8957 +29800 train 5.629164 (lr=8.3705e-06) (hash(x)=43935095) +26700 val loss 5.6468 +26700 val perplexity 283.3933 +26700 train 5.967397 (lr=1.5140e-05) (hash(x)=58836955) +29900 val loss 5.7960 +29900 val perplexity 328.9725 +29900 train 5.613548 (lr=8.3162e-06) (hash(x)=47165031) +26800 val loss 5.6446 +26800 val perplexity 282.7742 +26800 train 5.795603 (lr=1.5055e-05) (hash(x)=53665925) +30000 val loss 5.7977 +30000 val perplexity 329.5401 +30000 train 6.057730 (lr=8.2621e-06) (hash(x)=55187448) +26900 val loss 5.6464 +26900 val perplexity 283.2592 +26900 train 5.798853 (lr=1.4970e-05) (hash(x)=49535575) +30100 val loss 5.7909 +30100 val perplexity 327.3118 +30100 train 5.684449 (lr=8.2080e-06) (hash(x)=49106182) +27000 val loss 5.6421 +27000 val perplexity 282.0561 +27000 train 5.265175 (lr=1.4886e-05) (hash(x)=46799308) +30200 val loss 5.7917 +30200 val perplexity 327.5669 +30200 train 5.585877 (lr=8.1541e-06) (hash(x)=44482250) +27100 val loss 5.6408 +27100 val perplexity 281.6896 +27100 train 5.312801 (lr=1.4801e-05) (hash(x)=42191940) +30300 val loss 5.7943 +30300 val perplexity 328.4290 +30300 train 5.774892 (lr=8.1003e-06) (hash(x)=47186431) +27200 val loss 5.6486 +27200 val perplexity 283.9058 +27200 train 5.411514 (lr=1.4717e-05) (hash(x)=42152270) +30400 val loss 5.7892 +30400 val perplexity 326.7522 +30400 train 5.597364 (lr=8.0466e-06) (hash(x)=45796670) +27300 val loss 5.6413 +27300 val perplexity 281.8335 +27300 train 5.506339 (lr=1.4632e-05) (hash(x)=44970208) +30500 val loss 5.7874 +30500 val perplexity 326.1787 +30500 train 5.472888 (lr=7.9931e-06) (hash(x)=48580072) +27400 val loss 5.6367 +27400 val perplexity 280.5236 +27400 train 5.733561 (lr=1.4548e-05) (hash(x)=52832889) +30600 val loss 5.7889 +30600 val perplexity 326.6496 +30600 train 5.828889 (lr=7.9396e-06) (hash(x)=53974898) +27500 val loss 5.6335 +27500 val perplexity 279.6365 +30700 val loss 5.7861 +30700 val perplexity 325.7269 +30700 train 6.022291 (lr=7.8863e-06) (hash(x)=57628887) +27500 train 5.577798 (lr=1.4464e-05) (hash(x)=49318466) +30800 val loss 5.7864 +30800 val perplexity 325.8388 +30800 train 5.432889 (lr=7.8331e-06) (hash(x)=46526494) +27600 val loss 5.6325 +27600 val perplexity 279.3620 +27600 train 5.632758 (lr=1.4380e-05) (hash(x)=50216713) +30900 val loss 5.7823 +30900 val perplexity 324.5056 +30900 train 5.965190 (lr=7.7800e-06) (hash(x)=47535221) +27700 val loss 5.6279 +27700 val perplexity 278.0706 +27700 train 5.637230 (lr=1.4296e-05) (hash(x)=44599482) +31000 val loss 5.7838 +31000 val perplexity 325.0056 +31000 train 5.526973 (lr=7.7270e-06) (hash(x)=47825023) +27800 val loss 5.6223 +27800 val perplexity 276.5277 +27800 train 5.741809 (lr=1.4212e-05) (hash(x)=52853588) +31100 val loss 5.7865 +31100 val perplexity 325.8691 +31100 train 5.414888 (lr=7.6742e-06) (hash(x)=45348268) +27900 val loss 5.6490 +27900 val perplexity 284.0198 +27900 train 5.581344 (lr=1.4128e-05) (hash(x)=45297678) +31200 val loss 5.7826 +31200 val perplexity 324.6076 +31200 train 5.346751 (lr=7.6215e-06) (hash(x)=47201868) +28000 val loss 5.6221 +28000 val perplexity 276.4746 +28000 train 5.463049 (lr=1.4044e-05) (hash(x)=45648934) +31300 val loss 5.7833 +31300 val perplexity 324.8451 +31300 train 5.973914 (lr=7.5690e-06) (hash(x)=54166284) +28100 val loss 5.6211 +28100 val perplexity 276.1914 +28100 train 5.612262 (lr=1.3960e-05) (hash(x)=46498270) +31400 val loss 5.7853 +31400 val perplexity 325.4740 +31400 train 5.714035 (lr=7.5165e-06) (hash(x)=47387646) +28200 val loss 5.6235 +28200 val perplexity 276.8680 +28200 train 5.462210 (lr=1.3877e-05) (hash(x)=45066849) +31500 val loss 5.7840 +31500 val perplexity 325.0590 +31500 train 5.852404 (lr=7.4643e-06) (hash(x)=58341520) +28300 val loss 5.6184 +28300 val perplexity 275.4605 +28300 train 5.558880 (lr=1.3793e-05) (hash(x)=46542887) +31600 val loss 5.7794 +31600 val perplexity 323.5736 +31600 train 5.496296 (lr=7.4121e-06) (hash(x)=49020919) +31700 val loss 5.7754 +31700 val perplexity 322.2632 +31700 train 5.504386 (lr=7.3601e-06) (hash(x)=47216075) +28400 val loss 5.6160 +28400 val perplexity 274.7824 +28400 train 5.644979 (lr=1.3710e-05) (hash(x)=50854598) +31800 val loss 5.7720 +31800 val perplexity 321.1805 +31800 train 5.510058 (lr=7.3083e-06) (hash(x)=53041188) +28500 val loss 5.6134 +28500 val perplexity 274.0660 +28500 train 5.712130 (lr=1.3627e-05) (hash(x)=50005909) +31900 val loss 5.7697 +31900 val perplexity 320.4485 +31900 train 5.815587 (lr=7.2566e-06) (hash(x)=53464466) +28600 val loss 5.6166 +28600 val perplexity 274.9623 +28600 train 5.661458 (lr=1.3543e-05) (hash(x)=50777665) +32000 val loss 5.7649 +32000 val perplexity 318.9195 +32000 train 5.630099 (lr=7.2050e-06) (hash(x)=43954421) +28700 val loss 5.6136 +28700 val perplexity 274.1176 +28700 train 5.999113 (lr=1.3460e-05) (hash(x)=60124649) +32100 val loss 5.7676 +32100 val perplexity 319.7782 +32100 train 5.664749 (lr=7.1536e-06) (hash(x)=54546093) +28800 val loss 5.6152 +28800 val perplexity 274.5790 +28800 train 5.524859 (lr=1.3377e-05) (hash(x)=49592362) +32200 val loss 5.7645 +32200 val perplexity 318.7921 +32200 train 5.651332 (lr=7.1024e-06) (hash(x)=46553608) +28900 val loss 5.6129 +28900 val perplexity 273.9457 +28900 train 5.562673 (lr=1.3295e-05) (hash(x)=48094168) +32300 val loss 5.7631 +32300 val perplexity 318.3237 +32300 train 5.833776 (lr=7.0513e-06) (hash(x)=49034740) +29000 val loss 5.6111 +29000 val perplexity 273.4573 +29000 train 5.445677 (lr=1.3212e-05) (hash(x)=49283091) +32400 val loss 5.7602 +32400 val perplexity 317.4091 +32400 train 5.437687 (lr=7.0003e-06) (hash(x)=44275249) +29100 val loss 5.6091 +29100 val perplexity 272.8968 +29100 train 5.386805 (lr=1.3129e-05) (hash(x)=49704892) +32500 val loss 5.7612 +32500 val perplexity 317.7297 +32500 train 6.110644 (lr=6.9495e-06) (hash(x)=57029816) +29200 val loss 5.6079 +29200 val perplexity 272.5825 +29200 train 5.668581 (lr=1.3047e-05) (hash(x)=53661448) +32600 val loss 5.7577 +32600 val perplexity 316.6045 +32600 train 5.696491 (lr=6.8989e-06) (hash(x)=50987489) +29300 val loss 5.6040 +29300 val perplexity 271.5169 +29300 train 5.640203 (lr=1.2965e-05) (hash(x)=52935872) +32700 val loss 5.7607 +32700 val perplexity 317.5638 +32700 train 5.689260 (lr=6.8485e-06) (hash(x)=48694296) +29400 val loss 5.6004 +29400 val perplexity 270.5249 +29400 train 5.325976 (lr=1.2883e-05) (hash(x)=47385579) +32800 val loss 5.7598 +32800 val perplexity 317.2935 +32800 train 5.791007 (lr=6.7982e-06) (hash(x)=51471528) +32900 val loss 5.7552 +32900 val perplexity 315.8404 +32900 train 5.913572 (lr=6.7480e-06) (hash(x)=56096057) +29500 val loss 5.5991 +29500 val perplexity 270.1866 +29500 train 5.515272 (lr=1.2801e-05) (hash(x)=48758515) +33000 val loss 5.7555 +33000 val perplexity 315.9264 +33000 train 5.600418 (lr=6.6981e-06) (hash(x)=47079017) +29600 val loss 5.5936 +29600 val perplexity 268.7123 +29600 train 5.732780 (lr=1.2719e-05) (hash(x)=53373488) +33100 val loss 5.7577 +33100 val perplexity 316.6090 +33100 train 5.855866 (lr=6.6483e-06) (hash(x)=47759875) +29700 val loss 5.5938 +29700 val perplexity 268.7657 +29700 train 5.446488 (lr=1.2637e-05) (hash(x)=46924542) +33200 val loss 5.7566 +33200 val perplexity 316.2867 +33200 train 5.963431 (lr=6.5987e-06) (hash(x)=53977891) +29800 val loss 5.5879 +29800 val perplexity 267.1693 +29800 train 5.447593 (lr=1.2556e-05) (hash(x)=43935095) +33300 val loss 5.7534 +33300 val perplexity 315.2547 +33300 train 5.543291 (lr=6.5493e-06) (hash(x)=48728077) +29900 val loss 5.5899 +29900 val perplexity 267.7132 +29900 train 5.445040 (lr=1.2474e-05) (hash(x)=47165031) +33400 val loss 5.7548 +33400 val perplexity 315.7105 +33400 train 5.755011 (lr=6.5000e-06) (hash(x)=54923913) +30000 val loss 5.5918 +30000 val perplexity 268.2248 +30000 train 5.827078 (lr=1.2393e-05) (hash(x)=55187448) +33500 val loss 5.7548 +33500 val perplexity 315.6983 +33500 train 5.504075 (lr=6.4509e-06) (hash(x)=46623422) +30100 val loss 5.5820 +30100 val perplexity 265.5916 +30100 train 5.486887 (lr=1.2312e-05) (hash(x)=49106182) +33600 val loss 5.7570 +33600 val perplexity 316.3965 +33600 train 5.813240 (lr=6.4020e-06) (hash(x)=53759672) +30200 val loss 5.5845 +30200 val perplexity 266.2640 +30200 train 5.391113 (lr=1.2231e-05) (hash(x)=44482250) +33700 val loss 5.7535 +33700 val perplexity 315.2798 +33700 train 5.973464 (lr=6.3533e-06) (hash(x)=50177499) +30300 val loss 5.5836 +30300 val perplexity 266.0365 +30300 train 5.570301 (lr=1.2150e-05) (hash(x)=47186431) +33800 val loss 5.7536 +33800 val perplexity 315.3331 +33800 train 5.613473 (lr=6.3048e-06) (hash(x)=49208351) +30400 val loss 5.5816 +30400 val perplexity 265.4851 +30400 train 5.387383 (lr=1.2070e-05) (hash(x)=45796670) +33900 val loss 5.7539 +33900 val perplexity 315.4059 +33900 train 5.559392 (lr=6.2564e-06) (hash(x)=51136014) +34000 val loss 5.7518 +34000 val perplexity 314.7441 +34000 train 5.557000 (lr=6.2083e-06) (hash(x)=55323465) +30500 val loss 5.5797 +30500 val perplexity 265.0002 +30500 train 5.252225 (lr=1.1990e-05) (hash(x)=48580072) +34100 val loss 5.7466 +34100 val perplexity 313.1314 +34100 train 5.817913 (lr=6.1603e-06) (hash(x)=46512090) +30600 val loss 5.5848 +30600 val perplexity 266.3577 +30600 train 5.605900 (lr=1.1909e-05) (hash(x)=53974898) +34200 val loss 5.7516 +34200 val perplexity 314.6991 +34200 train 5.814545 (lr=6.1125e-06) (hash(x)=52513966) +30700 val loss 5.5760 +30700 val perplexity 264.0248 +30700 train 5.813101 (lr=1.1829e-05) (hash(x)=57628887) +34300 val loss 5.7438 +34300 val perplexity 312.2599 +34300 train 5.766311 (lr=6.0650e-06) (hash(x)=51181125) +30800 val loss 5.5800 +30800 val perplexity 265.0778 +30800 train 5.216499 (lr=1.1750e-05) (hash(x)=46526494) +34400 val loss 5.7396 +34400 val perplexity 310.9433 +34400 train 5.891932 (lr=6.0176e-06) (hash(x)=50979117) +30900 val loss 5.5751 +30900 val perplexity 263.7696 +30900 train 5.739004 (lr=1.1670e-05) (hash(x)=47535221) +34500 val loss 5.7540 +34500 val perplexity 315.4562 +34500 train 5.836828 (lr=5.9704e-06) (hash(x)=50394150) +31000 val loss 5.5753 +31000 val perplexity 263.8334 +31000 train 5.316584 (lr=1.1591e-05) (hash(x)=47825023) +34600 val loss 5.7406 +34600 val perplexity 311.2663 +34600 train 5.670851 (lr=5.9234e-06) (hash(x)=52666346) +31100 val loss 5.5799 +31100 val perplexity 265.0409 +31100 train 5.209934 (lr=1.1511e-05) (hash(x)=45348268) +34700 val loss 5.7349 +34700 val perplexity 309.4702 +34700 train 5.978181 (lr=5.8766e-06) (hash(x)=48275110) +31200 val loss 5.5752 +31200 val perplexity 263.8069 +31200 train 5.140681 (lr=1.1432e-05) (hash(x)=47201868) +34800 val loss 5.7343 +34800 val perplexity 309.2944 +34800 train 5.544619 (lr=5.8301e-06) (hash(x)=43282226) +31300 val loss 5.5744 +31300 val perplexity 263.5809 +31300 train 5.807706 (lr=1.1353e-05) (hash(x)=54166284) +34900 val loss 5.7328 +34900 val perplexity 308.8318 +34900 train 5.909778 (lr=5.7837e-06) (hash(x)=55252392) +31400 val loss 5.5750 +31400 val perplexity 263.7616 +31400 train 5.475660 (lr=1.1275e-05) (hash(x)=47387646) +35000 val loss 5.7319 +35000 val perplexity 308.5631 +35000 train 5.751704 (lr=5.7375e-06) (hash(x)=51882979) +31500 val loss 5.5778 +31500 val perplexity 264.4806 +31500 train 5.637706 (lr=1.1196e-05) (hash(x)=58341520) +35100 val loss 5.7310 +35100 val perplexity 308.2884 +35100 train 5.501493 (lr=5.6916e-06) (hash(x)=56053572) +31600 val loss 5.5723 +31600 val perplexity 263.0361 +31600 train 5.281992 (lr=1.1118e-05) (hash(x)=49020919) +35200 val loss 5.7310 +35200 val perplexity 308.2868 +35200 train 5.803996 (lr=5.6458e-06) (hash(x)=51979525) +31700 val loss 5.5687 +31700 val perplexity 262.0886 +31700 train 5.318092 (lr=1.1040e-05) (hash(x)=47216075) +35300 val loss 5.7320 +35300 val perplexity 308.5799 +35300 train 5.746668 (lr=5.6003e-06) (hash(x)=50873666) +31800 val loss 5.5656 +31800 val perplexity 261.2875 +31800 train 5.290979 (lr=1.0962e-05) (hash(x)=53041188) +35400 val loss 5.7332 +35400 val perplexity 308.9667 +35400 train 6.341711 (lr=5.5550e-06) (hash(x)=69087970) +31900 val loss 5.5596 +31900 val perplexity 259.7091 +31900 train 5.616481 (lr=1.0885e-05) (hash(x)=53464466) +35500 val loss 5.7288 +35500 val perplexity 307.6032 +35500 train 5.466399 (lr=5.5099e-06) (hash(x)=48654918) +32000 val loss 5.5578 +32000 val perplexity 259.2516 +32000 train 5.437356 (lr=1.0808e-05) (hash(x)=43954421) +35600 val loss 5.7263 +35600 val perplexity 306.8270 +35600 train 5.431338 (lr=5.4650e-06) (hash(x)=49987504) +35700 val loss 5.7274 +35700 val perplexity 307.1733 +35700 train 5.459028 (lr=5.4204e-06) (hash(x)=46372898) +32100 val loss 5.5546 +32100 val perplexity 258.4320 +32100 train 5.408111 (lr=1.0730e-05) (hash(x)=54546093) +35800 val loss 5.7292 +35800 val perplexity 307.7269 +35800 train 5.755822 (lr=5.3759e-06) (hash(x)=49996578) +32200 val loss 5.5540 +32200 val perplexity 258.2716 +32200 train 5.480125 (lr=1.0654e-05) (hash(x)=46553608) +35900 val loss 5.7299 +35900 val perplexity 307.9337 +35900 train 5.778760 (lr=5.3317e-06) (hash(x)=49932960) +32300 val loss 5.5517 +32300 val perplexity 257.6772 +32300 train 5.655649 (lr=1.0577e-05) (hash(x)=49034740) +36000 val loss 5.7305 +36000 val perplexity 308.1278 +36000 train 5.475699 (lr=5.2877e-06) (hash(x)=41502642) +32400 val loss 5.5481 +32400 val perplexity 256.7425 +32400 train 5.237797 (lr=1.0500e-05) (hash(x)=44275249) +36100 val loss 5.7298 +36100 val perplexity 307.9119 +36100 train 5.604872 (lr=5.2440e-06) (hash(x)=55380312) +32500 val loss 5.5495 +32500 val perplexity 257.0983 +32500 train 5.876234 (lr=1.0424e-05) (hash(x)=57029816) +36200 val loss 5.7295 +36200 val perplexity 307.8176 +36200 train 5.444288 (lr=5.2004e-06) (hash(x)=51323559) +32600 val loss 5.5474 +32600 val perplexity 256.5820 +32600 train 5.486413 (lr=1.0348e-05) (hash(x)=50987489) +36300 val loss 5.7295 +36300 val perplexity 307.8032 +36300 train 5.645262 (lr=5.1571e-06) (hash(x)=47058693) +32700 val loss 5.5514 +32700 val perplexity 257.5987 +32700 train 5.513605 (lr=1.0273e-05) (hash(x)=48694296) +36400 val loss 5.7299 +36400 val perplexity 307.9358 +36400 train 5.308862 (lr=5.1141e-06) (hash(x)=45928242) +32800 val loss 5.5472 +32800 val perplexity 256.5150 +32800 train 5.578935 (lr=1.0197e-05) (hash(x)=51471528) +36500 val loss 5.7209 +36500 val perplexity 305.1734 +36500 train 5.814166 (lr=5.0712e-06) (hash(x)=52834624) +36600 val loss 5.7193 +36600 val perplexity 304.6960 +36600 train 5.874860 (lr=5.0286e-06) (hash(x)=47195270) +32900 val loss 5.5450 +32900 val perplexity 255.9670 +32900 train 5.686032 (lr=1.0122e-05) (hash(x)=56096057) +36700 val loss 5.7198 +36700 val perplexity 304.8460 +36700 train 5.796463 (lr=4.9863e-06) (hash(x)=52501734) +33000 val loss 5.5444 +33000 val perplexity 255.8070 +33000 train 5.398359 (lr=1.0047e-05) (hash(x)=47079017) +36800 val loss 5.7193 +36800 val perplexity 304.6879 +36800 train 5.788925 (lr=4.9441e-06) (hash(x)=50333721) +33100 val loss 5.5459 +33100 val perplexity 256.1890 +33100 train 5.681064 (lr=9.9725e-06) (hash(x)=47759875) +36900 val loss 5.7146 +36900 val perplexity 303.2648 +36900 train 5.609434 (lr=4.9023e-06) (hash(x)=46841481) +33200 val loss 5.5436 +33200 val perplexity 255.5994 +33200 train 5.763474 (lr=9.8980e-06) (hash(x)=53977891) +37000 val loss 5.7132 +37000 val perplexity 302.8389 +37000 train 5.692466 (lr=4.8606e-06) (hash(x)=61838433) +33300 val loss 5.5401 +33300 val perplexity 254.7092 +33300 train 5.360518 (lr=9.8239e-06) (hash(x)=48728077) +37100 val loss 5.7118 +37100 val perplexity 302.4249 +37100 train 5.661762 (lr=4.8192e-06) (hash(x)=47537252) +33400 val loss 5.5387 +33400 val perplexity 254.3549 +33400 train 5.504649 (lr=9.7500e-06) (hash(x)=54923913) +37200 val loss 5.7104 +37200 val perplexity 302.0014 +37200 train 5.849042 (lr=4.7781e-06) (hash(x)=51840280) +33500 val loss 5.5426 +33500 val perplexity 255.3335 +33500 train 5.264757 (lr=9.6764e-06) (hash(x)=46623422) +37300 val loss 5.7117 +37300 val perplexity 302.3878 +37300 train 5.527365 (lr=4.7372e-06) (hash(x)=48870226) +33600 val loss 5.5444 +33600 val perplexity 255.8027 +33600 train 5.620683 (lr=9.6030e-06) (hash(x)=53759672) +37400 val loss 5.7113 +37400 val perplexity 302.2510 +37400 train 5.659022 (lr=4.6965e-06) (hash(x)=45398282) +33700 val loss 5.5427 +33700 val perplexity 255.3604 +33700 train 5.787031 (lr=9.5300e-06) (hash(x)=50177499) +37500 val loss 5.7089 +37500 val perplexity 301.5457 +37500 train 5.747249 (lr=4.6561e-06) (hash(x)=46278516) +33800 val loss 5.5433 +33800 val perplexity 255.5080 +33800 train 5.379657 (lr=9.4572e-06) (hash(x)=49208351) +37600 val loss 5.7084 +37600 val perplexity 301.3796 +37600 train 5.653144 (lr=4.6160e-06) (hash(x)=49553278) +37700 val loss 5.7090 +37700 val perplexity 301.5713 +37700 train 5.979125 (lr=4.5761e-06) (hash(x)=61653685) +33900 val loss 5.5437 +33900 val perplexity 255.6320 +33900 train 5.319641 (lr=9.3846e-06) (hash(x)=51136014) +37800 val loss 5.7076 +37800 val perplexity 301.1364 +37800 train 5.848799 (lr=4.5365e-06) (hash(x)=58942903) +34000 val loss 5.5386 +34000 val perplexity 254.3111 +34000 train 5.324093 (lr=9.3124e-06) (hash(x)=55323465) +37900 val loss 5.7084 +37900 val perplexity 301.3798 +37900 train 5.707924 (lr=4.4971e-06) (hash(x)=51529837) +34100 val loss 5.5355 +34100 val perplexity 253.5245 +34100 train 5.660663 (lr=9.2405e-06) (hash(x)=46512090) +38000 val loss 5.7097 +38000 val perplexity 301.7912 +38000 train 5.445275 (lr=4.4580e-06) (hash(x)=51047484) +34200 val loss 5.5390 +34200 val perplexity 254.4174 +34200 train 5.589602 (lr=9.1688e-06) (hash(x)=52513966) +38100 val loss 5.7085 +38100 val perplexity 301.4046 +38100 train 5.621894 (lr=4.4191e-06) (hash(x)=49117671) +34300 val loss 5.5279 +34300 val perplexity 251.6145 +34300 train 5.545732 (lr=9.0974e-06) (hash(x)=51181125) +38200 val loss 5.7095 +38200 val perplexity 301.7227 +38200 train 5.665784 (lr=4.3805e-06) (hash(x)=49501130) +34400 val loss 5.5272 +34400 val perplexity 251.4430 +34400 train 5.694310 (lr=9.0264e-06) (hash(x)=50979117) +38300 val loss 5.7096 +38300 val perplexity 301.7436 +38300 train 5.446144 (lr=4.3422e-06) (hash(x)=49878964) +34500 val loss 5.5377 +34500 val perplexity 254.0858 +34500 train 5.634828 (lr=8.9556e-06) (hash(x)=50394150) +38400 val loss 5.7081 +38400 val perplexity 301.3078 +38400 train 5.663073 (lr=4.3041e-06) (hash(x)=52855313) +34600 val loss 5.5224 +34600 val perplexity 250.2318 +34600 train 5.435220 (lr=8.8851e-06) (hash(x)=52666346) +38500 val loss 5.7068 +38500 val perplexity 300.9050 +38500 train 5.296468 (lr=4.2663e-06) (hash(x)=47213252) +34700 val loss 5.5200 +34700 val perplexity 249.6335 +34700 train 5.796903 (lr=8.8150e-06) (hash(x)=48275110) +38600 val loss 5.7066 +38600 val perplexity 300.8580 +38600 train 5.849113 (lr=4.2288e-06) (hash(x)=50001498) +38700 val loss 5.7086 +38700 val perplexity 301.4583 +38700 train 5.441696 (lr=4.1915e-06) (hash(x)=49541736) +34800 val loss 5.5225 +34800 val perplexity 250.2614 +34800 train 5.355512 (lr=8.7451e-06) (hash(x)=43282226) +38800 val loss 5.7047 +38800 val perplexity 300.2818 +38800 train 5.640159 (lr=4.1545e-06) (hash(x)=49712820) +34900 val loss 5.5191 +34900 val perplexity 249.4100 +34900 train 5.662911 (lr=8.6755e-06) (hash(x)=55252392) +38900 val loss 5.6991 +38900 val perplexity 298.6116 +38900 train 5.644313 (lr=4.1178e-06) (hash(x)=50913051) +35000 val loss 5.5189 +35000 val perplexity 249.3533 +35000 train 5.565717 (lr=8.6063e-06) (hash(x)=51882979) +39000 val loss 5.6999 +39000 val perplexity 298.8247 +39000 train 5.658595 (lr=4.0813e-06) (hash(x)=46110703) +35100 val loss 5.5167 +35100 val perplexity 248.8248 +35100 train 5.243548 (lr=8.5374e-06) (hash(x)=56053572) +39100 val loss 5.6976 +39100 val perplexity 298.1637 +39100 train 5.538100 (lr=4.0452e-06) (hash(x)=43887045) +35200 val loss 5.5161 +35200 val perplexity 248.6686 +35200 train 5.574739 (lr=8.4688e-06) (hash(x)=51979525) +39200 val loss 5.6974 +39200 val perplexity 298.0850 +39200 train 5.917207 (lr=4.0093e-06) (hash(x)=53102477) +35300 val loss 5.5145 +35300 val perplexity 248.2735 +35300 train 5.536997 (lr=8.4005e-06) (hash(x)=50873666) +39300 val loss 5.6965 +39300 val perplexity 297.8202 +39300 train 5.542694 (lr=3.9736e-06) (hash(x)=44607212) +35400 val loss 5.5195 +35400 val perplexity 249.5042 +35400 train 5.950989 (lr=8.3325e-06) (hash(x)=69087970) +39400 val loss 5.6994 +39400 val perplexity 298.6752 +39400 train 5.890422 (lr=3.9383e-06) (hash(x)=55054109) +39500 val loss 5.6963 +39500 val perplexity 297.7704 +39500 train 5.539431 (lr=3.9032e-06) (hash(x)=47529853) +35500 val loss 5.5137 +35500 val perplexity 248.0676 +35500 train 5.258649 (lr=8.2648e-06) (hash(x)=48654918) +39600 val loss 5.6948 +39600 val perplexity 297.3285 +39600 train 5.447984 (lr=3.8685e-06) (hash(x)=50783698) +35600 val loss 5.5109 +35600 val perplexity 247.3798 +35600 train 5.189189 (lr=8.1975e-06) (hash(x)=49987504) +39700 val loss 5.6960 +39700 val perplexity 297.6726 +39700 train 5.470551 (lr=3.8340e-06) (hash(x)=47104160) +35700 val loss 5.5132 +35700 val perplexity 247.9386 +35700 train 5.272167 (lr=8.1305e-06) (hash(x)=46372898) +39800 val loss 5.6954 +39800 val perplexity 297.4811 +39800 train 5.645205 (lr=3.7998e-06) (hash(x)=50941478) +35800 val loss 5.5121 +35800 val perplexity 247.6822 +35800 train 5.517332 (lr=8.0639e-06) (hash(x)=49996578) +39900 val loss 5.6948 +39900 val perplexity 297.3215 +39900 train 5.433103 (lr=3.7659e-06) (hash(x)=47374830) +35900 val loss 5.5186 +35900 val perplexity 249.2885 +35900 train 5.562427 (lr=7.9976e-06) (hash(x)=49932960) +40000 val loss 5.6959 +40000 val perplexity 297.6487 +40000 train 5.835879 (lr=3.7322e-06) (hash(x)=56749414) +36000 val loss 5.5145 +36000 val perplexity 248.2733 +36000 train 5.265144 (lr=7.9316e-06) (hash(x)=41502642) +40100 val loss 5.6951 +40100 val perplexity 297.4113 +40100 train 5.634251 (lr=3.6989e-06) (hash(x)=49256359) +36100 val loss 5.5123 +36100 val perplexity 247.7263 +36100 train 5.372263 (lr=7.8659e-06) (hash(x)=55380312) +40200 val loss 5.6932 +40200 val perplexity 296.8517 +40200 train 5.628206 (lr=3.6658e-06) (hash(x)=47932192) +36200 val loss 5.5135 +36200 val perplexity 248.0148 +36200 train 5.209120 (lr=7.8006e-06) (hash(x)=51323559) +40300 val loss 5.6944 +40300 val perplexity 297.1999 +40300 train 5.501371 (lr=3.6331e-06) (hash(x)=42473499) +36300 val loss 5.5156 +36300 val perplexity 248.5325 +36300 train 5.449249 (lr=7.7357e-06) (hash(x)=47058693) +40400 val loss 5.6953 +40400 val perplexity 297.4526 +40400 train 5.493920 (lr=3.6006e-06) (hash(x)=50469946) +40500 val loss 5.6916 +40500 val perplexity 296.3681 +40500 train 5.793636 (lr=3.5684e-06) (hash(x)=48818656) +36400 val loss 5.5117 +36400 val perplexity 247.5750 +36400 train 5.107434 (lr=7.6711e-06) (hash(x)=45928242) +40600 val loss 5.6874 +40600 val perplexity 295.1180 +40600 train 5.534501 (lr=3.5365e-06) (hash(x)=45215748) +36500 val loss 5.5026 +36500 val perplexity 245.3199 +36500 train 5.586413 (lr=7.6068e-06) (hash(x)=52834624) +40700 val loss 5.6864 +40700 val perplexity 294.8233 +40700 train 5.787674 (lr=3.5050e-06) (hash(x)=48057577) +36600 val loss 5.5013 +36600 val perplexity 245.0139 +36600 train 5.619504 (lr=7.5429e-06) (hash(x)=47195270) +40800 val loss 5.6853 +40800 val perplexity 294.4969 +40800 train 5.652356 (lr=3.4737e-06) (hash(x)=50103136) +36700 val loss 5.5031 +36700 val perplexity 245.4485 +36700 train 5.562798 (lr=7.4794e-06) (hash(x)=52501734) +40900 val loss 5.6835 +40900 val perplexity 293.9779 +40900 train 5.536093 (lr=3.4427e-06) (hash(x)=45764202) +36800 val loss 5.4998 +36800 val perplexity 244.6308 +36800 train 5.528430 (lr=7.4162e-06) (hash(x)=50333721) +41000 val loss 5.6827 +41000 val perplexity 293.7494 +41000 train 5.791839 (lr=3.4120e-06) (hash(x)=50724130) +36900 val loss 5.4974 +36900 val perplexity 244.0456 +36900 train 5.406976 (lr=7.3534e-06) (hash(x)=46841481) +41100 val loss 5.6820 +41100 val perplexity 293.5409 +41100 train 5.604681 (lr=3.3816e-06) (hash(x)=48739125) +37000 val loss 5.4944 +37000 val perplexity 243.3304 +37000 train 5.403760 (lr=7.2909e-06) (hash(x)=61838433) +41200 val loss 5.6823 +41200 val perplexity 293.6338 +41200 train 5.614425 (lr=3.3516e-06) (hash(x)=47473129) +37100 val loss 5.4932 +37100 val perplexity 243.0327 +37100 train 5.429741 (lr=7.2289e-06) (hash(x)=47537252) +41300 val loss 5.6818 +41300 val perplexity 293.4778 +41300 train 5.623785 (lr=3.3218e-06) (hash(x)=51032823) +37200 val loss 5.4920 +37200 val perplexity 242.7420 +37200 train 5.668897 (lr=7.1671e-06) (hash(x)=51840280) +41400 val loss 5.6871 +41400 val perplexity 295.0287 +41400 train 5.701999 (lr=3.2923e-06) (hash(x)=44411781) +37300 val loss 5.4929 +37300 val perplexity 242.9673 +37300 train 5.322398 (lr=7.1058e-06) (hash(x)=48870226) +41500 val loss 5.6816 +41500 val perplexity 293.4059 +41500 train 5.723105 (lr=3.2632e-06) (hash(x)=48630357) +37400 val loss 5.4910 +37400 val perplexity 242.4942 +37400 train 5.466489 (lr=7.0448e-06) (hash(x)=45398282) +41600 val loss 5.6794 +41600 val perplexity 292.7790 +41600 train 5.502401 (lr=3.2343e-06) (hash(x)=47836869) +41700 val loss 5.6789 +41700 val perplexity 292.6180 +41700 train 5.574647 (lr=3.2058e-06) (hash(x)=49143658) +37500 val loss 5.4907 +37500 val perplexity 242.4238 +37500 train 5.575109 (lr=6.9842e-06) (hash(x)=46278516) +41800 val loss 5.6791 +41800 val perplexity 292.6981 +41800 train 5.578754 (lr=3.1775e-06) (hash(x)=50426446) +37600 val loss 5.4910 +37600 val perplexity 242.5067 +37600 train 5.450679 (lr=6.9240e-06) (hash(x)=49553278) +41900 val loss 5.6784 +41900 val perplexity 292.4810 +41900 train 5.572034 (lr=3.1496e-06) (hash(x)=48134303) +37700 val loss 5.4908 +37700 val perplexity 242.4494 +37700 train 5.739932 (lr=6.8642e-06) (hash(x)=61653685) +42000 val loss 5.6782 +42000 val perplexity 292.4124 +42000 train 5.487007 (lr=3.1220e-06) (hash(x)=48151525) +37800 val loss 5.4895 +37800 val perplexity 242.1463 +37800 train 5.585812 (lr=6.8047e-06) (hash(x)=58942903) +42100 val loss 5.6785 +42100 val perplexity 292.5200 +42100 train 5.528406 (lr=3.0947e-06) (hash(x)=49984437) +37900 val loss 5.4904 +37900 val perplexity 242.3472 +37900 train 5.498044 (lr=6.7456e-06) (hash(x)=51529837) +42200 val loss 5.6780 +42200 val perplexity 292.3537 +42200 train 5.574935 (lr=3.0677e-06) (hash(x)=49103965) +38000 val loss 5.4899 +38000 val perplexity 242.2381 +38000 train 5.239496 (lr=6.6870e-06) (hash(x)=51047484) +42300 val loss 5.6793 +42300 val perplexity 292.7445 +42300 train 5.412384 (lr=3.0411e-06) (hash(x)=52134795) +38100 val loss 5.4874 +38100 val perplexity 241.6255 +38100 train 5.421770 (lr=6.6287e-06) (hash(x)=49117671) +42400 val loss 5.6802 +42400 val perplexity 293.0219 +42400 train 5.491049 (lr=3.0147e-06) (hash(x)=47289074) +38200 val loss 5.4871 +38200 val perplexity 241.5463 +38200 train 5.446860 (lr=6.5708e-06) (hash(x)=49501130) +42500 val loss 5.6788 +42500 val perplexity 292.6009 +42500 train 6.098083 (lr=2.9887e-06) (hash(x)=54213272) +38300 val loss 5.4911 +38300 val perplexity 242.5335 +38300 train 5.220195 (lr=6.5133e-06) (hash(x)=49878964) +42600 val loss 5.6766 +42600 val perplexity 291.9604 +42600 train 5.377842 (lr=2.9630e-06) (hash(x)=49347326) +38400 val loss 5.4881 +38400 val perplexity 241.7906 +38400 train 5.441985 (lr=6.4562e-06) (hash(x)=52855313) +42700 val loss 5.6774 +42700 val perplexity 292.1828 +42700 train 5.448061 (lr=2.9376e-06) (hash(x)=45762383) +42800 val loss 5.6756 +42800 val perplexity 291.6778 +38500 val loss 5.4894 +38500 val perplexity 242.1167 +42800 train 5.749851 (lr=2.9125e-06) (hash(x)=50377506) +38500 train 5.094051 (lr=6.3995e-06) (hash(x)=47213252) +42900 val loss 5.6714 +42900 val perplexity 290.4287 +42900 train 6.042685 (lr=2.8878e-06) (hash(x)=57332431) +38600 val loss 5.4854 +38600 val perplexity 241.1450 +38600 train 5.672081 (lr=6.3431e-06) (hash(x)=50001498) +43000 val loss 5.6715 +43000 val perplexity 290.4657 +43000 train 5.732287 (lr=2.8633e-06) (hash(x)=50599951) +38700 val loss 5.4883 +38700 val perplexity 241.8550 +38700 train 5.205518 (lr=6.2872e-06) (hash(x)=49541736) +43100 val loss 5.6710 +43100 val perplexity 290.3141 +43100 train 5.665458 (lr=2.8392e-06) (hash(x)=51139362) +38800 val loss 5.4848 +38800 val perplexity 240.9989 +38800 train 5.396961 (lr=6.2318e-06) (hash(x)=49712820) +43200 val loss 5.6697 +43200 val perplexity 289.9550 +43200 train 5.501375 (lr=2.8155e-06) (hash(x)=45759314) +38900 val loss 5.4786 +38900 val perplexity 239.5097 +38900 train 5.432703 (lr=6.1767e-06) (hash(x)=50913051) +43300 val loss 5.6702 +43300 val perplexity 290.0797 +43300 train 5.693662 (lr=2.7920e-06) (hash(x)=48251106) +39000 val loss 5.4807 +39000 val perplexity 240.0258 +39000 train 5.482339 (lr=6.1220e-06) (hash(x)=46110703) +43400 val loss 5.6694 +43400 val perplexity 289.8601 +43400 train 5.992477 (lr=2.7689e-06) (hash(x)=49941825) +39100 val loss 5.4761 +39100 val perplexity 238.9155 +39100 train 5.327767 (lr=6.0677e-06) (hash(x)=43887045) +43500 val loss 5.6688 +43500 val perplexity 289.6748 +43500 train 6.042636 (lr=2.7461e-06) (hash(x)=47282597) +39200 val loss 5.4760 +39200 val perplexity 238.8938 +39200 train 5.625887 (lr=6.0139e-06) (hash(x)=53102477) +43600 val loss 5.6687 +43600 val perplexity 289.6590 +43600 train 5.760139 (lr=2.7236e-06) (hash(x)=47311813) +43700 val loss 5.6677 +43700 val perplexity 289.3606 +43700 train 5.714488 (lr=2.7015e-06) (hash(x)=51739445) +39300 val loss 5.4751 +39300 val perplexity 238.6723 +39300 train 5.340345 (lr=5.9605e-06) (hash(x)=44607212) +43800 val loss 5.6667 +43800 val perplexity 289.0907 +43800 train 5.702378 (lr=2.6797e-06) (hash(x)=58597156) +39400 val loss 5.4777 +39400 val perplexity 239.2887 +39400 train 5.694132 (lr=5.9075e-06) (hash(x)=55054109) +43900 val loss 5.6674 +43900 val perplexity 289.2733 +43900 train 5.392615 (lr=2.6582e-06) (hash(x)=45351304) +39500 val loss 5.4767 +39500 val perplexity 239.0576 +39500 train 5.314574 (lr=5.8549e-06) (hash(x)=47529853) +44000 val loss 5.6675 +44000 val perplexity 289.3095 +44000 train 5.440125 (lr=2.6370e-06) (hash(x)=46603503) +39600 val loss 5.4746 +39600 val perplexity 238.5657 +39600 train 5.211457 (lr=5.8027e-06) (hash(x)=50783698) +44100 val loss 5.6662 +44100 val perplexity 288.9365 +44100 train 5.641182 (lr=2.6162e-06) (hash(x)=52189096) +39700 val loss 5.4746 +39700 val perplexity 238.5529 +39700 train 5.290993 (lr=5.7510e-06) (hash(x)=47104160) +44200 val loss 5.6670 +44200 val perplexity 289.1649 +44200 train 5.823875 (lr=2.5957e-06) (hash(x)=51527792) +39800 val loss 5.4758 +39800 val perplexity 238.8517 +39800 train 5.436923 (lr=5.6997e-06) (hash(x)=50941478) +44300 val loss 5.6653 +44300 val perplexity 288.6639 +44300 train 5.507483 (lr=2.5756e-06) (hash(x)=47982628) +39900 val loss 5.4755 +39900 val perplexity 238.7588 +39900 train 5.197717 (lr=5.6488e-06) (hash(x)=47374830) +44400 val loss 5.6661 +44400 val perplexity 288.8989 +44400 train 5.702556 (lr=2.5558e-06) (hash(x)=54902968) +40000 val loss 5.4759 +40000 val perplexity 238.8763 +40000 train 5.601263 (lr=5.5983e-06) (hash(x)=56749414) +44500 val loss 5.6655 +44500 val perplexity 288.7284 +44500 train 5.582448 (lr=2.5363e-06) (hash(x)=49008493) +40100 val loss 5.4750 +40100 val perplexity 238.6479 +40100 train 5.417756 (lr=5.5483e-06) (hash(x)=49256359) +44600 val loss 5.6664 +44600 val perplexity 288.9843 +44600 train 5.383864 (lr=2.5172e-06) (hash(x)=43643824) +44700 val loss 5.6677 +44700 val perplexity 289.3668 +44700 train 5.541717 (lr=2.4984e-06) (hash(x)=50465872) +40200 val loss 5.4715 +40200 val perplexity 237.8270 +40200 train 5.399641 (lr=5.4987e-06) (hash(x)=47932192) +44800 val loss 5.6661 +44800 val perplexity 288.9013 +44800 train 5.443906 (lr=2.4799e-06) (hash(x)=51962160) +40300 val loss 5.4740 +40300 val perplexity 238.4106 +40300 train 5.325750 (lr=5.4496e-06) (hash(x)=42473499) +44900 val loss 5.6669 +44900 val perplexity 289.1229 +44900 train 5.806747 (lr=2.4618e-06) (hash(x)=57370039) +40400 val loss 5.4724 +40400 val perplexity 238.0230 +40400 train 5.318840 (lr=5.4009e-06) (hash(x)=50469946) +45000 val loss 5.6668 +45000 val perplexity 289.1119 +45000 train 5.551552 (lr=2.4440e-06) (hash(x)=52648130) +40500 val loss 5.4685 +40500 val perplexity 237.1157 +40500 train 5.579536 (lr=5.3526e-06) (hash(x)=48818656) +45100 val loss 5.6835 +45100 val perplexity 293.9619 +45100 train 5.340375 (lr=2.4266e-06) (hash(x)=45155459) +40600 val loss 5.4664 +40600 val perplexity 236.6012 +40600 train 5.303145 (lr=5.3048e-06) (hash(x)=45215748) +45200 val loss 5.6621 +45200 val perplexity 287.7547 +45200 train 5.679145 (lr=2.4095e-06) (hash(x)=52304992) +40700 val loss 5.4643 +40700 val perplexity 236.1162 +40700 train 5.553162 (lr=5.2574e-06) (hash(x)=48057577) +45300 val loss 5.6613 +45300 val perplexity 287.5237 +45300 train 5.508339 (lr=2.3927e-06) (hash(x)=46456407) +40800 val loss 5.4638 +40800 val perplexity 235.9893 +40800 train 5.446899 (lr=5.2105e-06) (hash(x)=50103136) +45400 val loss 5.6591 +45400 val perplexity 286.8905 +45400 train 5.648170 (lr=2.3763e-06) (hash(x)=48325493) +40900 val loss 5.4610 +40900 val perplexity 235.3259 +40900 train 5.309520 (lr=5.1641e-06) (hash(x)=45764202) +45500 val loss 5.6582 +45500 val perplexity 286.6210 +45500 train 5.629743 (lr=2.3602e-06) (hash(x)=48869064) +41000 val loss 5.4581 +41000 val perplexity 234.6536 +41000 train 5.591974 (lr=5.1180e-06) (hash(x)=50724130) +45600 val loss 5.6585 +45600 val perplexity 286.7208 +45600 train 5.715761 (lr=2.3445e-06) (hash(x)=46077805) +41100 val loss 5.4584 +41100 val perplexity 234.7148 +41100 train 5.369337 (lr=5.0725e-06) (hash(x)=48739125) +45700 val loss 5.6572 +45700 val perplexity 286.3512 +45700 train 5.910591 (lr=2.3291e-06) (hash(x)=56831165) +45800 val loss 5.6579 +45800 val perplexity 286.5343 +45800 train 5.527074 (lr=2.3141e-06) (hash(x)=50124921) +41200 val loss 5.4596 +41200 val perplexity 235.0125 +41200 train 5.410974 (lr=5.0273e-06) (hash(x)=47473129) +45900 val loss 5.6584 +45900 val perplexity 286.6858 +45900 train 5.253099 (lr=2.2994e-06) (hash(x)=43227400) +41300 val loss 5.4583 +41300 val perplexity 234.7058 +41300 train 5.387813 (lr=4.9827e-06) (hash(x)=51032823) +46000 val loss 5.6561 +46000 val perplexity 286.0288 +46000 train 5.701273 (lr=2.2850e-06) (hash(x)=58308887) +41400 val loss 5.4648 +41400 val perplexity 236.2330 +41400 train 5.543021 (lr=4.9385e-06) (hash(x)=44411781) +46100 val loss 5.6558 +46100 val perplexity 285.9456 +46100 train 5.719070 (lr=2.2710e-06) (hash(x)=48571632) +41500 val loss 5.4559 +41500 val perplexity 234.1426 +41500 train 5.525509 (lr=4.8947e-06) (hash(x)=48630357) +46200 val loss 5.6573 +46200 val perplexity 286.3673 +46200 train 5.921195 (lr=2.2574e-06) (hash(x)=46368862) +41600 val loss 5.4568 +41600 val perplexity 234.3492 +41600 train 5.317335 (lr=4.8515e-06) (hash(x)=47836869) +46300 val loss 5.6563 +46300 val perplexity 286.0965 +46300 train 5.438584 (lr=2.2441e-06) (hash(x)=52820317) +41700 val loss 5.4543 +41700 val perplexity 233.7604 +41700 train 5.373765 (lr=4.8087e-06) (hash(x)=49143658) +46400 val loss 5.6555 +46400 val perplexity 285.8735 +46400 train 5.469899 (lr=2.2311e-06) (hash(x)=51871487) +41800 val loss 5.4554 +41800 val perplexity 234.0164 +41800 train 5.364164 (lr=4.7663e-06) (hash(x)=50426446) +46500 val loss 5.6554 +46500 val perplexity 285.8247 +46500 train 5.608011 (lr=2.2185e-06) (hash(x)=47867973) +41900 val loss 5.4543 +41900 val perplexity 233.7697 +41900 train 5.383428 (lr=4.7244e-06) (hash(x)=48134303) +46600 val loss 5.6547 +46600 val perplexity 285.6374 +46600 train 5.327915 (lr=2.2062e-06) (hash(x)=50685713) +42000 val loss 5.4535 +42000 val perplexity 233.5839 +42000 train 5.288549 (lr=4.6830e-06) (hash(x)=48151525) +46700 val loss 5.6542 +46700 val perplexity 285.4811 +46700 train 5.784547 (lr=2.1943e-06) (hash(x)=50975526) +46800 val loss 5.6553 +46800 val perplexity 285.8151 +46800 train 5.804982 (lr=2.1828e-06) (hash(x)=51012129) +42100 val loss 5.4546 +42100 val perplexity 233.8253 +42100 train 5.267159 (lr=4.6421e-06) (hash(x)=49984437) +46900 val loss 5.6566 +46900 val perplexity 286.1624 +46900 train 5.392484 (lr=2.1716e-06) (hash(x)=51205353) +42200 val loss 5.4550 +42200 val perplexity 233.9277 +42200 train 5.354021 (lr=4.6016e-06) (hash(x)=49103965) +47000 val loss 5.6572 +47000 val perplexity 286.3344 +47000 train 5.791560 (lr=2.1607e-06) (hash(x)=48939661) +42300 val loss 5.4557 +42300 val perplexity 234.0970 +42300 train 5.182218 (lr=4.5616e-06) (hash(x)=52134795) +47100 val loss 5.6559 +47100 val perplexity 285.9825 +47100 train 5.596141 (lr=2.1502e-06) (hash(x)=51002557) +42400 val loss 5.4556 +42400 val perplexity 234.0598 +42400 train 5.271979 (lr=4.5221e-06) (hash(x)=47289074) +47200 val loss 5.6549 +47200 val perplexity 285.6822 +47200 train 5.476904 (lr=2.1400e-06) (hash(x)=43422209) +42500 val loss 5.4553 +42500 val perplexity 233.9948 +42500 train 5.886106 (lr=4.4830e-06) (hash(x)=54213272) +47300 val loss 5.6557 +47300 val perplexity 285.9208 +47300 train 5.405007 (lr=2.1302e-06) (hash(x)=45870552) +42600 val loss 5.4529 +42600 val perplexity 233.4239 +42600 train 5.148093 (lr=4.4445e-06) (hash(x)=49347326) +47400 val loss 5.6559 +47400 val perplexity 285.9607 +47400 train 5.509630 (lr=2.1208e-06) (hash(x)=47001810) +42700 val loss 5.4532 +42700 val perplexity 233.4969 +42700 train 5.247355 (lr=4.4064e-06) (hash(x)=45762383) +47500 val loss 5.6522 +47500 val perplexity 284.9274 +47500 train 5.699212 (lr=2.1117e-06) (hash(x)=49146619) +42800 val loss 5.4508 +42800 val perplexity 232.9459 +42800 train 5.512878 (lr=4.3688e-06) (hash(x)=50377506) +47600 val loss 5.6514 +47600 val perplexity 284.6997 +47600 train 5.739098 (lr=2.1030e-06) (hash(x)=53742853) +42900 val loss 5.4462 +42900 val perplexity 231.8792 +42900 train 5.756657 (lr=4.3316e-06) (hash(x)=57332431) +47700 val loss 5.6507 +47700 val perplexity 284.4972 +47700 train 5.884995 (lr=2.0946e-06) (hash(x)=53019458) +43000 val loss 5.4452 +43000 val perplexity 231.6546 +43000 train 5.520500 (lr=4.2950e-06) (hash(x)=50599951) +47800 val loss 5.6509 +47800 val perplexity 284.5344 +47800 train 5.702728 (lr=2.0865e-06) (hash(x)=54497914) +43100 val loss 5.4440 +43100 val perplexity 231.3621 +43100 train 5.392635 (lr=4.2589e-06) (hash(x)=51139362) +47900 val loss 5.6499 +47900 val perplexity 284.2704 +47900 train 5.541301 (lr=2.0789e-06) (hash(x)=47109012) +43200 val loss 5.4437 +43200 val perplexity 231.3060 +43200 train 5.259689 (lr=4.2232e-06) (hash(x)=45759314) +48000 val loss 5.6481 +48000 val perplexity 283.7499 +48000 train 5.637836 (lr=2.0715e-06) (hash(x)=48429816) +48100 val loss 5.6484 +48100 val perplexity 283.8308 +48100 train 5.581638 (lr=2.0646e-06) (hash(x)=46500465) +43300 val loss 5.4432 +43300 val perplexity 231.1922 +43300 train 5.459748 (lr=4.1880e-06) (hash(x)=48251106) +48200 val loss 5.6473 +48200 val perplexity 283.5217 +48200 train 5.836461 (lr=2.0580e-06) (hash(x)=50888195) +43400 val loss 5.4429 +43400 val perplexity 231.1138 +43400 train 5.798908 (lr=4.1533e-06) (hash(x)=49941825) +48300 val loss 5.6485 +48300 val perplexity 283.8764 +48300 train 5.720854 (lr=2.0517e-06) (hash(x)=56406586) +43500 val loss 5.4414 +43500 val perplexity 230.7735 +43500 train 5.831889 (lr=4.1191e-06) (hash(x)=47282597) +48400 val loss 5.6467 +48400 val perplexity 283.3670 +48400 train 5.747758 (lr=2.0458e-06) (hash(x)=48713969) +43600 val loss 5.4416 +43600 val perplexity 230.8219 +43600 train 5.549546 (lr=4.0854e-06) (hash(x)=47311813) +48500 val loss 5.6468 +48500 val perplexity 283.3939 +48500 train 5.049244 (lr=2.0403e-06) (hash(x)=37691696) +43700 val loss 5.4418 +43700 val perplexity 230.8594 +43700 train 5.497785 (lr=4.0522e-06) (hash(x)=51739445) +48600 val loss 5.6477 +48600 val perplexity 283.6286 +48600 train 5.921036 (lr=2.0351e-06) (hash(x)=48712321) +43800 val loss 5.4404 +43800 val perplexity 230.5303 +43800 train 5.463239 (lr=4.0195e-06) (hash(x)=58597156) +48700 val loss 5.6468 +48700 val perplexity 283.3956 +48700 train 6.096560 (lr=2.0302e-06) (hash(x)=63513096) +43900 val loss 5.4396 +43900 val perplexity 230.3588 +43900 train 5.153522 (lr=3.9873e-06) (hash(x)=45351304) +48800 val loss 5.6471 +48800 val perplexity 283.4566 +48800 train 5.272035 (lr=2.0258e-06) (hash(x)=44917721) +44000 val loss 5.4412 +44000 val perplexity 230.7221 +44000 train 5.234890 (lr=3.9556e-06) (hash(x)=46603503) +48900 val loss 5.6466 +48900 val perplexity 283.3162 +48900 train 5.860542 (lr=2.0217e-06) (hash(x)=51426559) +44100 val loss 5.4382 +44100 val perplexity 230.0337 +44100 train 5.368792 (lr=3.9243e-06) (hash(x)=52189096) +49000 val loss 5.6461 +49000 val perplexity 283.1849 +49000 train 5.565176 (lr=2.0179e-06) (hash(x)=47909690) +49100 val loss 5.6466 +49100 val perplexity 283.3275 +49100 train 5.546085 (lr=2.0145e-06) (hash(x)=48329226) +44200 val loss 5.4398 +44200 val perplexity 230.3895 +44200 train 5.605023 (lr=3.8936e-06) (hash(x)=51527792) +49200 val loss 5.6457 +49200 val perplexity 283.0739 +49200 train 5.610006 (lr=2.0115e-06) (hash(x)=49512472) +44300 val loss 5.4381 +44300 val perplexity 230.0032 +44300 train 5.309308 (lr=3.8634e-06) (hash(x)=47982628) +49300 val loss 5.6468 +49300 val perplexity 283.3733 +49300 train 5.499342 (lr=2.0088e-06) (hash(x)=46541146) +44400 val loss 5.4394 +44400 val perplexity 230.2962 +44400 train 5.438356 (lr=3.8337e-06) (hash(x)=54902968) +49400 val loss 5.6474 +49400 val perplexity 283.5584 +49400 train 5.602351 (lr=2.0064e-06) (hash(x)=52192792) +44500 val loss 5.4380 +44500 val perplexity 229.9736 +44500 train 5.389482 (lr=3.8045e-06) (hash(x)=49008493) +49500 val loss 5.6466 +49500 val perplexity 283.3275 +49500 train 5.442019 (lr=2.0045e-06) (hash(x)=48954331) +44600 val loss 5.4405 +44600 val perplexity 230.5472 +44600 train 5.178605 (lr=3.7758e-06) (hash(x)=43643824) +49600 val loss 5.6456 +49600 val perplexity 283.0495 +49600 train 5.770230 (lr=2.0029e-06) (hash(x)=50364098) +44700 val loss 5.4393 +44700 val perplexity 230.2883 +49700 val loss 5.6453 +49700 val perplexity 282.9691 +49700 train 5.596664 (lr=2.0016e-06) (hash(x)=52016774) +44700 train 5.284790 (lr=3.7476e-06) (hash(x)=50465872) +49800 val loss 5.6447 +49800 val perplexity 282.7817 +49800 train 5.875211 (lr=2.0007e-06) (hash(x)=54182957) +44800 val loss 5.4400 +44800 val perplexity 230.4324 +44800 train 5.186536 (lr=3.7199e-06) (hash(x)=51962160) +49900 val loss 5.6430 +49900 val perplexity 282.2951 +49900 train 5.375270 (lr=2.0002e-06) (hash(x)=48188126) +44900 val loss 5.4388 +44900 val perplexity 230.1630 +44900 train 5.511966 (lr=3.6927e-06) (hash(x)=57370039) +49999 val loss 5.6534 +49999 val perplexity 285.2544 +45000 val loss 5.4385 +45000 val perplexity 230.0898 +45000 train 5.305811 (lr=3.6660e-06) (hash(x)=52648130) +45100 val loss 5.4585 +45100 val perplexity 234.7465 +45100 train 5.155573 (lr=3.6398e-06) (hash(x)=45155459) +45200 val loss 5.4339 +45200 val perplexity 229.0419 +45200 train 5.446787 (lr=3.6142e-06) (hash(x)=52304992) +45300 val loss 5.4336 +45300 val perplexity 228.9677 +45300 train 5.286599 (lr=3.5891e-06) (hash(x)=46456407) +45400 val loss 5.4316 +45400 val perplexity 228.5157 +45400 train 5.423881 (lr=3.5644e-06) (hash(x)=48325493) +45500 val loss 5.4296 +45500 val perplexity 228.0468 +45500 train 5.419282 (lr=3.5403e-06) (hash(x)=48869064) +45600 val loss 5.4293 +45600 val perplexity 227.9918 +45600 train 5.524062 (lr=3.5167e-06) (hash(x)=46077805) +45700 val loss 5.4285 +45700 val perplexity 227.8026 +45700 train 5.671704 (lr=3.4936e-06) (hash(x)=56831165) +45800 val loss 5.4286 +45800 val perplexity 227.8312 +45800 train 5.283981 (lr=3.4711e-06) (hash(x)=50124921) +45900 val loss 5.4283 +45900 val perplexity 227.7593 +45900 train 5.058990 (lr=3.4490e-06) (hash(x)=43227400) +46000 val loss 5.4285 +46000 val perplexity 227.8043 +46000 train 5.393814 (lr=3.4275e-06) (hash(x)=58308887) +46100 val loss 5.4270 +46100 val perplexity 227.4579 +46100 train 5.502267 (lr=3.4065e-06) (hash(x)=48571632) +46200 val loss 5.4279 +46200 val perplexity 227.6668 +46200 train 5.741102 (lr=3.3860e-06) (hash(x)=46368862) +46300 val loss 5.4267 +46300 val perplexity 227.3910 +46300 train 5.187996 (lr=3.3661e-06) (hash(x)=52820317) +46400 val loss 5.4267 +46400 val perplexity 227.4003 +46400 train 5.245609 (lr=3.3466e-06) (hash(x)=51871487) +46500 val loss 5.4269 +46500 val perplexity 227.4543 +46500 train 5.443164 (lr=3.3277e-06) (hash(x)=47867973) +46600 val loss 5.4260 +46600 val perplexity 227.2440 +46600 train 5.115084 (lr=3.3093e-06) (hash(x)=50685713) +46700 val loss 5.4269 +46700 val perplexity 227.4542 +46700 train 5.573802 (lr=3.2915e-06) (hash(x)=50975526) +46800 val loss 5.4274 +46800 val perplexity 227.5538 +46800 train 5.555970 (lr=3.2741e-06) (hash(x)=51012129) +46900 val loss 5.4278 +46900 val perplexity 227.6414 +46900 train 5.155362 (lr=3.2573e-06) (hash(x)=51205353) +47000 val loss 5.4297 +47000 val perplexity 228.0747 +47000 train 5.553667 (lr=3.2410e-06) (hash(x)=48939661) +47100 val loss 5.4262 +47100 val perplexity 227.2777 +47100 train 5.373991 (lr=3.2253e-06) (hash(x)=51002557) +47200 val loss 5.4273 +47200 val perplexity 227.5327 +47200 train 5.283518 (lr=3.2101e-06) (hash(x)=43422209) +47300 val loss 5.4271 +47300 val perplexity 227.4949 +47300 train 5.126106 (lr=3.1954e-06) (hash(x)=45870552) +47400 val loss 5.4274 +47400 val perplexity 227.5626 +47400 train 5.285728 (lr=3.1812e-06) (hash(x)=47001810) +47500 val loss 5.4242 +47500 val perplexity 226.8370 +47500 train 5.470029 (lr=3.1675e-06) (hash(x)=49146619) +47600 val loss 5.4219 +47600 val perplexity 226.3114 +47600 train 5.504127 (lr=3.1544e-06) (hash(x)=53742853) +47700 val loss 5.4214 +47700 val perplexity 226.1848 +47700 train 5.598899 (lr=3.1419e-06) (hash(x)=53019458) +47800 val loss 5.4200 +47800 val perplexity 225.8824 +47800 train 5.460047 (lr=3.1298e-06) (hash(x)=54497914) +47900 val loss 5.4213 +47900 val perplexity 226.1619 +47900 train 5.314313 (lr=3.1183e-06) (hash(x)=47109012) +48000 val loss 5.4181 +48000 val perplexity 225.4470 +48000 train 5.434885 (lr=3.1073e-06) (hash(x)=48429816) +48100 val loss 5.4178 +48100 val perplexity 225.3796 +48100 train 5.395557 (lr=3.0969e-06) (hash(x)=46500465) +48200 val loss 5.4163 +48200 val perplexity 225.0500 +48200 train 5.633785 (lr=3.0869e-06) (hash(x)=50888195) +48300 val loss 5.4163 +48300 val perplexity 225.0465 +48300 train 5.458313 (lr=3.0776e-06) (hash(x)=56406586) +48400 val loss 5.4162 +48400 val perplexity 225.0163 +48400 train 5.490232 (lr=3.0687e-06) (hash(x)=48713969) diff --git a/attention_kindselective_n_heads4_seed1339/model_02500.pt b/attention_kindselective_n_heads4_seed1339/model_02500.pt index c5c23b5330c3323c23f3dca08c9bd5e4428946a2..a380769216da4d1b33d086d703bed20d6a620fce 100644 --- a/attention_kindselective_n_heads4_seed1339/model_02500.pt +++ b/attention_kindselective_n_heads4_seed1339/model_02500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:2ec7853f39c020bf0817d8a9424f0786ad4b93fc0839e1f8734dc36c9b628b3c +oid sha256:3d5fec9045af2708a11b1103c045ec7a59a9b1e48b4bc322a7fb0d864f4fe143 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_05000.pt b/attention_kindselective_n_heads4_seed1339/model_05000.pt index 8389fc41740ac0d7ccd09639c00cfee3ba3ae3e0..737bab7bb024680ff2c623a2bc5b0e62626fa60c 100644 --- a/attention_kindselective_n_heads4_seed1339/model_05000.pt +++ b/attention_kindselective_n_heads4_seed1339/model_05000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:e7a3ea4733ac190d2c44f57c3fc804dc59f7571ad4f028f02fe4775d0b66df1b +oid sha256:6425c5316fad4e1ed4ca5670f21cacaad297f0b0c03153ad6258b436a6c27271 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_07500.pt b/attention_kindselective_n_heads4_seed1339/model_07500.pt index 52b87a5e7cb4c597ff25cefe62a4e9864b13af62..c3c2ae5541dc00a2715429f9f919d33550b02a6b 100644 --- a/attention_kindselective_n_heads4_seed1339/model_07500.pt +++ b/attention_kindselective_n_heads4_seed1339/model_07500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:5d9cba9958a987856e63371998d573139f4d7e12bf83d1dc6741e08c1656ae1e +oid sha256:fe251dc7a96a35307a4218cddf07ddef72e020cd1dcd16c06fd8a07f476d2b3d size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_10000.pt b/attention_kindselective_n_heads4_seed1339/model_10000.pt index a9e8b33a0b159e863b7454403ebb45916c2c0bb2..b5791faf7fb153898d276507ee61757e32ea6369 100644 --- a/attention_kindselective_n_heads4_seed1339/model_10000.pt +++ b/attention_kindselective_n_heads4_seed1339/model_10000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:75c9a83ca6838f6643993a01ade7f926d7dccd9aaabbaa9d1cc052a66d45ad6a +oid sha256:5fd919e371e6ea4001f6469599129e20ba9c35d2917a86faebec778aa5af7d62 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_12500.pt b/attention_kindselective_n_heads4_seed1339/model_12500.pt new file mode 100644 index 0000000000000000000000000000000000000000..6e0e11fb453a3df0e816733c9218b4245a458f8f --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_12500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e90652aa199388014ee309da42595a880d3c574892a0112cc9949e2b8638ce26 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_15000.pt b/attention_kindselective_n_heads4_seed1339/model_15000.pt new file mode 100644 index 0000000000000000000000000000000000000000..67396b76a2ddc2b7f3771f47ba25b5d8bd8372dc --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_15000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:51bc84a3106488453a5165f744f925a378d18cfbb260c8d6172c8542270e41b8 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_17500.pt b/attention_kindselective_n_heads4_seed1339/model_17500.pt new file mode 100644 index 0000000000000000000000000000000000000000..ce566936131372c700f872137f74185a2a9a6423 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_17500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b65470ea1dbc614d43c8967a5647ec824d520c4dcbc8bf8bb997eba0e8a8588 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_20000.pt b/attention_kindselective_n_heads4_seed1339/model_20000.pt new file mode 100644 index 0000000000000000000000000000000000000000..729e71d12ad64334d597d39e2bfad0edc7f0e06a --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_20000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cdd979d4e03f11a0456610a83b22b5e55182f12ee739d1b7f61c723f42c42f48 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_22500.pt b/attention_kindselective_n_heads4_seed1339/model_22500.pt new file mode 100644 index 0000000000000000000000000000000000000000..c33bb940a2257208f348d50bd4424bcd66902789 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_22500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a871162895775d404e04532e5cd853fb0824250f90d37bb1d22bd0f3c24629e +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_25000.pt b/attention_kindselective_n_heads4_seed1339/model_25000.pt new file mode 100644 index 0000000000000000000000000000000000000000..9f08ecfcc8eca77cf010618a9c2dca97475dcf59 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_25000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a83e359f8b0060fada80673e3c5ff586fda7a9daab15b2effb38085e33a9b42b +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_27500.pt b/attention_kindselective_n_heads4_seed1339/model_27500.pt new file mode 100644 index 0000000000000000000000000000000000000000..93f0b6ab3d3c71918446b703b4104b1452558dfb --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_27500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ffc6f7f6a861696e3c363d8c66639412eb48e3b6182b3c073248470ea7b0ffa +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_30000.pt b/attention_kindselective_n_heads4_seed1339/model_30000.pt new file mode 100644 index 0000000000000000000000000000000000000000..4cc2346401e5667fbe16af79e0bbe6ff149454fc --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_30000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d07b7c5cc6f74f1959a7fb18d77e008f804277c518ea7b34db9812b6e50c9907 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_32500.pt b/attention_kindselective_n_heads4_seed1339/model_32500.pt new file mode 100644 index 0000000000000000000000000000000000000000..c5ec66fa170ff618370b0846f2f1f8d5cf1a80a3 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_32500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:12d6178d05edaf2390b6f89ec425795f99ea95c3ed1e2b33f4a97309d1531997 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_35000.pt b/attention_kindselective_n_heads4_seed1339/model_35000.pt new file mode 100644 index 0000000000000000000000000000000000000000..09c9505d1a0387b53f0c9ab0fcc43b0683a4650b --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_35000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e3b75d837f318b1c757629de47cc740ccfa4f287f25ac4e0c4333eb9c78a5ee +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_37500.pt b/attention_kindselective_n_heads4_seed1339/model_37500.pt new file mode 100644 index 0000000000000000000000000000000000000000..429dc4708f7cbc481d84d63663254ec811403fa7 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/model_37500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca6ccb8895713bd20d9694a58e9172efa70ca0fedb143bf06d7407e5e1e38a5e +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_40000.pt b/attention_kindselective_n_heads4_seed1339/model_40000.pt index 61cf3a8d0c7a22bcdf5872161f38e8cf862c104b..ed8ce3c733b254493ba10ab91657ffd00656301b 100644 --- a/attention_kindselective_n_heads4_seed1339/model_40000.pt +++ b/attention_kindselective_n_heads4_seed1339/model_40000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:78a8fd9d6a750808245efe7c72bed034116984b1a3d236604a1e986d60e7ed58 +oid sha256:a533c25bdb31ab0d17973befa3c8c3dcdcb5d641b8659f6aa9f3b8facc3b2780 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_42500.pt b/attention_kindselective_n_heads4_seed1339/model_42500.pt index 287ee60663b097cfcc3fa512dfd065791250067a..16d19e89903a49fd1c24e01ff328350fa4da28d9 100644 --- a/attention_kindselective_n_heads4_seed1339/model_42500.pt +++ b/attention_kindselective_n_heads4_seed1339/model_42500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c3363e83947e45c3c64503ecbdcc0b270aa74e74aa9ec5438a596bd3c8156913 +oid sha256:a19c01365e93545b970a342c34ae9c9fa346575249ff72f728858556a727cb65 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_45000.pt b/attention_kindselective_n_heads4_seed1339/model_45000.pt index 13167ed7e3753e1dce10c3863c23a55886c2f532..f35e6e1f9f416dde8e3209bd2e5fe0478d1dd0db 100644 --- a/attention_kindselective_n_heads4_seed1339/model_45000.pt +++ b/attention_kindselective_n_heads4_seed1339/model_45000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:652267ce4a3c3eb24b3fdd1182cbeaad71435a28d6533b8d65f60dfd783e5411 +oid sha256:d5a557feb627ced00d55c191957ef9d595e11f01512e3b8d6ca00283170fe997 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_47500.pt b/attention_kindselective_n_heads4_seed1339/model_47500.pt index 0b23b1b9e8b4fdf6cdfd28ef2456715fdb311042..a924ae7fc88994bf89ce62accb8dfa8327e6df35 100644 --- a/attention_kindselective_n_heads4_seed1339/model_47500.pt +++ b/attention_kindselective_n_heads4_seed1339/model_47500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d198491e4dd117689476940ce9f230be59d99c41399afc7f16477ab0abe13f73 +oid sha256:2518192db956074cf791c8857f72dae7cb8006abf2e21173b4251ce253896af9 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/model_49999.pt b/attention_kindselective_n_heads4_seed1339/model_49999.pt index 6758f7a8409333654a03b9f237c32dbc3127bae7..703e775fb56e8bcf2529244f1121f60d52b382cf 100644 --- a/attention_kindselective_n_heads4_seed1339/model_49999.pt +++ b/attention_kindselective_n_heads4_seed1339/model_49999.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:bb07662f40aa23921523b0c43b25a496ce763a42d93ca236f9ab6e69e0f9407d +oid sha256:40c1c361a207f5c416fef30c79e06b4edc3315c07f7d7cb1c6deccac65695913 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_02500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_02500.pt index da1e1d9db4c678be7a760cf8da6afd420be81af1..d47d475fce67018eeac6c5d8e99663f037314d27 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_02500.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_02500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:6d0daaea20dfea170f7f2a3011fad8c68fd0d34d7837cfc7e1b82e10c8afac2e +oid sha256:cd90610bdfd7254edfb1fbed624dad17d4f7fb821d7e8099ff353419a048f679 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_05000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_05000.pt index 69faf7e5c34efc672e99edc1fb1a495bc42ded03..17771cd2e3450e5015d8f21eb4b8fa07c60626a2 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_05000.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_05000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d6b399697378d1367510c9f4104e498e29bd1617ea27a7878564f423ee87160d +oid sha256:418298de7332b6ca0f7c05bced21479834ce4b6d543e9849da08709fe1de4027 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_07500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_07500.pt index fdb9e4d5a20e0fb00db560871e225d5bce4c57ea..5bce51fcbe49dd797965f22c5f077a1352b9f4ee 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_07500.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_07500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:827fa057c0c4bae82594e7d86c915d39ef2208d8da842bda634d1d07f8a6f14c +oid sha256:6cff77e448d0e70d4643a97f8c9274780bb2ed5a0a6cce01879bf1ecf907295f size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_10000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_10000.pt index f387986da3d00d8ebe925f4f2e0efb29811de2c0..4b6caeb4612e586a0102b9a58a412b8478796375 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_10000.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_10000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:bdb56e2388c4d739ff044d95f0f99d11044dc7064e262ea9e0da4677e4b7369f +oid sha256:aa0a87a2a81beb289e8e65a4ab7fc951d274bd263127fca5d51aa21d13888550 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_12500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_12500.pt new file mode 100644 index 0000000000000000000000000000000000000000..55b11baf3e48a2adb18d71b4d45e2d0a866a6813 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_12500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d7a0ba73b14ef416ec076c5d37e507cfede5fbb34783c0f7b600aea887abb29 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_15000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_15000.pt new file mode 100644 index 0000000000000000000000000000000000000000..59d295d82961ac69458f68dd35588d75c6a7f675 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_15000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:32d28686675b5b55348834047127817f0e2c3dc674e7bd10f22d92b6f149a217 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_17500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_17500.pt new file mode 100644 index 0000000000000000000000000000000000000000..e93304d5ca51d78a02c4b605b9e21e3187677302 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_17500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:03b5777fddf8f8c914472aa5eafe22ceb2cfb59f26d567878cdd4884ab50c479 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_20000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_20000.pt new file mode 100644 index 0000000000000000000000000000000000000000..a26a38802e28b5834652d1d53f0e91c24123dbbf --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_20000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8292f83f4d2bc0af27f2aa4f429fcb06e230b506e186df8de062a95dfb5dfeb3 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_22500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_22500.pt new file mode 100644 index 0000000000000000000000000000000000000000..1254df95c44059df9eb574be7e74ec9b015eefac --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_22500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dfaac188309cc159a685e295d76251abe6a9406bfdd79ba55f666a22afd78108 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_25000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_25000.pt new file mode 100644 index 0000000000000000000000000000000000000000..20824bbd775c47ec47307a3a517125242c2a2246 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_25000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:606eede0cf315db37d8147a825918ca71a1293450df3a0536665e7f7d7229a74 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_27500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_27500.pt new file mode 100644 index 0000000000000000000000000000000000000000..77b4e9e024f602271efcbd599cfa680409cc65ff --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_27500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c87bdc27376baa7605f19f576e1e3b4cd128c4df6940579b3c3c3d083cea83f +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_30000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_30000.pt new file mode 100644 index 0000000000000000000000000000000000000000..e78fe8708485a9766f5a93bf3b587c2d071550e0 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_30000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3effc948bf127f1c27e7f4cb83d15688cd556614fc32ec8f5f3a2e192a0c8f97 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_32500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_32500.pt new file mode 100644 index 0000000000000000000000000000000000000000..cdc944c58688116b8b550d060c09e2c41941b343 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_32500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87f585f51d45083438bd55a44b333c2af592eebc7d797aaf73d8aa401c1082f9 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_35000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_35000.pt new file mode 100644 index 0000000000000000000000000000000000000000..4c7f15aa3a2049b84461498c85980fe8b96695fa --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_35000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:78f5449827bf9b6cf3ccb135222d294071d320102418282e681f5cb76d9b8932 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_37500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_37500.pt new file mode 100644 index 0000000000000000000000000000000000000000..187bc24f7c42e864b80cab4adc7db26434b03da8 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1339/optimizer_37500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1fe8ff6437f7738f84d63aea8ab7b27705ca04186f5bf66c11d59cdcbb788622 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_40000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_40000.pt index 8c00b4015e6276dc5f0601be8fc1acf704ef8334..97da2e2fd38cc413355cc71da38bcd3d104a3811 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_40000.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_40000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:7ecc8125e65f3bc259ac039c34b488e49892d0778ea59a14ffd86b25c567c0ac +oid sha256:4cdb59a59ed901dc5c16544ba72a3310a0fade3e062facba26115f016a6d092f size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_42500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_42500.pt index 27817d3f5dad4033eca10b12cd9e7eae29acf348..7c6182be156dc5188de51aa367bee4179851561e 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_42500.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_42500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a22bc9fe743fba8fc4b4560471a341de420ff16d5fb5457b184c26c331fefcdb +oid sha256:aa8fb26bf4fe7759cd2b76c5ed8febaccf16c0a57fb589c13a33a961029cde2e size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_45000.pt b/attention_kindselective_n_heads4_seed1339/optimizer_45000.pt index f61eae6efe238e0db1971efe138bbbcf68e43dae..329b6a906ecfe8e364e28a35b6effe46ec211aaa 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_45000.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_45000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:f9b0a4a35f4948303d8fd2c63401b8e68948d5d7ffed4e82c8947d42f9eace1f +oid sha256:6a30b197f02bc2f1a2ec7a7bccf42eed64f10856a638c92ee6cde0527c6cc9d3 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_47500.pt b/attention_kindselective_n_heads4_seed1339/optimizer_47500.pt index fab3936158d5bb1e0d581d0db080a7f91e416033..4b6b827aaeeb97a6ca3740d187d9b717af667af3 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_47500.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_47500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ca09c067aab98fdf7e31137500fc5459871aa21e3a7a17c7d424ef21a92d9905 +oid sha256:0ac68b3ad90d9064ad744aaad49bbb5e029c5ac5a5c18e021d06fb313c0b7a91 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1339/optimizer_49999.pt b/attention_kindselective_n_heads4_seed1339/optimizer_49999.pt index 98c436535c38b140178d522688dc8ce599f72756..cf9e20dd61772b4346930b901831b59b778575ac 100644 --- a/attention_kindselective_n_heads4_seed1339/optimizer_49999.pt +++ b/attention_kindselective_n_heads4_seed1339/optimizer_49999.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:67ac5a0dfae5b7e50ccf39b379788bfee75cb4871c0743d86aacdb6f3b54b940 +oid sha256:054ca24fbbaf1151d78a62b65441c4c2fdf18e5ea7b4247415bf7908d0a62dbd size 179406214