diff --git a/attention_kindselective_n_heads2_seed1341/args.json b/attention_kindselective_n_heads2_seed1341/args.json
index 7b7c266e28d627250970e1817493006b1c1e300d..851c1a076be2c0fb3ef71bd181fbff7fa689a696 100644
--- a/attention_kindselective_n_heads2_seed1341/args.json
+++ b/attention_kindselective_n_heads2_seed1341/args.json
@@ -1 +1 @@
-{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_6/attention_kindselective_n_heads2_seed1341", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 2, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 10000, "warmup_steps": 200, "group": "wider_is_better_6", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1341, "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": 0.00015, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "15e-5_10240_2_1341", "n_embd": 128}
\ No newline at end of file
+{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads2_seed1341", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 2, "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": 1341, "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_2_1341", "n_embd": 128}
\ No newline at end of file
diff --git a/attention_kindselective_n_heads2_seed1341/dataloader_10000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_10000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e1452fd777edd0ccb65d6b47710c50208c14b312
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/dataloader_10000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3858f6c832feea78a674d8c5c384061cc7d4f22cddbd0a2be6de33bc91e2c72
+size 964
diff --git a/attention_kindselective_n_heads2_seed1341/dataloader_12500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_12500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f63806424a0177a7f2d678c2c63138219ed021f3
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_15000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_15000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..0ea0f5a1bfab75667c4ebb0ca01b358cdc836a54
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_17500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_17500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..eb9392348b1209c827e3e376b05eeda80e779aa8
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_20000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_20000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ba32fca059aec40e4f758de96bdac6df23b9d9f5
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_22500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_22500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..498dc444f528d893090328a5bd1e2f37da46dc12
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_25000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_25000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b657bc134192d0ea956f984c289d0c682979a1f4
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_27500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_27500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d92d43d89390714f43db4f0782e49af0145b4a90
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_30000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_30000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..06856f53253b6c8cbefd7d595d9b9b7266b22621
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_32500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_32500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..dc0b4b40e57f41aa1046b3bd2697256635160c09
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_35000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_35000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..62647b540fa7925361626b9f8dfa3959eebb7608
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_37500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_37500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..475d8b538138a8e39b76a4cf04c8eaeac074d295
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/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_heads2_seed1341/dataloader_40000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_40000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..50ec00ef6be330bbdb4cdf88e9a1097345da0d4d
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/dataloader_40000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84f58ba3b4a1f9be7da4e697e06782f0e1ce4d3aca49f1997087fc83aa466dd9
+size 964
diff --git a/attention_kindselective_n_heads2_seed1341/dataloader_42500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_42500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..74e133f73a7293d5f4d6407784703c91f705d6e3
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/dataloader_42500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf6d24c78d89100d146bce9f26be940db3d71092473d9b55db97d6b35531eac2
+size 964
diff --git a/attention_kindselective_n_heads2_seed1341/dataloader_45000.pt b/attention_kindselective_n_heads2_seed1341/dataloader_45000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..718bc149d695b1e9498bdd0693053d7417207818
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/dataloader_45000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22bb90b43d81f3da5454f91a70e1ed29aeb2f470a727ce38390ff8a5c4924889
+size 964
diff --git a/attention_kindselective_n_heads2_seed1341/dataloader_47500.pt b/attention_kindselective_n_heads2_seed1341/dataloader_47500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..3eab97c7c6d4b86d90405bf1c4f3435727495da4
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/dataloader_47500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55507725e6988f190e4963078652fafa6b68e8d4f79221387612612babf3e1c1
+size 964
diff --git a/attention_kindselective_n_heads2_seed1341/dataloader_49999.pt b/attention_kindselective_n_heads2_seed1341/dataloader_49999.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d87f88b62a343a49411f8a6feee8f527879fcd1f
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/dataloader_49999.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47776cddb8021172f048a950b83f25b692cb340214b800ce3837c15ceb58907c
+size 964
diff --git a/attention_kindselective_n_heads2_seed1341/log2.txt b/attention_kindselective_n_heads2_seed1341/log2.txt
index 65dbc62cbef4deca89167d2bcaa829a84f6d86b3..f526ec0f8bbf8e11898172597eadef1a90cf0000 100644
--- a/attention_kindselective_n_heads2_seed1341/log2.txt
+++ b/attention_kindselective_n_heads2_seed1341/log2.txt
@@ -1,560 +1,4401 @@
-max_steps: 10000
-1500 val loss 7.3014
-1500 val perplexity 1482.3606
-1500 train 7.226598 (lr=9.6149e-05) (hash(x)=49016270)
-1600 val loss 7.2945
-1600 val perplexity 1472.1656
-1600 train 7.086260 (lr=9.5544e-05) (hash(x)=46100488)
+max_steps: 50000
+max_steps: 50000
 0 val loss 11.7712
 0 val perplexity 129466.1094
-1700 val loss 7.2754
-1700 val perplexity 1444.2773
-1700 train 7.331821 (lr=9.4897e-05) (hash(x)=49185350)
-1800 val loss 7.2540
-1800 val perplexity 1413.7415
-1800 train 7.142362 (lr=9.4209e-05) (hash(x)=48024574)
-1900 val loss 7.2813
-1900 val perplexity 1452.8243
-1900 train 7.092393 (lr=9.3481e-05) (hash(x)=45823189)
-0 train 11.770625 (lr=7.5000e-07) (hash(x)=47078120)
-2000 val loss 7.2275
-2000 val perplexity 1376.7460
-2000 train 7.047328 (lr=9.2714e-05) (hash(x)=45703932)
-100 val loss 9.4343
-100 val perplexity 12510.2705
-100 train 9.394177 (lr=7.5750e-05) (hash(x)=43429388)
-2100 val loss 7.1969
-2100 val perplexity 1335.3459
-2100 train 7.812517 (lr=9.1908e-05) (hash(x)=58570170)
-200 val loss 7.8396
-200 val perplexity 2539.0632
-200 train 8.055985 (lr=1.5000e-04) (hash(x)=52929681)
-2200 val loss 7.1753
-2200 val perplexity 1306.7522
-2200 train 7.213409 (lr=9.1064e-05) (hash(x)=55262880)
-300 val loss 7.6860
-300 val perplexity 2177.5486
-300 train 7.698088 (lr=1.4997e-04) (hash(x)=49930367)
-2300 val loss 7.1819
-2300 val perplexity 1315.4077
-2300 train 6.962734 (lr=9.0182e-05) (hash(x)=46415497)
-400 val loss 7.6539
-400 val perplexity 2108.8242
-2400 val loss 7.1252
-2400 val perplexity 1242.9492
-400 train 7.473444 (lr=1.4986e-04) (hash(x)=48542946)
-2400 train 7.056499 (lr=8.9265e-05) (hash(x)=49272278)
-2500 val loss 7.0784
-2500 val perplexity 1186.0667
-2500 train 6.929870 (lr=8.8313e-05) (hash(x)=48390803)
-500 val loss 7.5836
-500 val perplexity 1965.7809
-500 train 7.809551 (lr=1.4969e-04) (hash(x)=55286048)
-2600 val loss 7.0476
-2600 val perplexity 1150.1053
-2600 train 6.941715 (lr=8.7326e-05) (hash(x)=47450116)
-600 val loss 7.5162
-600 val perplexity 1837.5521
-600 train 7.756805 (lr=1.4945e-04) (hash(x)=51577760)
-2700 val loss 7.0309
-2700 val perplexity 1131.0837
-2700 train 7.128078 (lr=8.6306e-05) (hash(x)=52681152)
-700 val loss 7.4236
-700 val perplexity 1675.0077
-700 train 7.572020 (lr=1.4913e-04) (hash(x)=57433471)
-2800 val loss 6.9949
-2800 val perplexity 1091.0746
-2800 train 7.043819 (lr=8.5254e-05) (hash(x)=50664094)
-800 val loss 7.3456
-800 val perplexity 1549.4303
-800 train 7.255013 (lr=1.4876e-04) (hash(x)=49799291)
-2900 val loss 6.9631
-2900 val perplexity 1056.9043
-2900 train 6.776724 (lr=8.4170e-05) (hash(x)=47067144)
-900 val loss 7.2976
-900 val perplexity 1476.7948
-900 train 7.272717 (lr=1.4831e-04) (hash(x)=49502839)
-3000 val loss 6.9730
-3000 val perplexity 1067.3795
-3000 train 6.783483 (lr=8.3057e-05) (hash(x)=45015009)
-1000 val loss 7.2546
-1000 val perplexity 1414.6605
-1000 train 7.555295 (lr=1.4779e-04) (hash(x)=51142904)
-3100 val loss 6.9279
-3100 val perplexity 1020.3347
-3100 train 6.763106 (lr=8.1915e-05) (hash(x)=45245896)
-1100 val loss 7.2519
-1100 val perplexity 1410.7919
-1100 train 7.300293 (lr=1.4721e-04) (hash(x)=52751086)
-3200 val loss 6.9114
-3200 val perplexity 1003.6149
-3200 train 6.848857 (lr=8.0745e-05) (hash(x)=49995942)
-1200 val loss 7.2147
-1200 val perplexity 1359.3206
-1200 train 7.214197 (lr=1.4656e-04) (hash(x)=51538621)
-3300 val loss 6.8949
-3300 val perplexity 987.2419
-3300 train 6.779411 (lr=7.9549e-05) (hash(x)=52311504)
-3400 val loss 6.8761
-3400 val perplexity 968.8881
-3400 train 6.842221 (lr=7.8328e-05) (hash(x)=44332917)
-1300 val loss 7.1409
-1300 val perplexity 1262.5270
-1300 train 7.198671 (lr=1.4585e-04) (hash(x)=52034040)
-3500 val loss 6.8513
-3500 val perplexity 945.1112
-3500 train 6.955316 (lr=7.7082e-05) (hash(x)=56517159)
-1400 val loss 7.1240
-1400 val perplexity 1241.3577
-1400 train 7.189587 (lr=1.4507e-04) (hash(x)=50640105)
-3600 val loss 6.8735
-3600 val perplexity 966.3168
-3600 train 6.768483 (lr=7.5814e-05) (hash(x)=50720920)
-1500 val loss 7.0398
-1500 val perplexity 1141.1345
-1500 train 6.955029 (lr=1.4422e-04) (hash(x)=49016270)
-3700 val loss 6.8349
-3700 val perplexity 929.7078
-3700 train 7.343882 (lr=7.4525e-05) (hash(x)=62727701)
-1600 val loss 6.9730
-1600 val perplexity 1067.4009
-1600 train 6.752039 (lr=1.4332e-04) (hash(x)=46100488)
-3800 val loss 6.8337
-3800 val perplexity 928.6059
-3800 train 6.693852 (lr=7.3215e-05) (hash(x)=54772539)
-3900 val loss 6.8087
-3900 val perplexity 905.6780
-3900 train 6.816524 (lr=7.1887e-05) (hash(x)=52274485)
-1700 val loss 6.9253
-1700 val perplexity 1017.7491
-1700 train 6.972584 (lr=1.4235e-04) (hash(x)=49185350)
-4000 val loss 6.7848
-4000 val perplexity 884.2670
-4000 train 6.490056 (lr=7.0541e-05) (hash(x)=50118307)
-1800 val loss 6.9016
-1800 val perplexity 993.8886
-1800 train 6.798162 (lr=1.4131e-04) (hash(x)=48024574)
-4100 val loss 6.7726
-4100 val perplexity 873.5496
-4100 train 6.365652 (lr=6.9180e-05) (hash(x)=42771647)
-1900 val loss 6.8735
-1900 val perplexity 966.3652
-1900 train 6.644013 (lr=1.4022e-04) (hash(x)=45823189)
-4200 val loss 6.7686
-4200 val perplexity 870.0539
-4200 train 6.847145 (lr=6.7804e-05) (hash(x)=51748836)
-2000 val loss 6.8751
-2000 val perplexity 967.8617
-2000 train 6.675347 (lr=1.3907e-04) (hash(x)=45703932)
-4300 val loss 6.7499
-4300 val perplexity 853.9985
-4300 train 6.751280 (lr=6.6414e-05) (hash(x)=49021280)
-2100 val loss 6.8334
-2100 val perplexity 928.3230
-2100 train 7.513886 (lr=1.3786e-04) (hash(x)=58570170)
-4400 val loss 6.7124
-4400 val perplexity 822.5333
-4400 train 6.766258 (lr=6.5013e-05) (hash(x)=55200309)
-2200 val loss 6.8604
-2200 val perplexity 953.7519
-2200 train 6.891486 (lr=1.3660e-04) (hash(x)=55262880)
-4500 val loss 6.7050
-4500 val perplexity 816.5079
-4500 train 6.752545 (lr=6.3602e-05) (hash(x)=52085049)
-2300 val loss 6.8071
-2300 val perplexity 904.2579
-2300 train 6.533747 (lr=1.3527e-04) (hash(x)=46415497)
-4600 val loss 6.6968
-4600 val perplexity 809.8085
-4600 train 6.673974 (lr=6.2182e-05) (hash(x)=48935595)
-2400 val loss 6.8048
-2400 val perplexity 902.1807
-2400 train 6.739892 (lr=1.3390e-04) (hash(x)=49272278)
-4700 val loss 6.6875
-4700 val perplexity 802.3204
-4700 train 6.886833 (lr=6.0754e-05) (hash(x)=49182380)
-2500 val loss 6.8094
-2500 val perplexity 906.3088
-2500 train 6.669321 (lr=1.3247e-04) (hash(x)=48390803)
-4800 val loss 6.6605
-4800 val perplexity 780.9518
-4800 train 6.491550 (lr=5.9321e-05) (hash(x)=43941929)
-2600 val loss 6.7963
-2600 val perplexity 894.5168
-2600 train 6.670294 (lr=1.3099e-04) (hash(x)=47450116)
-4900 val loss 6.6521
-4900 val perplexity 774.4189
-4900 train 6.824918 (lr=5.7883e-05) (hash(x)=51852773)
-2700 val loss 6.7732
-2700 val perplexity 874.0683
-2700 train 6.866146 (lr=1.2946e-04) (hash(x)=52681152)
-5000 val loss 6.6418
-5000 val perplexity 766.5018
-5000 train 6.268013 (lr=5.6442e-05) (hash(x)=40509616)
-2800 val loss 6.7278
-2800 val perplexity 835.3218
-2800 train 6.812080 (lr=1.2788e-04) (hash(x)=50664094)
-5100 val loss 6.6423
-5100 val perplexity 766.8707
-5100 train 6.882291 (lr=5.5000e-05) (hash(x)=57585369)
-2900 val loss 6.7396
-2900 val perplexity 845.1917
-2900 train 6.555730 (lr=1.2626e-04) (hash(x)=47067144)
-5200 val loss 6.6396
-5200 val perplexity 764.7626
-5200 train 6.626321 (lr=5.3558e-05) (hash(x)=51042313)
-3000 val loss 6.7224
-3000 val perplexity 830.8346
-3000 train 6.528443 (lr=1.2459e-04) (hash(x)=45015009)
-5300 val loss 6.6180
-5300 val perplexity 748.4346
-5300 train 6.767743 (lr=5.2117e-05) (hash(x)=52001684)
-3100 val loss 6.7166
-3100 val perplexity 826.0220
-3100 train 6.519020 (lr=1.2287e-04) (hash(x)=45245896)
-5400 val loss 6.6201
-5400 val perplexity 750.0352
-5400 train 6.522018 (lr=5.0679e-05) (hash(x)=48831647)
-3200 val loss 6.7114
-3200 val perplexity 821.6827
-3200 train 6.646773 (lr=1.2112e-04) (hash(x)=49995942)
-5500 val loss 6.6321
-5500 val perplexity 759.0754
-5500 train 7.037599 (lr=4.9246e-05) (hash(x)=50192069)
-3300 val loss 6.6904
-3300 val perplexity 804.6768
-3300 train 6.555431 (lr=1.1932e-04) (hash(x)=52311504)
-5600 val loss 6.6034
-5600 val perplexity 737.5775
-5600 train 6.437107 (lr=4.7818e-05) (hash(x)=47208852)
-3400 val loss 6.6894
-3400 val perplexity 803.8519
-3400 train 6.636265 (lr=1.1749e-04) (hash(x)=44332917)
-5700 val loss 6.6001
-5700 val perplexity 735.1389
-5700 train 6.192906 (lr=4.6398e-05) (hash(x)=44061694)
-3500 val loss 6.6821
-3500 val perplexity 798.0287
-3500 train 6.763896 (lr=1.1562e-04) (hash(x)=56517159)
-5800 val loss 6.5920
-5800 val perplexity 729.2386
-5800 train 6.948812 (lr=4.4987e-05) (hash(x)=56513279)
-5900 val loss 6.5843
-5900 val perplexity 723.6619
-3600 val loss 6.6827
-3600 val perplexity 798.4920
-5900 train 6.729167 (lr=4.3586e-05) (hash(x)=50412818)
-3600 train 6.536693 (lr=1.1372e-04) (hash(x)=50720920)
-6000 val loss 6.5736
-6000 val perplexity 715.9265
-6000 train 6.400131 (lr=4.2196e-05) (hash(x)=47159634)
-3700 val loss 6.6724
-3700 val perplexity 790.3077
-3700 train 7.203786 (lr=1.1179e-04) (hash(x)=62727701)
-6100 val loss 6.5569
-6100 val perplexity 704.1142
-6100 train 6.635352 (lr=4.0820e-05) (hash(x)=54312795)
-3800 val loss 6.6596
-3800 val perplexity 780.2100
-3800 train 6.530156 (lr=1.0982e-04) (hash(x)=54772539)
-6200 val loss 6.5587
-6200 val perplexity 705.3464
-6200 train 6.674115 (lr=3.9459e-05) (hash(x)=54187587)
-3900 val loss 6.6374
-3900 val perplexity 763.1033
-3900 train 6.637521 (lr=1.0783e-04) (hash(x)=52274485)
-6300 val loss 6.5444
-6300 val perplexity 695.3438
-6300 train 6.631743 (lr=3.8113e-05) (hash(x)=53620387)
-4000 val loss 6.6360
-4000 val perplexity 762.0699
-4000 train 6.343906 (lr=1.0581e-04) (hash(x)=50118307)
-6400 val loss 6.5333
-6400 val perplexity 687.6434
-6400 train 6.477994 (lr=3.6785e-05) (hash(x)=48761774)
-4100 val loss 6.6389
-4100 val perplexity 764.2654
-4100 train 6.220299 (lr=1.0377e-04) (hash(x)=42771647)
-6500 val loss 6.5243
-6500 val perplexity 681.5228
-6500 train 6.738882 (lr=3.5475e-05) (hash(x)=56690281)
-4200 val loss 6.6299
-4200 val perplexity 757.4057
-4200 train 6.721921 (lr=1.0171e-04) (hash(x)=51748836)
-6600 val loss 6.5221
-6600 val perplexity 679.9893
-6600 train 6.334319 (lr=3.4186e-05) (hash(x)=42985269)
-4300 val loss 6.6093
-4300 val perplexity 741.9653
-4300 train 6.632129 (lr=9.9622e-05) (hash(x)=49021280)
-6700 val loss 6.5310
-6700 val perplexity 686.0870
-6700 train 6.588527 (lr=3.2918e-05) (hash(x)=53315447)
-4400 val loss 6.5829
-4400 val perplexity 722.6357
-4400 train 6.619546 (lr=9.7520e-05) (hash(x)=55200309)
-6800 val loss 6.5043
-6800 val perplexity 667.9774
-6800 train 6.952129 (lr=3.1672e-05) (hash(x)=61577166)
-4500 val loss 6.5829
-4500 val perplexity 722.6595
-4500 train 6.642871 (lr=9.5403e-05) (hash(x)=52085049)
-6900 val loss 6.5010
-6900 val perplexity 665.8378
-6900 train 6.649751 (lr=3.0451e-05) (hash(x)=54641005)
-4600 val loss 6.5709
-4600 val perplexity 714.0108
-4600 train 6.567044 (lr=9.3273e-05) (hash(x)=48935595)
-7000 val loss 6.4974
-7000 val perplexity 663.4460
-7000 train 7.009356 (lr=2.9255e-05) (hash(x)=60579512)
-4700 val loss 6.5487
-4700 val perplexity 698.3546
-4700 train 6.793222 (lr=9.1132e-05) (hash(x)=49182380)
-7100 val loss 6.4923
-7100 val perplexity 660.0073
-7100 train 6.391061 (lr=2.8085e-05) (hash(x)=53151549)
-4800 val loss 6.5477
-4800 val perplexity 697.6464
-4800 train 6.381069 (lr=8.8982e-05) (hash(x)=43941929)
-7200 val loss 6.4895
-7200 val perplexity 658.1876
-7200 train 7.420212 (lr=2.6943e-05) (hash(x)=71842455)
-4900 val loss 6.5418
-4900 val perplexity 693.5468
-4900 train 6.709642 (lr=8.6825e-05) (hash(x)=51852773)
-7300 val loss 6.4850
-7300 val perplexity 655.2300
-7300 train 6.246530 (lr=2.5830e-05) (hash(x)=44516452)
-5000 val loss 6.5309
-5000 val perplexity 686.0157
-5000 train 6.152505 (lr=8.4663e-05) (hash(x)=40509616)
-7400 val loss 6.4924
-7400 val perplexity 660.0942
-7400 train 6.132852 (lr=2.4746e-05) (hash(x)=42667710)
-5100 val loss 6.5262
-5100 val perplexity 682.7995
-5100 train 6.765361 (lr=8.2500e-05) (hash(x)=57585369)
-7500 val loss 6.4816
-7500 val perplexity 653.0245
-7500 train 6.229507 (lr=2.3694e-05) (hash(x)=47050797)
-5200 val loss 6.5454
-5200 val perplexity 696.0334
-5200 train 6.526903 (lr=8.0337e-05) (hash(x)=51042313)
-7600 val loss 6.4780
-7600 val perplexity 650.6663
-7600 train 6.379726 (lr=2.2674e-05) (hash(x)=49785056)
-5300 val loss 6.5289
-5300 val perplexity 684.6187
-5300 train 6.668643 (lr=7.8175e-05) (hash(x)=52001684)
-7700 val loss 6.4767
-7700 val perplexity 649.8347
-7700 train 6.284160 (lr=2.1687e-05) (hash(x)=53232030)
-5400 val loss 6.5403
-5400 val perplexity 692.4857
-5400 train 6.453519 (lr=7.6018e-05) (hash(x)=48831647)
-7800 val loss 6.4654
-7800 val perplexity 642.5144
-7800 train 6.285869 (lr=2.0735e-05) (hash(x)=48049749)
-5500 val loss 6.5345
-5500 val perplexity 688.4803
-5500 train 6.949864 (lr=7.3868e-05) (hash(x)=50192069)
-7900 val loss 6.4628
-7900 val perplexity 640.8224
-7900 train 6.296690 (lr=1.9818e-05) (hash(x)=44768513)
-5600 val loss 6.5215
-5600 val perplexity 679.6085
-5600 train 6.368748 (lr=7.1727e-05) (hash(x)=47208852)
-8000 val loss 6.4540
-8000 val perplexity 635.2330
-8000 train 6.295214 (lr=1.8936e-05) (hash(x)=46228039)
-5700 val loss 6.5209
-5700 val perplexity 679.1575
-5700 train 6.097320 (lr=6.9597e-05) (hash(x)=44061694)
-8100 val loss 6.4509
-8100 val perplexity 633.2829
-8100 train 6.780039 (lr=1.8092e-05) (hash(x)=60017091)
-5800 val loss 6.5331
-5800 val perplexity 687.5231
-5800 train 6.887940 (lr=6.7480e-05) (hash(x)=56513279)
-8200 val loss 6.4417
-8200 val perplexity 627.4558
-8200 train 6.451499 (lr=1.7286e-05) (hash(x)=49910198)
-5900 val loss 6.5202
-5900 val perplexity 678.6848
-8300 val loss 6.4402
-8300 val perplexity 626.5251
-5900 train 6.650718 (lr=6.5378e-05) (hash(x)=50412818)
-8300 train 6.745976 (lr=1.6519e-05) (hash(x)=57919055)
-8400 val loss 6.4405
-8400 val perplexity 626.7328
-8400 train 6.515765 (lr=1.5791e-05) (hash(x)=49694964)
-6000 val loss 6.5113
-6000 val perplexity 672.6966
-6000 train 6.349458 (lr=6.3294e-05) (hash(x)=47159634)
-8500 val loss 6.4337
-8500 val perplexity 622.4640
-8500 train 6.469025 (lr=1.5103e-05) (hash(x)=53762585)
-6100 val loss 6.4893
-6100 val perplexity 658.0756
-6100 train 6.571654 (lr=6.1230e-05) (hash(x)=54312795)
-8600 val loss 6.4334
-8600 val perplexity 622.3082
-8600 train 6.465587 (lr=1.4456e-05) (hash(x)=51166973)
-6200 val loss 6.4849
-6200 val perplexity 655.2000
-6200 train 6.599206 (lr=5.9188e-05) (hash(x)=54187587)
-8700 val loss 6.4311
-8700 val perplexity 620.8644
-8700 train 6.505373 (lr=1.3851e-05) (hash(x)=53968049)
-6300 val loss 6.4821
-6300 val perplexity 653.3412
-6300 train 6.581801 (lr=5.7169e-05) (hash(x)=53620387)
-8800 val loss 6.4265
-8800 val perplexity 617.9882
-8800 train 6.509236 (lr=1.3289e-05) (hash(x)=59231056)
-6400 val loss 6.4599
-6400 val perplexity 638.9672
-6400 train 6.395986 (lr=5.5177e-05) (hash(x)=48761774)
-8900 val loss 6.4245
-8900 val perplexity 616.7817
-8900 train 6.301777 (lr=1.2769e-05) (hash(x)=50488048)
-6500 val loss 6.4580
-6500 val perplexity 637.7895
-6500 train 6.647865 (lr=5.3213e-05) (hash(x)=56690281)
-9000 val loss 6.4237
-9000 val perplexity 616.2996
-9000 train 6.162522 (lr=1.2292e-05) (hash(x)=44492956)
-6600 val loss 6.4526
-6600 val perplexity 634.3764
-6600 train 6.277388 (lr=5.1279e-05) (hash(x)=42985269)
-9100 val loss 6.4321
-9100 val perplexity 621.4722
-9100 train 6.460136 (lr=1.1860e-05) (hash(x)=51134989)
-6700 val loss 6.4537
-6700 val perplexity 635.0168
-6700 train 6.495444 (lr=4.9377e-05) (hash(x)=53315447)
-9200 val loss 6.4255
-9200 val perplexity 617.3776
-9200 train 6.232402 (lr=1.1472e-05) (hash(x)=48636056)
-6800 val loss 6.4483
-6800 val perplexity 631.6550
-6800 train 6.872110 (lr=4.7509e-05) (hash(x)=61577166)
-9300 val loss 6.4248
-9300 val perplexity 616.9818
-9300 train 6.371828 (lr=1.1128e-05) (hash(x)=50200551)
-6900 val loss 6.4408
-6900 val perplexity 626.9297
-6900 train 6.579897 (lr=4.5676e-05) (hash(x)=54641005)
-9400 val loss 6.4224
-9400 val perplexity 615.4691
-9400 train 6.256755 (lr=1.0830e-05) (hash(x)=48057228)
-7000 val loss 6.4378
-7000 val perplexity 625.0102
-7000 train 6.968445 (lr=4.3882e-05) (hash(x)=60579512)
-9500 val loss 6.4180
-9500 val perplexity 612.8030
-9500 train 6.189470 (lr=1.0577e-05) (hash(x)=48125171)
-7100 val loss 6.4326
-7100 val perplexity 621.8027
-7100 train 6.320305 (lr=4.2128e-05) (hash(x)=53151549)
-9600 val loss 6.4210
-9600 val perplexity 614.6295
-9600 train 6.340978 (lr=1.0369e-05) (hash(x)=53375853)
-7200 val loss 6.4300
-7200 val perplexity 620.1948
-7200 train 7.380653 (lr=4.0414e-05) (hash(x)=71842455)
-9700 val loss 6.4135
-9700 val perplexity 610.0555
-9700 train 7.223229 (lr=1.0208e-05) (hash(x)=53924631)
-7300 val loss 6.4297
-7300 val perplexity 620.0109
-7300 train 6.191623 (lr=3.8745e-05) (hash(x)=44516452)
-9800 val loss 6.4102
-9800 val perplexity 608.0391
-9800 train 6.454649 (lr=1.0092e-05) (hash(x)=48895047)
-7400 val loss 6.4289
-7400 val perplexity 619.4763
-7400 train 6.074793 (lr=3.7120e-05) (hash(x)=42667710)
-9900 val loss 6.4071
-9900 val perplexity 606.1632
-9900 train 6.266374 (lr=1.0023e-05) (hash(x)=44269923)
-7500 val loss 6.4316
-7500 val perplexity 621.1490
-7500 train 6.169018 (lr=3.5541e-05) (hash(x)=47050797)
-9999 val loss 6.4119
-9999 val perplexity 609.0312
-7600 val loss 6.4237
-7600 val perplexity 616.2653
-7600 train 6.329450 (lr=3.4011e-05) (hash(x)=49785056)
-7700 val loss 6.4194
-7700 val perplexity 613.6569
-7700 train 6.231349 (lr=3.2531e-05) (hash(x)=53232030)
-7800 val loss 6.4186
-7800 val perplexity 613.1287
-7800 train 6.243061 (lr=3.1102e-05) (hash(x)=48049749)
-7900 val loss 6.4043
-7900 val perplexity 604.4531
-7900 train 6.246456 (lr=2.9726e-05) (hash(x)=44768513)
-8000 val loss 6.4115
-8000 val perplexity 608.7827
-8000 train 6.265230 (lr=2.8405e-05) (hash(x)=46228039)
-8100 val loss 6.4030
-8100 val perplexity 603.6253
-8100 train 6.710750 (lr=2.7138e-05) (hash(x)=60017091)
-8200 val loss 6.3952
-8200 val perplexity 598.9441
-8200 train 6.407417 (lr=2.5929e-05) (hash(x)=49910198)
-8300 val loss 6.3936
-8300 val perplexity 598.0146
-8300 train 6.694312 (lr=2.4778e-05) (hash(x)=57919055)
-8400 val loss 6.3967
-8400 val perplexity 599.8673
-8400 train 6.466394 (lr=2.3686e-05) (hash(x)=49694964)
-8500 val loss 6.3893
-8500 val perplexity 595.4279
-8500 train 6.434396 (lr=2.2655e-05) (hash(x)=53762585)
-8600 val loss 6.3880
-8600 val perplexity 594.6692
-8600 train 6.429276 (lr=2.1685e-05) (hash(x)=51166973)
-8700 val loss 6.3826
-8700 val perplexity 591.4665
-8700 train 6.472390 (lr=2.0777e-05) (hash(x)=53968049)
-8800 val loss 6.3809
-8800 val perplexity 590.4394
-8800 train 6.468944 (lr=1.9933e-05) (hash(x)=59231056)
-8900 val loss 6.3768
-8900 val perplexity 588.0292
-8900 train 6.261299 (lr=1.9153e-05) (hash(x)=50488048)
-9000 val loss 6.3776
-9000 val perplexity 588.4850
-9000 train 6.119425 (lr=1.8439e-05) (hash(x)=44492956)
-9100 val loss 6.3821
-9100 val perplexity 591.1530
-9100 train 6.412546 (lr=1.7790e-05) (hash(x)=51134989)
-9200 val loss 6.3779
-9200 val perplexity 588.6931
-9200 train 6.194124 (lr=1.7208e-05) (hash(x)=48636056)
-9300 val loss 6.3750
-9300 val perplexity 587.0090
-9300 train 6.325751 (lr=1.6692e-05) (hash(x)=50200551)
-9400 val loss 6.3751
-9400 val perplexity 587.0705
-9400 train 6.198377 (lr=1.6245e-05) (hash(x)=48057228)
-9500 val loss 6.3753
-9500 val perplexity 587.1808
-9500 train 6.169206 (lr=1.5865e-05) (hash(x)=48125171)
-9600 val loss 6.3750
-9600 val perplexity 587.0090
-9600 train 6.290072 (lr=1.5554e-05) (hash(x)=53375853)
-9700 val loss 6.3654
-9700 val perplexity 581.3495
-9700 train 7.198069 (lr=1.5312e-05) (hash(x)=53924631)
-9800 val loss 6.3617
-9800 val perplexity 579.2410
-9800 train 6.411182 (lr=1.5139e-05) (hash(x)=48895047)
-9900 val loss 6.3613
-9900 val perplexity 579.0135
-9900 train 6.211933 (lr=1.5035e-05) (hash(x)=44269923)
-9999 val loss 6.3711
-9999 val perplexity 584.7159
+0 val loss 11.7712
+0 val perplexity 129466.1094
+0 val loss 11.7712
+0 val perplexity 129466.1094
+0 train 11.770625 (lr=2.5000e-07) (hash(x)=47078120)
+0 train 11.770625 (lr=3.5000e-07) (hash(x)=47078120)
+0 train 11.770625 (lr=5.0000e-07) (hash(x)=47078120)
+100 val loss 10.2033
+100 val perplexity 26992.5645
+100 train 10.206360 (lr=2.5250e-05) (hash(x)=43429388)
+100 val loss 10.0565
+100 val perplexity 23306.7559
+100 train 10.048087 (lr=3.5350e-05) (hash(x)=43429388)
+100 val loss 9.8951
+100 val perplexity 19833.2090
+100 train 9.880125 (lr=5.0500e-05) (hash(x)=43429388)
+200 val loss 8.5846
+200 val perplexity 5348.3901
+200 train 8.746173 (lr=5.0000e-05) (hash(x)=52929681)
+200 val loss 8.3626
+200 val perplexity 4284.0156
+200 train 8.533326 (lr=7.0000e-05) (hash(x)=52929681)
+200 val loss 8.3264
+200 val perplexity 4131.3823
+200 train 8.494742 (lr=1.0000e-04) (hash(x)=52929681)
+300 val loss 7.8316
+300 val perplexity 2518.9448
+300 train 7.851068 (lr=5.0000e-05) (hash(x)=49930367)
+300 val loss 7.8070
+300 val perplexity 2457.8367
+300 train 7.819790 (lr=6.9999e-05) (hash(x)=49930367)
+300 val loss 7.8235
+300 val perplexity 2498.6895
+300 train 7.841080 (lr=9.9999e-05) (hash(x)=49930367)
+400 val loss 7.6793
+400 val perplexity 2163.0315
+400 train 7.493921 (lr=4.9998e-05) (hash(x)=48542946)
+400 val loss 7.6705
+400 val perplexity 2144.2366
+400 train 7.483693 (lr=6.9997e-05) (hash(x)=48542946)
+400 val loss 7.6975
+400 val perplexity 2202.8335
+400 train 7.526471 (lr=9.9996e-05) (hash(x)=48542946)
+500 val loss 7.6384
+500 val perplexity 2076.3428
+500 train 7.835117 (lr=4.9996e-05) (hash(x)=55286048)
+500 val loss 7.6544
+500 val perplexity 2109.8975
+500 train 7.862077 (lr=6.9994e-05) (hash(x)=55286048)
+500 val loss 7.6798
+500 val perplexity 2164.1704
+500 train 7.879671 (lr=9.9992e-05) (hash(x)=55286048)
+600 val loss 7.6124
+600 val perplexity 2023.1598
+600 train 7.864288 (lr=4.9993e-05) (hash(x)=51577760)
+600 val loss 7.6192
+600 val perplexity 2036.9033
+600 train 7.893260 (lr=6.9990e-05) (hash(x)=51577760)
+600 val loss 7.6376
+600 val perplexity 2074.8047
+600 train 7.888408 (lr=9.9986e-05) (hash(x)=51577760)
+700 val loss 7.6056
+700 val perplexity 2009.3633
+700 train 7.753090 (lr=4.9989e-05) (hash(x)=57433471)
+700 val loss 7.6017
+700 val perplexity 2001.4990
+700 train 7.751354 (lr=6.9984e-05) (hash(x)=57433471)
+700 val loss 7.6035
+700 val perplexity 2005.2074
+700 train 7.749254 (lr=9.9978e-05) (hash(x)=57433471)
+800 val loss 7.6036
+800 val perplexity 2005.4731
+800 train 7.522329 (lr=4.9984e-05) (hash(x)=49799291)
+800 val loss 7.6126
+800 val perplexity 2023.5988
+800 train 7.552643 (lr=6.9977e-05) (hash(x)=49799291)
+800 val loss 7.5791
+800 val perplexity 1956.7739
+800 train 7.513973 (lr=9.9968e-05) (hash(x)=49799291)
+900 val loss 7.6089
+900 val perplexity 2016.0988
+900 train 7.618979 (lr=4.9978e-05) (hash(x)=49502839)
+900 val loss 7.6029
+900 val perplexity 2004.0411
+900 train 7.605560 (lr=6.9969e-05) (hash(x)=49502839)
+900 val loss 7.5441
+900 val perplexity 1889.4799
+900 train 7.559354 (lr=9.9956e-05) (hash(x)=49502839)
+1000 val loss 7.5819
+1000 val perplexity 1962.3719
+1000 train 7.842342 (lr=4.9971e-05) (hash(x)=51142904)
+1000 val loss 7.6253
+1000 val perplexity 2049.4331
+1000 train 7.876435 (lr=6.9960e-05) (hash(x)=51142904)
+1100 val loss 7.5698
+1100 val perplexity 1938.7089
+1100 train 7.650688 (lr=4.9964e-05) (hash(x)=52751086)
+1000 val loss 7.5053
+1000 val perplexity 1817.6809
+1000 train 7.764091 (lr=9.9943e-05) (hash(x)=51142904)
+1100 val loss 7.5860
+1100 val perplexity 1970.4799
+1100 train 7.674089 (lr=6.9949e-05) (hash(x)=52751086)
+1200 val loss 7.5671
+1200 val perplexity 1933.4513
+1200 train 7.568090 (lr=4.9955e-05) (hash(x)=51538621)
+1200 val loss 7.5782
+1200 val perplexity 1955.1324
+1200 train 7.595935 (lr=6.9937e-05) (hash(x)=51538621)
+1100 val loss 7.4754
+1100 val perplexity 1764.1299
+1100 train 7.548927 (lr=9.9927e-05) (hash(x)=52751086)
+1300 val loss 7.5720
+1300 val perplexity 1942.9651
+1300 train 7.642082 (lr=4.9946e-05) (hash(x)=52034040)
+1300 val loss 7.5800
+1300 val perplexity 1958.6017
+1300 train 7.652497 (lr=6.9924e-05) (hash(x)=52034040)
+1200 val loss 7.4633
+1200 val perplexity 1742.9686
+1200 train 7.467877 (lr=9.9910e-05) (hash(x)=51538621)
+1400 val loss 7.5476
+1400 val perplexity 1896.1561
+1400 train 7.586929 (lr=4.9936e-05) (hash(x)=50640105)
+1400 val loss 7.5772
+1400 val perplexity 1953.2203
+1400 train 7.614067 (lr=6.9910e-05) (hash(x)=50640105)
+1300 val loss 7.4409
+1300 val perplexity 1704.1986
+1300 train 7.514560 (lr=9.9892e-05) (hash(x)=52034040)
+1500 val loss 7.5311
+1500 val perplexity 1865.0898
+1500 train 7.483514 (lr=4.9924e-05) (hash(x)=49016270)
+1500 val loss 7.5687
+1500 val perplexity 1936.6226
+1500 train 7.509407 (lr=6.9894e-05) (hash(x)=49016270)
+1400 val loss 7.4407
+1400 val perplexity 1704.0084
+1400 train 7.479903 (lr=9.9871e-05) (hash(x)=50640105)
+1600 val loss 7.5091
+1600 val perplexity 1824.6228
+1600 train 7.283084 (lr=4.9912e-05) (hash(x)=46100488)
+1600 val loss 7.5822
+1600 val perplexity 1962.8669
+1600 train 7.371550 (lr=6.9877e-05) (hash(x)=46100488)
+1500 val loss 7.4013
+1500 val perplexity 1638.1898
+1500 train 7.332983 (lr=9.9849e-05) (hash(x)=49016270)
+1700 val loss 7.4860
+1700 val perplexity 1782.9028
+1700 train 7.526432 (lr=4.9899e-05) (hash(x)=49185350)
+1700 val loss 7.5687
+1700 val perplexity 1936.6531
+1700 train 7.600106 (lr=6.9859e-05) (hash(x)=49185350)
+1600 val loss 7.3796
+1600 val perplexity 1602.8821
+1600 train 7.147182 (lr=9.9825e-05) (hash(x)=46100488)
+1800 val loss 7.4800
+1800 val perplexity 1772.1639
+1800 train 7.367154 (lr=4.9885e-05) (hash(x)=48024574)
+1800 val loss 7.5553
+1800 val perplexity 1910.8547
+1800 train 7.461845 (lr=6.9840e-05) (hash(x)=48024574)
+1700 val loss 7.3547
+1700 val perplexity 1563.4885
+1700 train 7.391090 (lr=9.9799e-05) (hash(x)=49185350)
+1900 val loss 7.4720
+1900 val perplexity 1758.0442
+1900 train 7.292343 (lr=4.9871e-05) (hash(x)=45823189)
+1900 val loss 7.5419
+1900 val perplexity 1885.3346
+1900 train 7.362976 (lr=6.9819e-05) (hash(x)=45823189)
+1800 val loss 7.3457
+1800 val perplexity 1549.4635
+1800 train 7.229240 (lr=9.9771e-05) (hash(x)=48024574)
+2000 val loss 7.4559
+2000 val perplexity 1730.0703
+2000 train 7.296420 (lr=4.9855e-05) (hash(x)=45703932)
+2000 val loss 7.5337
+2000 val perplexity 1869.9216
+2000 train 7.385001 (lr=6.9797e-05) (hash(x)=45703932)
+2100 val loss 7.4495
+2100 val perplexity 1719.0724
+2100 train 8.025798 (lr=4.9839e-05) (hash(x)=58570170)
+1900 val loss 7.3341
+1900 val perplexity 1531.6666
+2100 val loss 7.4990
+2100 val perplexity 1806.2311
+1900 train 7.152208 (lr=9.9741e-05) (hash(x)=45823189)
+2100 train 8.082744 (lr=6.9774e-05) (hash(x)=58570170)
+2200 val loss 7.4700
+2200 val perplexity 1754.6849
+2200 train 7.494737 (lr=4.9821e-05) (hash(x)=55262880)
+2200 val loss 7.4779
+2200 val perplexity 1768.5037
+2200 train 7.508498 (lr=6.9750e-05) (hash(x)=55262880)
+2000 val loss 7.3298
+2000 val perplexity 1525.1315
+2000 train 7.174179 (lr=9.9710e-05) (hash(x)=45703932)
+2300 val loss 7.4234
+2300 val perplexity 1674.7913
+2300 train 7.192656 (lr=4.9803e-05) (hash(x)=46415497)
+2300 val loss 7.4373
+2300 val perplexity 1698.2014
+2300 train 7.212271 (lr=6.9724e-05) (hash(x)=46415497)
+2100 val loss 7.3045
+2100 val perplexity 1486.9984
+2100 train 7.906702 (lr=9.9677e-05) (hash(x)=58570170)
+2400 val loss 7.4013
+2400 val perplexity 1638.1156
+2400 train 7.342383 (lr=4.9784e-05) (hash(x)=49272278)
+2400 val loss 7.3972
+2400 val perplexity 1631.4749
+2400 train 7.336201 (lr=6.9697e-05) (hash(x)=49272278)
+2200 val loss 7.3089
+2200 val perplexity 1493.5055
+2200 train 7.352629 (lr=9.9642e-05) (hash(x)=55262880)
+2500 val loss 7.3858
+2500 val perplexity 1612.8698
+2500 train 7.269034 (lr=4.9764e-05) (hash(x)=48390803)
+2500 val loss 7.3657
+2500 val perplexity 1580.8145
+2500 train 7.247119 (lr=6.9669e-05) (hash(x)=48390803)
+2300 val loss 7.3411
+2300 val perplexity 1542.3964
+2300 train 7.163018 (lr=9.9606e-05) (hash(x)=46415497)
+2600 val loss 7.3594
+2600 val perplexity 1570.8942
+2600 train 7.243450 (lr=4.9743e-05) (hash(x)=47450116)
+2600 val loss 7.3575
+2600 val perplexity 1567.9031
+2600 train 7.243428 (lr=6.9640e-05) (hash(x)=47450116)
+2400 val loss 7.3191
+2400 val perplexity 1508.7885
+2400 train 7.264038 (lr=9.9567e-05) (hash(x)=49272278)
+2700 val loss 7.3460
+2700 val perplexity 1549.9852
+2700 train 7.436283 (lr=4.9721e-05) (hash(x)=52681152)
+2700 val loss 7.3453
+2700 val perplexity 1548.8527
+2700 train 7.419794 (lr=6.9609e-05) (hash(x)=52681152)
+2500 val loss 7.2346
+2500 val perplexity 1386.6477
+2500 train 7.094385 (lr=9.9527e-05) (hash(x)=48390803)
+2800 val loss 7.3179
+2800 val perplexity 1507.0958
+2800 train 7.384782 (lr=4.9698e-05) (hash(x)=50664094)
+2800 val loss 7.3382
+2800 val perplexity 1537.8748
+2800 train 7.414604 (lr=6.9577e-05) (hash(x)=50664094)
+2900 val loss 7.3156
+2900 val perplexity 1503.5679
+2900 train 7.132506 (lr=4.9674e-05) (hash(x)=47067144)
+2600 val loss 7.2024
+2600 val perplexity 1342.6188
+2600 train 7.080667 (lr=9.9485e-05) (hash(x)=47450116)
+2900 val loss 7.3248
+2900 val perplexity 1517.5225
+2900 train 7.166916 (lr=6.9544e-05) (hash(x)=47067144)
+3000 val loss 7.2966
+3000 val perplexity 1475.3104
+3000 train 7.111231 (lr=4.9650e-05) (hash(x)=45015009)
+2700 val loss 7.1974
+2700 val perplexity 1335.9701
+2700 train 7.291355 (lr=9.9442e-05) (hash(x)=52681152)
+3000 val loss 7.3050
+3000 val perplexity 1487.7729
+3000 train 7.111160 (lr=6.9510e-05) (hash(x)=45015009)
+3100 val loss 7.2757
+3100 val perplexity 1444.8168
+3100 train 7.088307 (lr=4.9625e-05) (hash(x)=45245896)
+3100 val loss 7.2994
+3100 val perplexity 1479.3949
+3100 train 7.128348 (lr=6.9474e-05) (hash(x)=45245896)
+2800 val loss 7.1475
+2800 val perplexity 1270.9235
+2800 train 7.205905 (lr=9.9396e-05) (hash(x)=50664094)
+3200 val loss 7.2659
+3200 val perplexity 1430.6354
+3200 train 7.239969 (lr=4.9598e-05) (hash(x)=49995942)
+3200 val loss 7.2771
+3200 val perplexity 1446.7656
+3200 train 7.227925 (lr=6.9438e-05) (hash(x)=49995942)
+2900 val loss 7.1312
+2900 val perplexity 1250.3710
+2900 train 6.949361 (lr=9.9349e-05) (hash(x)=47067144)
+3300 val loss 7.2506
+3300 val perplexity 1408.8812
+3300 train 7.149277 (lr=4.9571e-05) (hash(x)=52311504)
+3300 val loss 7.2734
+3300 val perplexity 1441.4346
+3300 train 7.174790 (lr=6.9400e-05) (hash(x)=52311504)
+3000 val loss 7.1105
+3000 val perplexity 1224.7368
+3000 train 6.942707 (lr=9.9300e-05) (hash(x)=45015009)
+3400 val loss 7.2187
+3400 val perplexity 1364.7137
+3400 train 7.215257 (lr=4.9543e-05) (hash(x)=44332917)
+3400 val loss 7.2563
+3400 val perplexity 1417.0701
+3400 train 7.229104 (lr=6.9360e-05) (hash(x)=44332917)
+3100 val loss 7.0858
+3100 val perplexity 1194.9200
+3100 train 6.913786 (lr=9.9249e-05) (hash(x)=45245896)
+3500 val loss 7.1893
+3500 val perplexity 1325.1709
+3500 train 7.291782 (lr=4.9514e-05) (hash(x)=56517159)
+3500 val loss 7.2113
+3500 val perplexity 1354.6410
+3500 train 7.321904 (lr=6.9320e-05) (hash(x)=56517159)
+3200 val loss 7.0656
+3200 val perplexity 1170.9994
+3200 train 7.020722 (lr=9.9197e-05) (hash(x)=49995942)
+3600 val loss 7.1847
+3600 val perplexity 1319.0482
+3600 train 7.071852 (lr=4.9484e-05) (hash(x)=50720920)
+3600 val loss 7.2092
+3600 val perplexity 1351.8141
+3600 train 7.138223 (lr=6.9278e-05) (hash(x)=50720920)
+3300 val loss 7.0669
+3300 val perplexity 1172.4705
+3300 train 6.962692 (lr=9.9142e-05) (hash(x)=52311504)
+3700 val loss 7.1715
+3700 val perplexity 1301.7917
+3700 train 7.680056 (lr=4.9454e-05) (hash(x)=62727701)
+3700 val loss 7.1758
+3700 val perplexity 1307.4110
+3700 train 7.640621 (lr=6.9235e-05) (hash(x)=62727701)
+3400 val loss 7.0259
+3400 val perplexity 1125.4165
+3400 train 7.003755 (lr=9.9086e-05) (hash(x)=44332917)
+3800 val loss 7.1476
+3800 val perplexity 1271.0247
+3800 train 6.989200 (lr=4.9422e-05) (hash(x)=54772539)
+3800 val loss 7.1592
+3800 val perplexity 1285.8239
+3800 train 7.013257 (lr=6.9191e-05) (hash(x)=54772539)
+3500 val loss 7.0122
+3500 val perplexity 1110.0574
+3500 train 7.132928 (lr=9.9028e-05) (hash(x)=56517159)
+3900 val loss 7.1487
+3900 val perplexity 1272.4905
+3900 train 7.162163 (lr=4.9390e-05) (hash(x)=52274485)
+3900 val loss 7.1598
+3900 val perplexity 1286.6543
+3900 train 7.188324 (lr=6.9146e-05) (hash(x)=52274485)
+4000 val loss 7.1395
+4000 val perplexity 1260.7595
+4000 train 6.859540 (lr=4.9357e-05) (hash(x)=50118307)
+3600 val loss 6.9864
+3600 val perplexity 1081.8583
+3600 train 6.872036 (lr=9.8969e-05) (hash(x)=50720920)
+4000 val loss 7.1493
+4000 val perplexity 1273.2048
+4000 train 6.882396 (lr=6.9099e-05) (hash(x)=50118307)
+4100 val loss 7.1080
+4100 val perplexity 1221.7352
+4100 train 6.696227 (lr=4.9322e-05) (hash(x)=42771647)
+3700 val loss 6.9570
+3700 val perplexity 1050.4807
+3700 train 7.451654 (lr=9.8908e-05) (hash(x)=62727701)
+4100 val loss 7.1660
+4100 val perplexity 1294.6802
+4100 train 6.767435 (lr=6.9051e-05) (hash(x)=42771647)
+4200 val loss 7.0914
+4200 val perplexity 1201.6149
+4200 train 7.170551 (lr=4.9287e-05) (hash(x)=51748836)
+3800 val loss 6.9337
+3800 val perplexity 1026.3269
+3800 train 6.792619 (lr=9.8845e-05) (hash(x)=54772539)
+4200 val loss 7.1143
+4200 val perplexity 1229.4664
+4200 train 7.180784 (lr=6.9002e-05) (hash(x)=51748836)
+4300 val loss 7.0601
+4300 val perplexity 1164.5834
+4300 train 7.057050 (lr=4.9252e-05) (hash(x)=49021280)
+3900 val loss 6.9030
+3900 val perplexity 995.2720
+3900 train 6.928247 (lr=9.8780e-05) (hash(x)=52274485)
+4300 val loss 7.1378
+4300 val perplexity 1258.7052
+4300 train 7.148933 (lr=6.8952e-05) (hash(x)=49021280)
+4400 val loss 7.0244
+4400 val perplexity 1123.7247
+4400 train 7.103628 (lr=4.9215e-05) (hash(x)=55200309)
+4000 val loss 6.8878
+4000 val perplexity 980.2460
+4000 train 6.603559 (lr=9.8713e-05) (hash(x)=50118307)
+4400 val loss 7.0874
+4400 val perplexity 1196.8223
+4400 train 7.168321 (lr=6.8901e-05) (hash(x)=55200309)
+4500 val loss 7.0041
+4500 val perplexity 1101.1179
+4500 train 7.044929 (lr=4.9177e-05) (hash(x)=52085049)
+4500 val loss 7.0589
+4500 val perplexity 1163.2164
+4500 train 7.096609 (lr=6.8848e-05) (hash(x)=52085049)
+4100 val loss 6.8764
+4100 val perplexity 969.1635
+4100 train 6.470846 (lr=9.8645e-05) (hash(x)=42771647)
+4600 val loss 7.0027
+4600 val perplexity 1099.6326
+4600 train 6.979061 (lr=4.9139e-05) (hash(x)=48935595)
+4600 val loss 7.0364
+4600 val perplexity 1137.3259
+4600 train 6.995081 (lr=6.8794e-05) (hash(x)=48935595)
+4200 val loss 6.8608
+4200 val perplexity 954.1253
+4200 train 6.947160 (lr=9.8575e-05) (hash(x)=51748836)
+4700 val loss 6.9578
+4700 val perplexity 1051.2834
+4700 train 7.129683 (lr=4.9099e-05) (hash(x)=49182380)
+4700 val loss 7.0110
+4700 val perplexity 1108.7338
+4700 train 7.189000 (lr=6.8739e-05) (hash(x)=49182380)
+4300 val loss 6.8324
+4300 val perplexity 927.4540
+4300 train 6.841238 (lr=9.8503e-05) (hash(x)=49021280)
+4800 val loss 6.9223
+4800 val perplexity 1014.6456
+4800 train 6.725775 (lr=4.9059e-05) (hash(x)=43941929)
+4800 val loss 7.0088
+4800 val perplexity 1106.3794
+4800 train 6.818189 (lr=6.8683e-05) (hash(x)=43941929)
+4400 val loss 6.8061
+4400 val perplexity 903.3645
+4400 train 6.879786 (lr=9.8430e-05) (hash(x)=55200309)
+4900 val loss 6.8891
+4900 val perplexity 981.5084
+4900 train 7.054278 (lr=4.9018e-05) (hash(x)=51852773)
+4900 val loss 6.9903
+4900 val perplexity 1086.0919
+4900 train 7.134183 (lr=6.8626e-05) (hash(x)=51852773)
+4500 val loss 6.8024
+4500 val perplexity 899.9898
+4500 train 6.836329 (lr=9.8355e-05) (hash(x)=52085049)
+5000 val loss 6.8573
+5000 val perplexity 950.8198
+5000 train 6.482135 (lr=4.8976e-05) (hash(x)=40509616)
+5000 val loss 6.9693
+5000 val perplexity 1063.5103
+5000 train 6.612258 (lr=6.8567e-05) (hash(x)=40509616)
+5100 val loss 6.8343
+5100 val perplexity 929.1343
+5100 train 7.126708 (lr=4.8934e-05) (hash(x)=57585369)
+4600 val loss 6.7937
+4600 val perplexity 892.2471
+4600 train 6.779639 (lr=9.8278e-05) (hash(x)=48935595)
+5100 val loss 6.9601
+5100 val perplexity 1053.7472
+5100 train 7.232564 (lr=6.8507e-05) (hash(x)=57585369)
+5200 val loss 6.8134
+5200 val perplexity 910.0000
+5200 train 6.828725 (lr=4.8890e-05) (hash(x)=51042313)
+4700 val loss 6.7569
+4700 val perplexity 859.9806
+4700 train 6.949156 (lr=9.8199e-05) (hash(x)=49182380)
+5200 val loss 6.9511
+5200 val perplexity 1044.2822
+5200 train 6.954596 (lr=6.8446e-05) (hash(x)=51042313)
+5300 val loss 6.7903
+5300 val perplexity 889.1386
+5300 train 6.950994 (lr=4.8846e-05) (hash(x)=52001684)
+4800 val loss 6.7365
+4800 val perplexity 842.6316
+4800 train 6.560075 (lr=9.8119e-05) (hash(x)=43941929)
+5300 val loss 6.9428
+5300 val perplexity 1035.6433
+5300 train 7.068288 (lr=6.8384e-05) (hash(x)=52001684)
+5400 val loss 6.7679
+5400 val perplexity 869.5019
+5400 train 6.695953 (lr=4.8800e-05) (hash(x)=48831647)
+4900 val loss 6.7258
+4900 val perplexity 833.6343
+4900 train 6.891153 (lr=9.8036e-05) (hash(x)=51852773)
+5400 val loss 6.9532
+5400 val perplexity 1046.4855
+5400 train 6.901043 (lr=6.8320e-05) (hash(x)=48831647)
+5500 val loss 6.7634
+5500 val perplexity 865.5666
+5500 train 7.144577 (lr=4.8754e-05) (hash(x)=50192069)
+5000 val loss 6.7134
+5000 val perplexity 823.3696
+5500 val loss 6.9024
+5500 val perplexity 994.6591
+5500 train 7.260624 (lr=6.8256e-05) (hash(x)=50192069)
+5000 train 6.339698 (lr=9.7953e-05) (hash(x)=40509616)
+5600 val loss 6.7351
+5600 val perplexity 841.4648
+5600 train 6.562131 (lr=4.8707e-05) (hash(x)=47208852)
+5600 val loss 6.9064
+5600 val perplexity 998.6816
+5600 train 6.739335 (lr=6.8190e-05) (hash(x)=47208852)
+5100 val loss 6.6933
+5100 val perplexity 807.0188
+5100 train 6.913021 (lr=9.7867e-05) (hash(x)=57585369)
+5700 val loss 6.7383
+5700 val perplexity 844.1457
+5700 train 6.331465 (lr=4.8659e-05) (hash(x)=44061694)
+5700 val loss 6.8881
+5700 val perplexity 980.5087
+5700 train 6.497312 (lr=6.8123e-05) (hash(x)=44061694)
+5200 val loss 6.6851
+5200 val perplexity 800.4271
+5200 train 6.668004 (lr=9.7780e-05) (hash(x)=51042313)
+5800 val loss 6.7127
+5800 val perplexity 822.7604
+5800 train 7.059242 (lr=4.8611e-05) (hash(x)=56513279)
+5800 val loss 6.8917
+5800 val perplexity 984.1093
+5800 train 7.228807 (lr=6.8055e-05) (hash(x)=56513279)
+5300 val loss 6.6836
+5300 val perplexity 799.2283
+5300 train 6.820241 (lr=9.7691e-05) (hash(x)=52001684)
+5900 val loss 6.6940
+5900 val perplexity 807.5531
+5900 train 6.839691 (lr=4.8561e-05) (hash(x)=50412818)
+5900 val loss 6.8624
+5900 val perplexity 955.6297
+5900 train 6.993732 (lr=6.7985e-05) (hash(x)=50412818)
+5400 val loss 6.6707
+5400 val perplexity 788.9334
+5400 train 6.575157 (lr=9.7600e-05) (hash(x)=48831647)
+6000 val loss 6.6806
+6000 val perplexity 796.7877
+6000 train 6.477222 (lr=4.8511e-05) (hash(x)=47159634)
+6000 val loss 6.8573
+6000 val perplexity 950.7881
+6000 train 6.679455 (lr=6.7915e-05) (hash(x)=47159634)
+5500 val loss 6.6925
+5500 val perplexity 806.3149
+5500 train 7.105333 (lr=9.7508e-05) (hash(x)=50192069)
+6100 val loss 6.6626
+6100 val perplexity 782.5916
+6100 train 6.761533 (lr=4.8459e-05) (hash(x)=54312795)
+6100 val loss 6.8378
+6100 val perplexity 932.4249
+6100 train 6.926886 (lr=6.7843e-05) (hash(x)=54312795)
+6200 val loss 6.6512
+6200 val perplexity 773.7161
+6200 train 6.779728 (lr=4.8407e-05) (hash(x)=54187587)
+5600 val loss 6.6484
+5600 val perplexity 771.5285
+5600 train 6.471294 (lr=9.7414e-05) (hash(x)=47208852)
+6200 val loss 6.8263
+6200 val perplexity 921.8000
+6200 train 6.935552 (lr=6.7770e-05) (hash(x)=54187587)
+6300 val loss 6.6467
+6300 val perplexity 770.2537
+6300 train 6.743858 (lr=4.8355e-05) (hash(x)=53620387)
+5700 val loss 6.6725
+5700 val perplexity 790.3552
+5700 train 6.252309 (lr=9.7318e-05) (hash(x)=44061694)
+6300 val loss 6.8196
+6300 val perplexity 915.6463
+6300 train 6.899407 (lr=6.7696e-05) (hash(x)=53620387)
+6400 val loss 6.6123
+6400 val perplexity 744.2178
+6400 train 6.562149 (lr=4.8301e-05) (hash(x)=48761774)
+5800 val loss 6.6491
+5800 val perplexity 772.0687
+5800 train 6.999820 (lr=9.7221e-05) (hash(x)=56513279)
+6400 val loss 6.7858
+6400 val perplexity 885.2179
+6400 train 6.744049 (lr=6.7621e-05) (hash(x)=48761774)
+6500 val loss 6.6012
+6500 val perplexity 735.9983
+6500 train 6.817581 (lr=4.8246e-05) (hash(x)=56690281)
+6500 val loss 6.7862
+6500 val perplexity 885.5020
+6500 train 7.016493 (lr=6.7545e-05) (hash(x)=56690281)
+5900 val loss 6.6350
+5900 val perplexity 761.2691
+5900 train 6.762035 (lr=9.7122e-05) (hash(x)=50412818)
+6600 val loss 6.5925
+6600 val perplexity 729.6180
+6600 train 6.397733 (lr=4.8191e-05) (hash(x)=42985269)
+6600 val loss 6.7616
+6600 val perplexity 863.9914
+6600 train 6.561972 (lr=6.7467e-05) (hash(x)=42985269)
+6000 val loss 6.6299
+6000 val perplexity 757.4346
+6000 train 6.458057 (lr=9.7021e-05) (hash(x)=47159634)
+6700 val loss 6.6130
+6700 val perplexity 744.6924
+6700 train 6.649519 (lr=4.8135e-05) (hash(x)=53315447)
+6700 val loss 6.8267
+6700 val perplexity 922.1284
+6700 train 6.879135 (lr=6.7389e-05) (hash(x)=53315447)
+6100 val loss 6.5997
+6100 val perplexity 734.8572
+6100 train 6.667270 (lr=9.6919e-05) (hash(x)=54312795)
+6800 val loss 6.5657
+6800 val perplexity 710.2944
+6800 train 7.060288 (lr=4.8078e-05) (hash(x)=61577166)
+6800 val loss 6.7405
+6800 val perplexity 845.9574
+6800 train 7.170722 (lr=6.7309e-05) (hash(x)=61577166)
+6200 val loss 6.6275
+6200 val perplexity 755.5559
+6200 train 6.753746 (lr=9.6815e-05) (hash(x)=54187587)
+6900 val loss 6.5577
+6900 val perplexity 704.6382
+6900 train 6.695563 (lr=4.8020e-05) (hash(x)=54641005)
+6900 val loss 6.7316
+6900 val perplexity 838.4496
+6900 train 6.866562 (lr=6.7228e-05) (hash(x)=54641005)
+7000 val loss 6.5502
+7000 val perplexity 699.4157
+7000 train 7.104306 (lr=4.7961e-05) (hash(x)=60579512)
+6300 val loss 6.5743
+6300 val perplexity 716.4271
+6300 train 6.656224 (lr=9.6709e-05) (hash(x)=53620387)
+7000 val loss 6.7276
+7000 val perplexity 835.1517
+7000 train 7.248949 (lr=6.7146e-05) (hash(x)=60579512)
+7100 val loss 6.5454
+7100 val perplexity 696.0660
+7100 train 6.414138 (lr=4.7902e-05) (hash(x)=53151549)
+6400 val loss 6.5723
+6400 val perplexity 715.0309
+6400 train 6.517745 (lr=9.6602e-05) (hash(x)=48761774)
+7100 val loss 6.7227
+7100 val perplexity 831.0735
+7100 train 6.612114 (lr=6.7063e-05) (hash(x)=53151549)
+7200 val loss 6.5428
+7200 val perplexity 694.2211
+7200 train 7.489408 (lr=4.7842e-05) (hash(x)=71842455)
+6500 val loss 6.5639
+6500 val perplexity 709.0126
+6500 train 6.793872 (lr=9.6493e-05) (hash(x)=56690281)
+7200 val loss 6.7122
+7200 val perplexity 822.3823
+7200 train 7.659578 (lr=6.6978e-05) (hash(x)=71842455)
+7300 val loss 6.5334
+7300 val perplexity 687.7208
+7300 train 6.287797 (lr=4.7781e-05) (hash(x)=44516452)
+6600 val loss 6.5889
+6600 val perplexity 727.0155
+6600 train 6.406914 (lr=9.6382e-05) (hash(x)=42985269)
+7300 val loss 6.7050
+7300 val perplexity 816.5002
+7300 train 6.460290 (lr=6.6893e-05) (hash(x)=44516452)
+7400 val loss 6.5283
+7400 val perplexity 684.2463
+7400 train 6.159605 (lr=4.7719e-05) (hash(x)=42667710)
+6700 val loss 6.6013
+6700 val perplexity 736.0158
+6700 train 6.640049 (lr=9.6270e-05) (hash(x)=53315447)
+7400 val loss 6.7047
+7400 val perplexity 816.1969
+7400 train 6.357360 (lr=6.6806e-05) (hash(x)=42667710)
+7500 val loss 6.5114
+7500 val perplexity 672.7684
+7500 train 6.242702 (lr=4.7656e-05) (hash(x)=47050797)
+6800 val loss 6.5671
+6800 val perplexity 711.2753
+6800 train 7.001522 (lr=9.6156e-05) (hash(x)=61577166)
+7500 val loss 6.6879
+7500 val perplexity 802.6105
+7500 train 6.419368 (lr=6.6718e-05) (hash(x)=47050797)
+7600 val loss 6.5051
+7600 val perplexity 668.5150
+7600 train 6.399702 (lr=4.7593e-05) (hash(x)=49785056)
+6900 val loss 6.5791
+6900 val perplexity 719.8727
+7600 val loss 6.7069
+7600 val perplexity 818.0633
+6900 train 6.707396 (lr=9.6040e-05) (hash(x)=54641005)
+7600 train 6.603954 (lr=6.6630e-05) (hash(x)=49785056)
+7700 val loss 6.4963
+7700 val perplexity 662.7036
+7700 train 6.285435 (lr=4.7528e-05) (hash(x)=53232030)
+7700 val loss 6.6720
+7700 val perplexity 790.0093
+7700 train 6.461568 (lr=6.6540e-05) (hash(x)=53232030)
+7000 val loss 6.5579
+7000 val perplexity 704.8250
+7000 train 7.086350 (lr=9.5923e-05) (hash(x)=60579512)
+7800 val loss 6.4855
+7800 val perplexity 655.5969
+7800 train 6.298987 (lr=4.7463e-05) (hash(x)=48049749)
+7800 val loss 6.6668
+7800 val perplexity 785.8955
+7800 train 6.494149 (lr=6.6448e-05) (hash(x)=48049749)
+7100 val loss 6.5446
+7100 val perplexity 695.4625
+7100 train 6.413547 (lr=9.5804e-05) (hash(x)=53151549)
+7900 val loss 6.4717
+7900 val perplexity 646.5621
+7900 train 6.300259 (lr=4.7397e-05) (hash(x)=44768513)
+7900 val loss 6.6829
+7900 val perplexity 798.5929
+7900 train 6.505653 (lr=6.6356e-05) (hash(x)=44768513)
+7200 val loss 6.5303
+7200 val perplexity 685.6115
+7200 train 7.385260 (lr=9.5683e-05) (hash(x)=71842455)
+8000 val loss 6.4533
+8000 val perplexity 634.7982
+8000 train 6.299805 (lr=4.7331e-05) (hash(x)=46228039)
+8000 val loss 6.6378
+8000 val perplexity 763.4367
+8000 train 6.497655 (lr=6.6263e-05) (hash(x)=46228039)
+8100 val loss 6.4586
+8100 val perplexity 638.1406
+8100 train 6.806199 (lr=4.7263e-05) (hash(x)=60017091)
+7300 val loss 6.5274
+7300 val perplexity 683.5856
+7300 train 6.305147 (lr=9.5561e-05) (hash(x)=44516452)
+8100 val loss 6.6301
+8100 val perplexity 757.5613
+8100 train 6.976375 (lr=6.6169e-05) (hash(x)=60017091)
+8200 val loss 6.4392
+8200 val perplexity 625.8813
+8200 train 6.448041 (lr=4.7195e-05) (hash(x)=49910198)
+7400 val loss 6.5183
+7400 val perplexity 677.4284
+7400 train 6.144113 (lr=9.5437e-05) (hash(x)=42667710)
+8200 val loss 6.6033
+8200 val perplexity 737.5395
+8200 train 6.593031 (lr=6.6073e-05) (hash(x)=49910198)
+8300 val loss 6.4276
+8300 val perplexity 618.6628
+8300 train 6.752931 (lr=4.7126e-05) (hash(x)=57919055)
+7500 val loss 6.5126
+7500 val perplexity 673.5985
+7500 train 6.239816 (lr=9.5312e-05) (hash(x)=47050797)
+8300 val loss 6.6012
+8300 val perplexity 735.9520
+8300 train 6.932031 (lr=6.5976e-05) (hash(x)=57919055)
+8400 val loss 6.4415
+8400 val perplexity 627.3317
+8400 train 6.512486 (lr=4.7056e-05) (hash(x)=49694964)
+7600 val loss 6.5044
+7600 val perplexity 668.0494
+7600 train 6.396592 (lr=9.5185e-05) (hash(x)=49785056)
+8400 val loss 6.5869
+8400 val perplexity 725.5267
+8400 train 6.661754 (lr=6.5879e-05) (hash(x)=49694964)
+8500 val loss 6.4245
+8500 val perplexity 616.7479
+8500 train 6.462635 (lr=4.6986e-05) (hash(x)=53762585)
+7700 val loss 6.5081
+7700 val perplexity 670.5755
+7700 train 6.280765 (lr=9.5057e-05) (hash(x)=53232030)
+8500 val loss 6.5770
+8500 val perplexity 718.4002
+8500 train 6.605002 (lr=6.5780e-05) (hash(x)=53762585)
+8600 val loss 6.4246
+8600 val perplexity 616.8341
+8600 train 6.470189 (lr=4.6914e-05) (hash(x)=51166973)
+7800 val loss 6.5016
+7800 val perplexity 666.1758
+7800 train 6.321699 (lr=9.4926e-05) (hash(x)=48049749)
+8600 val loss 6.5822
+8600 val perplexity 722.0977
+8600 train 6.605167 (lr=6.5680e-05) (hash(x)=51166973)
+8700 val loss 6.4141
+8700 val perplexity 610.3843
+8700 train 6.485646 (lr=4.6842e-05) (hash(x)=53968049)
+7900 val loss 6.4816
+7900 val perplexity 653.0329
+7900 train 6.317730 (lr=9.4795e-05) (hash(x)=44768513)
+8700 val loss 6.5770
+8700 val perplexity 718.3944
+8700 train 6.651061 (lr=6.5579e-05) (hash(x)=53968049)
+8800 val loss 6.4053
+8800 val perplexity 605.0353
+8800 train 6.495617 (lr=4.6769e-05) (hash(x)=59231056)
+8000 val loss 6.4775
+8000 val perplexity 650.3123
+8000 train 6.328767 (lr=9.4661e-05) (hash(x)=46228039)
+8800 val loss 6.5543
+8800 val perplexity 702.2714
+8800 train 6.639665 (lr=6.5477e-05) (hash(x)=59231056)
+8900 val loss 6.4061
+8900 val perplexity 605.5363
+8900 train 6.274081 (lr=4.6696e-05) (hash(x)=50488048)
+8100 val loss 6.5112
+8100 val perplexity 672.6090
+8100 train 6.866035 (lr=9.4526e-05) (hash(x)=60017091)
+8900 val loss 6.5646
+8900 val perplexity 709.5159
+8900 train 6.441329 (lr=6.5374e-05) (hash(x)=50488048)
+9000 val loss 6.4001
+9000 val perplexity 601.9169
+9000 train 6.120277 (lr=4.6621e-05) (hash(x)=44492956)
+8200 val loss 6.4541
+8200 val perplexity 635.3290
+9000 val loss 6.5750
+9000 val perplexity 716.9766
+8200 train 6.460249 (lr=9.4390e-05) (hash(x)=49910198)
+9000 train 6.295722 (lr=6.5270e-05) (hash(x)=44492956)
+9100 val loss 6.4258
+9100 val perplexity 617.5504
+9100 train 6.443644 (lr=4.6546e-05) (hash(x)=51134989)
+9100 val loss 6.5854
+9100 val perplexity 724.4301
+9100 train 6.569467 (lr=6.5164e-05) (hash(x)=51134989)
+8300 val loss 6.4611
+8300 val perplexity 639.7861
+8300 train 6.775478 (lr=9.4252e-05) (hash(x)=57919055)
+9200 val loss 6.3941
+9200 val perplexity 598.3110
+9200 train 6.185190 (lr=4.6470e-05) (hash(x)=48636056)
+9200 val loss 6.5649
+9200 val perplexity 709.7270
+9200 train 6.369600 (lr=6.5058e-05) (hash(x)=48636056)
+8400 val loss 6.4673
+8400 val perplexity 643.7472
+8400 train 6.531004 (lr=9.4112e-05) (hash(x)=49694964)
+9300 val loss 6.3914
+9300 val perplexity 596.6716
+9300 train 6.323339 (lr=4.6393e-05) (hash(x)=50200551)
+9300 val loss 6.5857
+9300 val perplexity 724.6741
+9300 train 6.534290 (lr=6.4951e-05) (hash(x)=50200551)
+9400 val loss 6.3927
+9400 val perplexity 597.4857
+9400 train 6.197774 (lr=4.6316e-05) (hash(x)=48057228)
+8500 val loss 6.4442
+8500 val perplexity 629.0693
+8500 train 6.469104 (lr=9.3971e-05) (hash(x)=53762585)
+9400 val loss 6.5753
+9400 val perplexity 717.1948
+9400 train 6.404617 (lr=6.4842e-05) (hash(x)=48057228)
+9500 val loss 6.3774
+9500 val perplexity 588.3885
+9500 train 6.152433 (lr=4.6238e-05) (hash(x)=48125171)
+8600 val loss 6.4367
+8600 val perplexity 624.3689
+8600 train 6.454939 (lr=9.3828e-05) (hash(x)=51166973)
+9500 val loss 6.5685
+9500 val perplexity 712.2860
+9500 train 6.320446 (lr=6.4733e-05) (hash(x)=48125171)
+9600 val loss 6.3833
+9600 val perplexity 591.8793
+9600 train 6.289352 (lr=4.6159e-05) (hash(x)=53375853)
+8700 val loss 6.4376
+8700 val perplexity 624.8827
+8700 train 6.501090 (lr=9.3684e-05) (hash(x)=53968049)
+9600 val loss 6.5832
+9600 val perplexity 722.8222
+9600 train 6.485303 (lr=6.4622e-05) (hash(x)=53375853)
+9700 val loss 6.3611
+9700 val perplexity 578.8823
+9700 train 7.147769 (lr=4.6079e-05) (hash(x)=53924631)
+8800 val loss 6.4275
+8800 val perplexity 618.6161
+8800 train 6.505811 (lr=9.3538e-05) (hash(x)=59231056)
+9700 val loss 6.5385
+9700 val perplexity 691.2393
+9700 train 7.267034 (lr=6.4511e-05) (hash(x)=53924631)
+9800 val loss 6.3427
+9800 val perplexity 568.3274
+9800 train 6.385958 (lr=4.5999e-05) (hash(x)=48895047)
+8900 val loss 6.4225
+8900 val perplexity 615.5196
+8900 train 6.296882 (lr=9.3391e-05) (hash(x)=50488048)
+9800 val loss 6.5275
+9800 val perplexity 683.7016
+9800 train 6.557533 (lr=6.4398e-05) (hash(x)=48895047)
+9900 val loss 6.3405
+9900 val perplexity 567.0519
+9900 train 6.189950 (lr=4.5917e-05) (hash(x)=44269923)
+9000 val loss 6.4359
+9000 val perplexity 623.8701
+9000 train 6.163263 (lr=9.3242e-05) (hash(x)=44492956)
+9900 val loss 6.5423
+9900 val perplexity 693.8755
+9900 train 6.399669 (lr=6.4284e-05) (hash(x)=44269923)
+10000 val loss 6.3566
+10000 val perplexity 576.3036
+10000 train 6.248072 (lr=4.5835e-05) (hash(x)=49666623)
+9100 val loss 6.4585
+9100 val perplexity 638.0742
+9100 train 6.373821 (lr=9.3092e-05) (hash(x)=51134989)
+10000 val loss 6.5586
+10000 val perplexity 705.2722
+10000 train 6.450226 (lr=6.4170e-05) (hash(x)=49666623)
+10100 val loss 6.3390
+10100 val perplexity 566.2075
+10100 train 5.839239 (lr=4.5753e-05) (hash(x)=39732485)
+9200 val loss 6.4227
+9200 val perplexity 615.6746
+9200 train 6.224191 (lr=9.2940e-05) (hash(x)=48636056)
+10100 val loss 6.5345
+10100 val perplexity 688.5092
+10100 train 6.054434 (lr=6.4054e-05) (hash(x)=39732485)
+10200 val loss 6.3173
+10200 val perplexity 554.0823
+10200 train 7.417423 (lr=4.5669e-05) (hash(x)=45176813)
+9300 val loss 6.4372
+9300 val perplexity 624.6777
+9300 train 6.348765 (lr=9.2786e-05) (hash(x)=50200551)
+10200 val loss 6.5042
+10200 val perplexity 667.9411
+10200 train 7.562184 (lr=6.3937e-05) (hash(x)=45176813)
+10300 val loss 6.3193
+10300 val perplexity 555.1849
+10300 train 6.179464 (lr=4.5585e-05) (hash(x)=47779031)
+10300 val loss 6.4966
+10300 val perplexity 662.9116
+10300 train 6.374511 (lr=6.3820e-05) (hash(x)=47779031)
+9400 val loss 6.4242
+9400 val perplexity 616.5568
+9400 train 6.235063 (lr=9.2632e-05) (hash(x)=48057228)
+10400 val loss 6.3183
+10400 val perplexity 554.6057
+10400 train 6.190692 (lr=4.5501e-05) (hash(x)=47802708)
+10400 val loss 6.4972
+10400 val perplexity 663.3062
+10400 train 6.364641 (lr=6.3701e-05) (hash(x)=47802708)
+9500 val loss 6.3993
+9500 val perplexity 601.4312
+9500 train 6.173857 (lr=9.2475e-05) (hash(x)=48125171)
+10500 val loss 6.3165
+10500 val perplexity 553.6346
+10500 train 6.168552 (lr=4.5415e-05) (hash(x)=47245542)
+10500 val loss 6.4939
+10500 val perplexity 661.0754
+10500 train 6.316882 (lr=6.3581e-05) (hash(x)=47245542)
+9600 val loss 6.4629
+9600 val perplexity 640.9107
+9600 train 6.345786 (lr=9.2317e-05) (hash(x)=53375853)
+10600 val loss 6.3047
+10600 val perplexity 547.1288
+10600 train 6.354541 (lr=4.5329e-05) (hash(x)=50902224)
+10600 val loss 6.4867
+10600 val perplexity 656.3573
+10600 train 6.530337 (lr=6.3460e-05) (hash(x)=50902224)
+9700 val loss 6.4034
+9700 val perplexity 603.8904
+9700 train 7.180225 (lr=9.2158e-05) (hash(x)=53924631)
+10700 val loss 6.3065
+10700 val perplexity 548.1271
+10700 train 6.179185 (lr=4.5242e-05) (hash(x)=51581765)
+10700 val loss 6.5164
+10700 val perplexity 676.1608
+10700 train 6.390903 (lr=6.3339e-05) (hash(x)=51581765)
+9800 val loss 6.3846
+9800 val perplexity 592.6229
+9800 train 6.434079 (lr=9.1997e-05) (hash(x)=48895047)
+10800 val loss 6.2876
+10800 val perplexity 537.8709
+10800 train 6.323123 (lr=4.5154e-05) (hash(x)=50137616)
+10800 val loss 6.4807
+10800 val perplexity 652.4384
+10800 train 6.485484 (lr=6.3216e-05) (hash(x)=50137616)
+10900 val loss 6.2884
+10900 val perplexity 538.2848
+10900 train 5.996392 (lr=4.5066e-05) (hash(x)=45446745)
+9900 val loss 6.4061
+9900 val perplexity 605.5161
+9900 train 6.268369 (lr=9.1835e-05) (hash(x)=44269923)
+10900 val loss 6.4753
+10900 val perplexity 648.9175
+10900 train 6.172770 (lr=6.3092e-05) (hash(x)=45446745)
+11000 val loss 6.2933
+11000 val perplexity 540.9515
+11000 train 6.114547 (lr=4.4977e-05) (hash(x)=50403725)
+10000 val loss 6.4034
+10000 val perplexity 603.8922
+10000 train 6.308901 (lr=9.1671e-05) (hash(x)=49666623)
+11000 val loss 6.4815
+11000 val perplexity 652.9236
+11000 train 6.322019 (lr=6.2968e-05) (hash(x)=50403725)
+11100 val loss 6.2851
+11100 val perplexity 536.5330
+11100 train 5.760800 (lr=4.4887e-05) (hash(x)=42015206)
+11100 val loss 6.4642
+11100 val perplexity 641.7599
+11100 train 5.946821 (lr=6.2842e-05) (hash(x)=42015206)
+10100 val loss 6.3869
+10100 val perplexity 594.0414
+10100 train 5.912936 (lr=9.1506e-05) (hash(x)=39732485)
+11200 val loss 6.2854
+11200 val perplexity 536.6940
+11200 train 6.147406 (lr=4.4797e-05) (hash(x)=52280259)
+11200 val loss 6.4705
+11200 val perplexity 645.8220
+11200 train 6.342646 (lr=6.2715e-05) (hash(x)=52280259)
+10200 val loss 6.3806
+10200 val perplexity 590.3018
+10200 train 7.562238 (lr=9.1339e-05) (hash(x)=45176813)
+11300 val loss 6.2873
+11300 val perplexity 537.6758
+11300 train 6.038785 (lr=4.4706e-05) (hash(x)=50264744)
+11300 val loss 6.4712
+11300 val perplexity 646.2400
+11300 train 6.238277 (lr=6.2588e-05) (hash(x)=50264744)
+10300 val loss 6.3753
+10300 val perplexity 587.1789
+10300 train 6.239228 (lr=9.1171e-05) (hash(x)=47779031)
+11400 val loss 6.2907
+11400 val perplexity 539.5180
+11400 train 6.176596 (lr=4.4614e-05) (hash(x)=49817375)
+11400 val loss 6.4692
+11400 val perplexity 644.9919
+11400 train 6.366453 (lr=6.2459e-05) (hash(x)=49817375)
+10400 val loss 6.3915
+10400 val perplexity 596.7308
+10400 train 6.279093 (lr=9.1001e-05) (hash(x)=47802708)
+11500 val loss 6.2874
+11500 val perplexity 537.7419
+11500 train 6.267765 (lr=4.4521e-05) (hash(x)=51662160)
+11500 val loss 6.4767
+11500 val perplexity 649.7969
+11500 train 6.471603 (lr=6.2330e-05) (hash(x)=51662160)
+10500 val loss 6.3759
+10500 val perplexity 587.5049
+10500 train 6.220656 (lr=9.0830e-05) (hash(x)=47245542)
+11600 val loss 6.2850
+11600 val perplexity 536.4857
+11600 train 6.226350 (lr=4.4428e-05) (hash(x)=48891495)
+11600 val loss 6.4661
+11600 val perplexity 642.9567
+11600 train 6.398019 (lr=6.2199e-05) (hash(x)=48891495)
+10600 val loss 6.3739
+10600 val perplexity 586.3585
+10600 train 6.414763 (lr=9.0658e-05) (hash(x)=50902224)
+11700 val loss 6.2578
+11700 val perplexity 522.0562
+11700 train 6.255913 (lr=4.4334e-05) (hash(x)=46306682)
+11700 val loss 6.4435
+11700 val perplexity 628.6244
+11700 train 6.432802 (lr=6.2068e-05) (hash(x)=46306682)
+10700 val loss 6.3873
+10700 val perplexity 594.2664
+10700 train 6.234374 (lr=9.0484e-05) (hash(x)=51581765)
+11800 val loss 6.2535
+11800 val perplexity 519.8424
+11800 train 6.217044 (lr=4.4240e-05) (hash(x)=46605293)
+11800 val loss 6.4419
+11800 val perplexity 627.6228
+11800 train 6.387120 (lr=6.1936e-05) (hash(x)=46605293)
+10800 val loss 6.3749
+10800 val perplexity 586.9379
+10800 train 6.412324 (lr=9.0308e-05) (hash(x)=50137616)
+11900 val loss 6.2447
+11900 val perplexity 515.2994
+11900 train 6.179983 (lr=4.4145e-05) (hash(x)=53746201)
+11900 val loss 6.4192
+11900 val perplexity 613.4951
+11900 train 6.365595 (lr=6.1802e-05) (hash(x)=53746201)
+12000 val loss 6.2409
+12000 val perplexity 513.2961
+12000 train 6.334722 (lr=4.4049e-05) (hash(x)=48587180)
+10900 val loss 6.3581
+10900 val perplexity 577.1487
+10900 train 6.080709 (lr=9.0132e-05) (hash(x)=45446745)
+12000 val loss 6.4090
+12000 val perplexity 607.2846
+12000 train 6.480982 (lr=6.1668e-05) (hash(x)=48587180)
+12100 val loss 6.2393
+12100 val perplexity 512.4902
+12100 train 6.029146 (lr=4.3952e-05) (hash(x)=45786729)
+11000 val loss 6.3604
+11000 val perplexity 578.4651
+11000 train 6.196943 (lr=8.9954e-05) (hash(x)=50403725)
+12100 val loss 6.4212
+12100 val perplexity 614.7646
+12100 train 6.189254 (lr=6.1533e-05) (hash(x)=45786729)
+12200 val loss 6.2346
+12200 val perplexity 510.0944
+12200 train 6.160137 (lr=4.3855e-05) (hash(x)=42858336)
+11100 val loss 6.3505
+11100 val perplexity 572.7651
+11100 train 5.849023 (lr=8.9774e-05) (hash(x)=42015206)
+12200 val loss 6.4066
+12200 val perplexity 605.8104
+12200 train 6.332368 (lr=6.1397e-05) (hash(x)=42858336)
+12300 val loss 6.2333
+12300 val perplexity 509.4522
+12300 train 6.406464 (lr=4.3757e-05) (hash(x)=51944313)
+11200 val loss 6.3685
+11200 val perplexity 583.1981
+11200 train 6.236219 (lr=8.9593e-05) (hash(x)=52280259)
+12300 val loss 6.4023
+12300 val perplexity 603.2524
+12300 train 6.598873 (lr=6.1260e-05) (hash(x)=51944313)
+12400 val loss 6.2126
+12400 val perplexity 498.9987
+12400 train 6.714368 (lr=4.3659e-05) (hash(x)=59357191)
+12400 val loss 6.3986
+12400 val perplexity 601.0114
+12400 train 6.895336 (lr=6.1122e-05) (hash(x)=59357191)
+11300 val loss 6.3664
+11300 val perplexity 581.9650
+11300 train 6.100276 (lr=8.9411e-05) (hash(x)=50264744)
+12500 val loss 6.2195
+12500 val perplexity 502.4362
+12500 train 6.334024 (lr=4.3560e-05) (hash(x)=49770062)
+12500 val loss 6.4069
+12500 val perplexity 606.0037
+12500 train 6.522009 (lr=6.0984e-05) (hash(x)=49770062)
+11400 val loss 6.3802
+11400 val perplexity 590.0237
+11400 train 6.277043 (lr=8.9227e-05) (hash(x)=49817375)
+12600 val loss 6.2030
+12600 val perplexity 494.2460
+12600 train 5.980834 (lr=4.3460e-05) (hash(x)=47235788)
+12600 val loss 6.3836
+12600 val perplexity 592.0518
+12600 train 6.151051 (lr=6.0844e-05) (hash(x)=47235788)
+11500 val loss 6.3619
+11500 val perplexity 579.3739
+11500 train 6.346138 (lr=8.9043e-05) (hash(x)=51662160)
+12700 val loss 6.2019
+12700 val perplexity 493.7014
+12700 train 6.207354 (lr=4.3360e-05) (hash(x)=51700943)
+12700 val loss 6.3822
+12700 val perplexity 591.2164
+12700 train 6.418068 (lr=6.0703e-05) (hash(x)=51700943)
+11600 val loss 6.3612
+11600 val perplexity 578.9637
+11600 train 6.309485 (lr=8.8856e-05) (hash(x)=48891495)
+12800 val loss 6.2139
+12800 val perplexity 499.6537
+12800 train 6.169513 (lr=4.3259e-05) (hash(x)=50532124)
+12800 val loss 6.3777
+12800 val perplexity 588.5892
+12800 train 6.337963 (lr=6.0562e-05) (hash(x)=50532124)
+12900 val loss 6.1960
+12900 val perplexity 490.7574
+12900 train 6.306409 (lr=4.3157e-05) (hash(x)=51948273)
+11700 val loss 6.3349
+11700 val perplexity 563.9124
+11700 train 6.321611 (lr=8.8668e-05) (hash(x)=46306682)
+12900 val loss 6.3772
+12900 val perplexity 588.2719
+12900 train 6.483133 (lr=6.0420e-05) (hash(x)=51948273)
+13000 val loss 6.2050
+13000 val perplexity 495.2059
+13000 train 6.372721 (lr=4.3055e-05) (hash(x)=46732738)
+11800 val loss 6.3301
+11800 val perplexity 561.2119
+11800 train 6.298628 (lr=8.8479e-05) (hash(x)=46605293)
+13000 val loss 6.3865
+13000 val perplexity 593.7670
+13000 train 6.531954 (lr=6.0277e-05) (hash(x)=46732738)
+13100 val loss 6.1958
+13100 val perplexity 490.7025
+13100 train 6.221768 (lr=4.2952e-05) (hash(x)=49837250)
+11900 val loss 6.3153
+11900 val perplexity 552.9648
+11900 train 6.255790 (lr=8.8289e-05) (hash(x)=53746201)
+13100 val loss 6.3670
+13100 val perplexity 582.2797
+13100 train 6.390731 (lr=6.0133e-05) (hash(x)=49837250)
+13200 val loss 6.2022
+13200 val perplexity 493.8576
+13200 train 6.488420 (lr=4.2848e-05) (hash(x)=57646893)
+12000 val loss 6.3122
+12000 val perplexity 551.2727
+12000 train 6.383875 (lr=8.8097e-05) (hash(x)=48587180)
+13200 val loss 6.3708
+13200 val perplexity 584.5456
+13200 train 6.642096 (lr=5.9988e-05) (hash(x)=57646893)
+13300 val loss 6.1930
+13300 val perplexity 489.3283
+13300 train 5.744956 (lr=4.2744e-05) (hash(x)=42055305)
+12100 val loss 6.3312
+12100 val perplexity 561.8511
+12100 train 6.104850 (lr=8.7904e-05) (hash(x)=45786729)
+13300 val loss 6.3628
+13300 val perplexity 579.8581
+13300 train 5.922172 (lr=5.9842e-05) (hash(x)=42055305)
+13400 val loss 6.2072
+13400 val perplexity 496.3196
+13400 train 5.900372 (lr=4.2640e-05) (hash(x)=44684844)
+12200 val loss 6.3282
+12200 val perplexity 560.1316
+12200 train 6.244182 (lr=8.7710e-05) (hash(x)=42858336)
+13400 val loss 6.3918
+13400 val perplexity 596.9084
+13400 train 6.109066 (lr=5.9695e-05) (hash(x)=44684844)
+13500 val loss 6.1840
+13500 val perplexity 484.9107
+13500 train 5.862307 (lr=4.2534e-05) (hash(x)=46971715)
+12300 val loss 6.3317
+12300 val perplexity 562.0834
+12300 train 6.538304 (lr=8.7515e-05) (hash(x)=51944313)
+13500 val loss 6.3781
+13500 val perplexity 588.7840
+13500 train 6.054822 (lr=5.9548e-05) (hash(x)=46971715)
+13600 val loss 6.1918
+13600 val perplexity 488.7330
+13600 train 6.017723 (lr=4.2428e-05) (hash(x)=48728506)
+12400 val loss 6.3200
+12400 val perplexity 555.5686
+12400 train 6.784978 (lr=8.7318e-05) (hash(x)=59357191)
+13600 val loss 6.3750
+13600 val perplexity 586.9799
+13600 train 6.203294 (lr=5.9400e-05) (hash(x)=48728506)
+13700 val loss 6.1925
+13700 val perplexity 489.0685
+13700 train 5.980932 (lr=4.2322e-05) (hash(x)=50739726)
+12500 val loss 6.3089
+12500 val perplexity 549.4297
+13700 val loss 6.3813
+13700 val perplexity 590.6945
+13700 train 6.179378 (lr=5.9251e-05) (hash(x)=50739726)
+12500 train 6.419203 (lr=8.7119e-05) (hash(x)=49770062)
+13800 val loss 6.1862
+13800 val perplexity 486.0093
+13800 train 6.287631 (lr=4.2215e-05) (hash(x)=52174318)
+13800 val loss 6.3732
+13800 val perplexity 585.9345
+13800 train 6.482311 (lr=5.9101e-05) (hash(x)=52174318)
+12600 val loss 6.3111
+12600 val perplexity 550.6550
+12600 train 6.092916 (lr=8.6920e-05) (hash(x)=47235788)
+13900 val loss 6.1776
+13900 val perplexity 481.8454
+13900 train 5.557216 (lr=4.2107e-05) (hash(x)=44141266)
+13900 val loss 6.3569
+13900 val perplexity 576.4279
+13900 train 5.761450 (lr=5.8950e-05) (hash(x)=44141266)
+12700 val loss 6.3178
+12700 val perplexity 554.3524
+12700 train 6.334719 (lr=8.6719e-05) (hash(x)=51700943)
+14000 val loss 6.1711
+14000 val perplexity 478.7227
+14000 train 6.133667 (lr=4.1999e-05) (hash(x)=50880921)
+14000 val loss 6.3533
+14000 val perplexity 574.3917
+14000 train 6.342307 (lr=5.8799e-05) (hash(x)=50880921)
+12800 val loss 6.3147
+12800 val perplexity 552.6302
+12800 train 6.270829 (lr=8.6517e-05) (hash(x)=50532124)
+14100 val loss 6.1645
+14100 val perplexity 475.5845
+14100 train 6.102219 (lr=4.1890e-05) (hash(x)=44179328)
+14100 val loss 6.3473
+14100 val perplexity 570.9208
+14100 train 6.280088 (lr=5.8646e-05) (hash(x)=44179328)
+12900 val loss 6.3020
+12900 val perplexity 545.6871
+12900 train 6.406723 (lr=8.6314e-05) (hash(x)=51948273)
+14200 val loss 6.1537
+14200 val perplexity 470.4647
+14200 train 6.252338 (lr=4.1781e-05) (hash(x)=52989772)
+14200 val loss 6.3343
+14200 val perplexity 563.5675
+14200 train 6.417709 (lr=5.8493e-05) (hash(x)=52989772)
+14300 val loss 6.1503
+14300 val perplexity 468.8498
+14300 train 6.385016 (lr=4.1671e-05) (hash(x)=55406567)
+13000 val loss 6.3186
+13000 val perplexity 554.7871
+13000 train 6.466562 (lr=8.6110e-05) (hash(x)=46732738)
+14300 val loss 6.3292
+14300 val perplexity 560.6941
+14300 train 6.554401 (lr=5.8339e-05) (hash(x)=55406567)
+14400 val loss 6.1478
+14400 val perplexity 467.6949
+14400 train 6.203402 (lr=4.1560e-05) (hash(x)=47274361)
+13100 val loss 6.3065
+13100 val perplexity 548.1247
+13100 train 6.335435 (lr=8.5904e-05) (hash(x)=49837250)
+14400 val loss 6.3307
+14400 val perplexity 561.5286
+14400 train 6.420453 (lr=5.8184e-05) (hash(x)=47274361)
+14500 val loss 6.1361
+14500 val perplexity 462.2333
+14500 train 6.565048 (lr=4.1449e-05) (hash(x)=53333322)
+13200 val loss 6.3183
+13200 val perplexity 554.6125
+13200 train 6.586292 (lr=8.5697e-05) (hash(x)=57646893)
+14500 val loss 6.3286
+14500 val perplexity 560.3875
+14500 train 6.732544 (lr=5.8029e-05) (hash(x)=53333322)
+14600 val loss 6.1396
+14600 val perplexity 463.8659
+14600 train 6.265152 (lr=4.1337e-05) (hash(x)=48578234)
+13300 val loss 6.3036
+13300 val perplexity 546.5611
+13300 train 5.868500 (lr=8.5489e-05) (hash(x)=42055305)
+14600 val loss 6.3275
+14600 val perplexity 559.7396
+14600 train 6.437446 (lr=5.7872e-05) (hash(x)=48578234)
+14700 val loss 6.1651
+14700 val perplexity 475.8719
+14700 train 5.846576 (lr=4.1225e-05) (hash(x)=45694052)
+13400 val loss 6.3046
+13400 val perplexity 547.0742
+13400 train 6.021350 (lr=8.5279e-05) (hash(x)=44684844)
+14700 val loss 6.3525
+14700 val perplexity 573.9468
+14700 train 6.048125 (lr=5.7715e-05) (hash(x)=45694052)
+14800 val loss 6.1441
+14800 val perplexity 465.9411
+14800 train 5.925498 (lr=4.1113e-05) (hash(x)=46245798)
+14800 val loss 6.3277
+14800 val perplexity 559.8675
+14800 train 6.138581 (lr=5.7558e-05) (hash(x)=46245798)
+13500 val loss 6.3102
+13500 val perplexity 550.1745
+13500 train 5.992841 (lr=8.5069e-05) (hash(x)=46971715)
+14900 val loss 6.1411
+14900 val perplexity 464.5530
+14900 train 6.283280 (lr=4.0999e-05) (hash(x)=55970820)
+14900 val loss 6.3363
+14900 val perplexity 564.6882
+14900 train 6.498343 (lr=5.7399e-05) (hash(x)=55970820)
+13600 val loss 6.2936
+13600 val perplexity 541.0870
+13600 train 6.112821 (lr=8.4857e-05) (hash(x)=48728506)
+15000 val loss 6.1492
+15000 val perplexity 468.3618
+15000 train 6.255603 (lr=4.0885e-05) (hash(x)=55635805)
+15000 val loss 6.3258
+15000 val perplexity 558.8156
+15000 train 6.419089 (lr=5.7240e-05) (hash(x)=55635805)
+13700 val loss 6.2896
+13700 val perplexity 538.9217
+13700 train 6.094644 (lr=8.4644e-05) (hash(x)=50739726)
+15100 val loss 6.1270
+15100 val perplexity 458.0396
+15100 train 6.044188 (lr=4.0771e-05) (hash(x)=48884573)
+15100 val loss 6.3032
+15100 val perplexity 546.3135
+15100 train 6.197279 (lr=5.7079e-05) (hash(x)=48884573)
+13800 val loss 6.3031
+13800 val perplexity 546.2784
+13800 train 6.412201 (lr=8.4430e-05) (hash(x)=52174318)
+15200 val loss 6.1233
+15200 val perplexity 456.3787
+15200 train 6.300706 (lr=4.0656e-05) (hash(x)=59283076)
+15200 val loss 6.3089
+15200 val perplexity 549.4617
+15200 train 6.510170 (lr=5.6919e-05) (hash(x)=59283076)
+13900 val loss 6.2823
+13900 val perplexity 535.0106
+13900 train 5.696769 (lr=8.4214e-05) (hash(x)=44141266)
+15300 val loss 6.1188
+15300 val perplexity 454.2971
+15300 train 6.042745 (lr=4.0541e-05) (hash(x)=48258741)
+15300 val loss 6.3095
+15300 val perplexity 549.7542
+15300 train 6.225156 (lr=5.6757e-05) (hash(x)=48258741)
+14000 val loss 6.2851
+14000 val perplexity 536.5412
+14000 train 6.278280 (lr=8.3998e-05) (hash(x)=50880921)
+15400 val loss 6.1319
+15400 val perplexity 460.3272
+15400 train 6.158960 (lr=4.0425e-05) (hash(x)=49433427)
+15400 val loss 6.3035
+15400 val perplexity 546.4634
+15400 train 6.343267 (lr=5.6595e-05) (hash(x)=49433427)
+14100 val loss 6.2732
+14100 val perplexity 530.1518
+14100 train 6.208812 (lr=8.3780e-05) (hash(x)=44179328)
+15500 val loss 6.1150
+15500 val perplexity 452.6169
+15500 val loss 6.3211
+15500 val perplexity 556.1944
+15500 train 5.899845 (lr=4.0308e-05) (hash(x)=42504793)
+15500 train 6.069526 (lr=5.6432e-05) (hash(x)=42504793)
+14200 val loss 6.2535
+14200 val perplexity 519.8256
+14200 train 6.344846 (lr=8.3561e-05) (hash(x)=52989772)
+15600 val loss 6.1173
+15600 val perplexity 453.6441
+15600 train 6.015007 (lr=4.0191e-05) (hash(x)=48860388)
+15600 val loss 6.2998
+15600 val perplexity 544.4466
+15600 train 6.159374 (lr=5.6268e-05) (hash(x)=48860388)
+14300 val loss 6.2454
+14300 val perplexity 515.6539
+14300 train 6.448986 (lr=8.3341e-05) (hash(x)=55406567)
+15700 val loss 6.2799
+15700 val perplexity 533.7578
+15700 val loss 6.0992
+15700 val perplexity 445.5022
+15700 train 6.086686 (lr=5.6104e-05) (hash(x)=49229381)
+15700 train 5.920973 (lr=4.0074e-05) (hash(x)=49229381)
+14400 val loss 6.2421
+14400 val perplexity 513.9424
+14400 train 6.370075 (lr=8.3120e-05) (hash(x)=47274361)
+15800 val loss 6.1038
+15800 val perplexity 447.5447
+15800 train 5.934508 (lr=3.9956e-05) (hash(x)=48471211)
+15800 val loss 6.2983
+15800 val perplexity 543.6398
+15800 train 6.073653 (lr=5.5938e-05) (hash(x)=48471211)
+14500 val loss 6.2622
+14500 val perplexity 524.3958
+14500 train 6.735986 (lr=8.2898e-05) (hash(x)=53333322)
+15900 val loss 6.1048
+15900 val perplexity 447.9807
+15900 train 5.814809 (lr=3.9838e-05) (hash(x)=47555877)
+15900 val loss 6.2992
+15900 val perplexity 544.1287
+15900 train 6.006874 (lr=5.5773e-05) (hash(x)=47555877)
+16000 val loss 6.1117
+16000 val perplexity 451.1030
+16000 train 5.804303 (lr=3.9719e-05) (hash(x)=48397986)
+14600 val loss 6.2519
+14600 val perplexity 519.0102
+14600 train 6.356515 (lr=8.2675e-05) (hash(x)=48578234)
+16000 val loss 6.2971
+16000 val perplexity 542.9921
+16000 train 5.997636 (lr=5.5606e-05) (hash(x)=48397986)
+16100 val loss 6.1077
+16100 val perplexity 449.2983
+16100 train 6.201589 (lr=3.9599e-05) (hash(x)=53299004)
+14700 val loss 6.2862
+14700 val perplexity 537.1274
+14700 train 5.999903 (lr=8.2451e-05) (hash(x)=45694052)
+16100 val loss 6.2833
+16100 val perplexity 535.5461
+16100 train 6.334311 (lr=5.5439e-05) (hash(x)=53299004)
+16200 val loss 6.1119
+16200 val perplexity 451.2162
+16200 train 5.900352 (lr=3.9479e-05) (hash(x)=45763277)
+16200 val loss 6.2982
+16200 val perplexity 543.6183
+16200 train 6.074014 (lr=5.5271e-05) (hash(x)=45763277)
+14800 val loss 6.2675
+14800 val perplexity 527.1354
+14800 train 6.069706 (lr=8.2225e-05) (hash(x)=46245798)
+16300 val loss 6.1061
+16300 val perplexity 448.5692
+16300 train 5.591532 (lr=3.9359e-05) (hash(x)=39675609)
+16300 val loss 6.2827
+16300 val perplexity 535.2198
+16300 train 5.740005 (lr=5.5102e-05) (hash(x)=39675609)
+14900 val loss 6.2519
+14900 val perplexity 519.0052
+14900 train 6.398389 (lr=8.1998e-05) (hash(x)=55970820)
+16400 val loss 6.0969
+16400 val perplexity 444.4576
+16400 train 6.162115 (lr=3.9238e-05) (hash(x)=48973557)
+16400 val loss 6.2844
+16400 val perplexity 536.1591
+16400 train 6.375087 (lr=5.4933e-05) (hash(x)=48973557)
+15000 val loss 6.2376
+15000 val perplexity 511.6510
+15000 train 6.338455 (lr=8.1771e-05) (hash(x)=55635805)
+16500 val loss 6.0883
+16500 val perplexity 440.6679
+16500 train 6.090487 (lr=3.9117e-05) (hash(x)=46121211)
+16500 val loss 6.2651
+16500 val perplexity 525.9087
+16500 train 6.226743 (lr=5.4763e-05) (hash(x)=46121211)
+15100 val loss 6.2410
+15100 val perplexity 513.3783
+15100 train 6.131569 (lr=8.1542e-05) (hash(x)=48884573)
+16600 val loss 6.0918
+16600 val perplexity 442.2100
+16600 train 6.352380 (lr=3.8995e-05) (hash(x)=49181055)
+16600 val loss 6.2623
+16600 val perplexity 524.4390
+16600 train 6.502903 (lr=5.4593e-05) (hash(x)=49181055)
+15200 val loss 6.2412
+15200 val perplexity 513.4949
+15200 train 6.447044 (lr=8.1312e-05) (hash(x)=59283076)
+16700 val loss 6.0925
+16700 val perplexity 442.5309
+16700 train 6.040877 (lr=3.8873e-05) (hash(x)=49361213)
+16700 val loss 6.2583
+16700 val perplexity 522.3468
+16700 train 6.198591 (lr=5.4422e-05) (hash(x)=49361213)
+15300 val loss 6.2669
+15300 val perplexity 526.8291
+15300 train 6.188086 (lr=8.1082e-05) (hash(x)=48258741)
+16800 val loss 6.0950
+16800 val perplexity 443.6194
+16800 train 6.086593 (lr=3.8750e-05) (hash(x)=52739987)
+16800 val loss 6.2646
+16800 val perplexity 525.6302
+16800 train 6.249082 (lr=5.4250e-05) (hash(x)=52739987)
+15400 val loss 6.2680
+15400 val perplexity 527.4374
+15400 train 6.287550 (lr=8.0850e-05) (hash(x)=49433427)
+16900 val loss 6.2472
+16900 val perplexity 516.5728
+16900 train 6.389314 (lr=5.4078e-05) (hash(x)=55994797)
+16900 val loss 6.0762
+16900 val perplexity 435.3646
+16900 train 6.241480 (lr=3.8627e-05) (hash(x)=55994797)
+15500 val loss 6.2683
+15500 val perplexity 527.6044
+15500 train 6.028718 (lr=8.0617e-05) (hash(x)=42504793)
+17000 val loss 6.0823
+17000 val perplexity 438.0379
+17000 train 6.104437 (lr=3.8503e-05) (hash(x)=50109331)
+17000 val loss 6.2478
+17000 val perplexity 516.8850
+17000 train 6.280750 (lr=5.3905e-05) (hash(x)=50109331)
+17100 val loss 6.2370
+17100 val perplexity 511.3335
+17100 train 6.350257 (lr=5.3731e-05) (hash(x)=50292169)
+17100 val loss 6.0777
+17100 val perplexity 436.0165
+17100 train 6.172071 (lr=3.8379e-05) (hash(x)=50292169)
+15600 val loss 6.2485
+15600 val perplexity 517.2164
+15600 train 6.143048 (lr=8.0383e-05) (hash(x)=48860388)
+17200 val loss 6.0749
+17200 val perplexity 434.8055
+17200 train 6.132131 (lr=3.8255e-05) (hash(x)=50722386)
+17200 val loss 6.2532
+17200 val perplexity 519.6669
+17200 train 6.294482 (lr=5.3557e-05) (hash(x)=50722386)
+15700 val loss 6.2353
+15700 val perplexity 510.4428
+15700 train 6.055554 (lr=8.0148e-05) (hash(x)=49229381)
+17300 val loss 6.0610
+17300 val perplexity 428.7892
+17300 train 6.203306 (lr=3.8130e-05) (hash(x)=54010042)
+17300 val loss 6.2296
+17300 val perplexity 507.5507
+17300 train 6.329121 (lr=5.3382e-05) (hash(x)=54010042)
+15800 val loss 6.2615
+15800 val perplexity 524.0236
+15800 train 6.045529 (lr=7.9912e-05) (hash(x)=48471211)
+17400 val loss 6.0610
+17400 val perplexity 428.7897
+17400 train 6.018238 (lr=3.8005e-05) (hash(x)=53970990)
+17400 val loss 6.2369
+17400 val perplexity 511.2510
+17400 train 6.211182 (lr=5.3206e-05) (hash(x)=53970990)
+15900 val loss 6.2596
+15900 val perplexity 523.0138
+15900 train 5.964798 (lr=7.9675e-05) (hash(x)=47555877)
+17500 val loss 6.0548
+17500 val perplexity 426.1338
+17500 val loss 6.2278
+17500 val perplexity 506.6382
+17500 train 6.134178 (lr=3.7879e-05) (hash(x)=52544667)
+17500 train 6.304132 (lr=5.3030e-05) (hash(x)=52544667)
+16000 val loss 6.2763
+16000 val perplexity 531.7962
+16000 train 5.966957 (lr=7.9437e-05) (hash(x)=48397986)
+17600 val loss 6.0604
+17600 val perplexity 428.5526
+17600 train 6.025463 (lr=3.7753e-05) (hash(x)=50735496)
+17600 val loss 6.2313
+17600 val perplexity 508.4021
+17600 train 6.178463 (lr=5.2854e-05) (hash(x)=50735496)
+16100 val loss 6.2599
+16100 val perplexity 523.1580
+16100 train 6.310643 (lr=7.9198e-05) (hash(x)=53299004)
+17700 val loss 6.0583
+17700 val perplexity 427.6525
+17700 train 6.439931 (lr=3.7626e-05) (hash(x)=55662551)
+17700 val loss 6.2334
+17700 val perplexity 509.4889
+17700 train 6.560091 (lr=5.2677e-05) (hash(x)=55662551)
+16200 val loss 6.2676
+16200 val perplexity 527.2297
+16200 train 6.036007 (lr=7.8959e-05) (hash(x)=45763277)
+17800 val loss 6.0554
+17800 val perplexity 426.3960
+17800 train 5.926167 (lr=3.7499e-05) (hash(x)=51137009)
+17800 val loss 6.2278
+17800 val perplexity 506.6174
+17800 train 6.111446 (lr=5.2499e-05) (hash(x)=51137009)
+16300 val loss 6.2505
+16300 val perplexity 518.2463
+16300 train 5.719383 (lr=7.8718e-05) (hash(x)=39675609)
+17900 val loss 6.0569
+17900 val perplexity 427.0648
+17900 train 6.852966 (lr=3.7372e-05) (hash(x)=73004834)
+17900 val loss 6.2341
+17900 val perplexity 509.8237
+17900 train 7.072075 (lr=5.2321e-05) (hash(x)=73004834)
+16400 val loss 6.2352
+16400 val perplexity 510.4156
+16400 train 6.316364 (lr=7.8476e-05) (hash(x)=48973557)
+18000 val loss 6.0514
+18000 val perplexity 424.7275
+18000 train 6.259719 (lr=3.7244e-05) (hash(x)=51900245)
+18000 val loss 6.2339
+18000 val perplexity 509.7630
+18000 train 6.430896 (lr=5.2142e-05) (hash(x)=51900245)
+18100 val loss 6.0417
+18100 val perplexity 420.6046
+18100 train 6.320068 (lr=3.7116e-05) (hash(x)=56278625)
+18100 val loss 6.2255
+18100 val perplexity 505.4630
+18100 train 6.494106 (lr=5.1962e-05) (hash(x)=56278625)
+16500 val loss 6.2256
+16500 val perplexity 505.5129
+16500 train 6.207320 (lr=7.8233e-05) (hash(x)=46121211)
+18200 val loss 6.0523
+18200 val perplexity 425.0752
+18200 train 5.916048 (lr=3.6987e-05) (hash(x)=50478164)
+18200 val loss 6.2215
+18200 val perplexity 503.4511
+18200 train 6.062120 (lr=5.1782e-05) (hash(x)=50478164)
+16600 val loss 6.2345
+16600 val perplexity 510.0611
+16600 train 6.504384 (lr=7.7990e-05) (hash(x)=49181055)
+18300 val loss 6.0451
+18300 val perplexity 422.0386
+18300 train 6.029937 (lr=3.6859e-05) (hash(x)=47837565)
+18300 val loss 6.2258
+18300 val perplexity 505.6358
+18300 train 6.176018 (lr=5.1602e-05) (hash(x)=47837565)
+16700 val loss 6.2452
+16700 val perplexity 515.5415
+16700 train 6.206596 (lr=7.7745e-05) (hash(x)=49361213)
+18400 val loss 6.0513
+18400 val perplexity 424.6834
+18400 train 5.898037 (lr=3.6729e-05) (hash(x)=49622704)
+18400 val loss 6.2292
+18400 val perplexity 507.3617
+18400 train 6.062944 (lr=5.1421e-05) (hash(x)=49622704)
+16800 val loss 6.2345
+16800 val perplexity 510.0693
+16800 train 6.248514 (lr=7.7500e-05) (hash(x)=52739987)
+18500 val loss 6.0398
+18500 val perplexity 419.7979
+18500 train 5.899223 (lr=3.6600e-05) (hash(x)=49741203)
+18500 val loss 6.2203
+18500 val perplexity 502.8362
+18500 train 6.031091 (lr=5.1240e-05) (hash(x)=49741203)
+16900 val loss 6.2274
+16900 val perplexity 506.4341
+16900 train 6.382916 (lr=7.7254e-05) (hash(x)=55994797)
+18600 val loss 6.0501
+18600 val perplexity 424.1730
+18600 train 6.520254 (lr=3.6470e-05) (hash(x)=54690634)
+18600 val loss 6.2280
+18600 val perplexity 506.7186
+18600 train 6.650096 (lr=5.1058e-05) (hash(x)=54690634)
+17000 val loss 6.2225
+17000 val perplexity 503.9757
+17000 train 6.270326 (lr=7.7007e-05) (hash(x)=50109331)
+18700 val loss 6.2301
+18700 val perplexity 507.7843
+18700 train 6.106448 (lr=5.0875e-05) (hash(x)=53778909)
+18700 val loss 6.0491
+18700 val perplexity 423.7169
+18700 train 5.904674 (lr=3.6339e-05) (hash(x)=53778909)
+17100 val loss 6.2159
+17100 val perplexity 500.6514
+17100 train 6.332056 (lr=7.6758e-05) (hash(x)=50292169)
+18800 val loss 6.2064
+18800 val perplexity 495.9312
+18800 train 6.366983 (lr=5.0692e-05) (hash(x)=52196791)
+18800 val loss 6.0363
+18800 val perplexity 418.3224
+18800 train 6.203441 (lr=3.6209e-05) (hash(x)=52196791)
+17200 val loss 6.2173
+17200 val perplexity 501.3246
+17200 train 6.273157 (lr=7.6510e-05) (hash(x)=50722386)
+18900 val loss 6.1859
+18900 val perplexity 485.8321
+18900 train 6.120074 (lr=5.0509e-05) (hash(x)=49452227)
+18900 val loss 6.0208
+18900 val perplexity 411.8900
+18900 train 5.965091 (lr=3.6078e-05) (hash(x)=49452227)
+17300 val loss 6.2028
+17300 val perplexity 494.1235
+17300 train 6.313985 (lr=7.6260e-05) (hash(x)=54010042)
+19000 val loss 6.2000
+19000 val perplexity 492.7433
+19000 train 6.192492 (lr=5.0325e-05) (hash(x)=54220718)
+19000 val loss 6.0268
+19000 val perplexity 414.3761
+19000 train 6.013001 (lr=3.5946e-05) (hash(x)=54220718)
+17400 val loss 6.2195
+17400 val perplexity 502.4561
+17400 train 6.203558 (lr=7.6009e-05) (hash(x)=53970990)
+19100 val loss 6.2064
+19100 val perplexity 495.8910
+19100 train 6.357291 (lr=5.0140e-05) (hash(x)=54899289)
+19100 val loss 6.0357
+19100 val perplexity 418.0925
+19100 train 6.170717 (lr=3.5814e-05) (hash(x)=54899289)
+19200 val loss 6.1756
+19200 val perplexity 480.8738
+19200 train 6.314922 (lr=4.9955e-05) (hash(x)=52676298)
+17500 val loss 6.1983
+17500 val perplexity 491.9195
+17500 train 6.253818 (lr=7.5758e-05) (hash(x)=52544667)
+19200 val loss 6.0158
+19200 val perplexity 409.8688
+19200 train 6.156129 (lr=3.5682e-05) (hash(x)=52676298)
+19300 val loss 6.1720
+19300 val perplexity 479.1528
+19300 train 6.221705 (lr=4.9770e-05) (hash(x)=54877992)
+19300 val loss 6.0196
+19300 val perplexity 411.4324
+19300 train 6.096160 (lr=3.5550e-05) (hash(x)=54877992)
+17600 val loss 6.2033
+17600 val perplexity 494.3816
+17600 train 6.163414 (lr=7.5505e-05) (hash(x)=50735496)
+19400 val loss 6.2017
+19400 val perplexity 493.5873
+19400 train 6.025438 (lr=4.9584e-05) (hash(x)=46432347)
+19400 val loss 6.0182
+19400 val perplexity 410.8215
+19400 train 5.834083 (lr=3.5417e-05) (hash(x)=46432347)
+17700 val loss 6.2076
+17700 val perplexity 496.4959
+17700 train 6.534688 (lr=7.5252e-05) (hash(x)=55662551)
+19500 val loss 6.1630
+19500 val perplexity 474.8275
+19500 train 6.253082 (lr=4.9398e-05) (hash(x)=50677944)
+19500 val loss 6.0098
+19500 val perplexity 407.4086
+19500 train 6.096924 (lr=3.5284e-05) (hash(x)=50677944)
+17800 val loss 6.2070
+17800 val perplexity 496.2302
+17800 train 6.056177 (lr=7.4998e-05) (hash(x)=51137009)
+19600 val loss 6.1729
+19600 val perplexity 479.5898
+19600 train 6.003708 (lr=4.9211e-05) (hash(x)=48855410)
+19600 val loss 6.0119
+19600 val perplexity 408.2649
+19600 train 5.852933 (lr=3.5151e-05) (hash(x)=48855410)
+17900 val loss 6.2096
+17900 val perplexity 497.5264
+17900 train 7.001251 (lr=7.4744e-05) (hash(x)=73004834)
+19700 val loss 6.1954
+19700 val perplexity 490.4861
+19700 train 6.463104 (lr=4.9024e-05) (hash(x)=57942870)
+19700 val loss 6.0211
+19700 val perplexity 412.0120
+19700 train 6.308620 (lr=3.5017e-05) (hash(x)=57942870)
+18000 val loss 6.1937
+18000 val perplexity 489.6782
+18000 train 6.391912 (lr=7.4488e-05) (hash(x)=51900245)
+19800 val loss 6.1841
+19800 val perplexity 484.9574
+19800 train 6.157583 (lr=4.8837e-05) (hash(x)=56663417)
+19800 val loss 6.0072
+19800 val perplexity 406.3322
+19800 train 5.984298 (lr=3.4883e-05) (hash(x)=56663417)
+18100 val loss 6.2003
+18100 val perplexity 492.9055
+18100 train 6.469229 (lr=7.4232e-05) (hash(x)=56278625)
+19900 val loss 6.1664
+19900 val perplexity 476.4777
+19900 train 6.036130 (lr=4.8649e-05) (hash(x)=41160884)
+19900 val loss 6.0034
+19900 val perplexity 404.8085
+19900 train 5.876379 (lr=3.4749e-05) (hash(x)=41160884)
+18200 val loss 6.1977
+18200 val perplexity 491.6086
+18200 train 6.040568 (lr=7.3975e-05) (hash(x)=50478164)
+20000 val loss 6.1748
+20000 val perplexity 480.4738
+20000 train 6.101958 (lr=4.8461e-05) (hash(x)=53380308)
+20000 val loss 6.0028
+20000 val perplexity 404.5619
+20000 train 5.943692 (lr=3.4615e-05) (hash(x)=53380308)
+18300 val loss 6.2111
+18300 val perplexity 498.2598
+18300 train 6.170772 (lr=7.3717e-05) (hash(x)=47837565)
+20100 val loss 6.1677
+20100 val perplexity 477.1066
+20100 train 6.002728 (lr=4.8272e-05) (hash(x)=46345322)
+20100 val loss 5.9981
+20100 val perplexity 402.6804
+20100 train 5.834086 (lr=3.4480e-05) (hash(x)=46345322)
+18400 val loss 6.2019
+18400 val perplexity 493.6668
+18400 train 6.035745 (lr=7.3459e-05) (hash(x)=49622704)
+20200 val loss 6.1732
+20200 val perplexity 479.7181
+20200 train 5.977310 (lr=4.8083e-05) (hash(x)=45030198)
+20200 val loss 6.0033
+20200 val perplexity 404.7531
+20200 train 5.833225 (lr=3.4345e-05) (hash(x)=45030198)
+18500 val loss 6.2105
+18500 val perplexity 497.9562
+18500 train 6.022415 (lr=7.3199e-05) (hash(x)=49741203)
+20300 val loss 6.1696
+20300 val perplexity 477.9764
+20300 train 5.636063 (lr=4.7893e-05) (hash(x)=36613892)
+20300 val loss 6.0073
+20300 val perplexity 406.3669
+20300 train 5.474130 (lr=3.4209e-05) (hash(x)=36613892)
+18600 val loss 6.2128
+18600 val perplexity 499.1172
+18600 train 6.630300 (lr=7.2939e-05) (hash(x)=54690634)
+20400 val loss 6.1657
+20400 val perplexity 476.1436
+20400 train 5.811815 (lr=4.7703e-05) (hash(x)=52987869)
+20400 val loss 6.0032
+20400 val perplexity 404.7357
+20400 train 5.619619 (lr=3.4074e-05) (hash(x)=52987869)
+18700 val loss 6.2068
+18700 val perplexity 496.0882
+18700 train 6.075917 (lr=7.2679e-05) (hash(x)=53778909)
+20500 val loss 6.1605
+20500 val perplexity 473.6513
+20500 train 6.055586 (lr=4.7513e-05) (hash(x)=51868861)
+20500 val loss 5.9988
+20500 val perplexity 402.9633
+20500 train 5.896246 (lr=3.3938e-05) (hash(x)=51868861)
+20600 val loss 6.1605
+20600 val perplexity 473.6872
+20600 train 6.005601 (lr=4.7323e-05) (hash(x)=48914729)
+18800 val loss 6.1777
+18800 val perplexity 481.8716
+18800 train 6.328458 (lr=7.2417e-05) (hash(x)=52196791)
+20600 val loss 6.0050
+20600 val perplexity 405.4350
+20600 train 5.827343 (lr=3.3802e-05) (hash(x)=48914729)
+20700 val loss 6.1518
+20700 val perplexity 469.5664
+20700 train 6.068443 (lr=4.7132e-05) (hash(x)=52260642)
+18900 val loss 6.1715
+18900 val perplexity 478.9191
+18900 train 6.102306 (lr=7.2155e-05) (hash(x)=49452227)
+20700 val loss 5.9881
+20700 val perplexity 398.6614
+20700 train 5.907519 (lr=3.3665e-05) (hash(x)=52260642)
+20800 val loss 6.1540
+20800 val perplexity 470.5742
+20800 train 6.611380 (lr=4.6940e-05) (hash(x)=67146535)
+19000 val loss 6.1746
+19000 val perplexity 480.4142
+19000 train 6.141204 (lr=7.1892e-05) (hash(x)=54220718)
+20800 val loss 5.9822
+20800 val perplexity 396.2974
+20800 train 6.462006 (lr=3.3529e-05) (hash(x)=67146535)
+20900 val loss 6.1529
+20900 val perplexity 470.0727
+20900 train 6.217502 (lr=4.6749e-05) (hash(x)=54553323)
+20900 val loss 5.9839
+20900 val perplexity 396.9801
+20900 train 6.038366 (lr=3.3392e-05) (hash(x)=54553323)
+19100 val loss 6.1787
+19100 val perplexity 482.3727
+19100 train 6.311283 (lr=7.1629e-05) (hash(x)=54899289)
+21000 val loss 6.1387
+21000 val perplexity 463.4713
+21000 train 6.117680 (lr=4.6557e-05) (hash(x)=48875330)
+21000 val loss 5.9773
+21000 val perplexity 394.3770
+21000 train 5.963885 (lr=3.3255e-05) (hash(x)=48875330)
+19200 val loss 6.1619
+19200 val perplexity 474.3357
+19200 train 6.279456 (lr=7.1365e-05) (hash(x)=52676298)
+21100 val loss 6.1434
+21100 val perplexity 465.6494
+21100 train 6.347168 (lr=4.6365e-05) (hash(x)=51341973)
+21100 val loss 5.9764
+21100 val perplexity 394.0028
+21100 train 6.203042 (lr=3.3118e-05) (hash(x)=51341973)
+19300 val loss 6.1691
+19300 val perplexity 477.7597
+19300 train 6.234082 (lr=7.1100e-05) (hash(x)=54877992)
+21200 val loss 6.1406
+21200 val perplexity 464.3428
+21200 train 6.099993 (lr=4.6172e-05) (hash(x)=49271158)
+21200 val loss 5.9711
+21200 val perplexity 391.9223
+21200 train 5.963935 (lr=3.2980e-05) (hash(x)=49271158)
+19400 val loss 6.1688
+19400 val perplexity 477.6094
+19400 train 6.003720 (lr=7.0835e-05) (hash(x)=46432347)
+21300 val loss 6.1358
+21300 val perplexity 462.1079
+21300 train 5.959718 (lr=4.5979e-05) (hash(x)=45591170)
+21300 val loss 5.9673
+21300 val perplexity 390.4627
+21300 train 5.794493 (lr=3.2842e-05) (hash(x)=45591170)
+19500 val loss 6.1838
+19500 val perplexity 484.8080
+19500 train 6.281179 (lr=7.0569e-05) (hash(x)=50677944)
+21400 val loss 6.1415
+21400 val perplexity 464.7537
+21400 train 6.203477 (lr=4.5786e-05) (hash(x)=53118930)
+21400 val loss 5.9793
+21400 val perplexity 395.1776
+21400 train 6.052376 (lr=3.2704e-05) (hash(x)=53118930)
+19600 val loss 6.1680
+19600 val perplexity 477.2160
+19600 train 6.011826 (lr=7.0302e-05) (hash(x)=48855410)
+21500 val loss 6.1434
+21500 val perplexity 465.6130
+21500 train 5.880425 (lr=4.5592e-05) (hash(x)=52415119)
+21500 val loss 5.9685
+21500 val perplexity 390.9249
+21500 train 5.735579 (lr=3.2566e-05) (hash(x)=52415119)
+21600 val loss 6.1402
+21600 val perplexity 464.1633
+21600 train 5.977185 (lr=4.5399e-05) (hash(x)=48585730)
+19700 val loss 6.1685
+19700 val perplexity 477.4916
+19700 train 6.435672 (lr=7.0035e-05) (hash(x)=57942870)
+21600 val loss 5.9629
+21600 val perplexity 388.7413
+21600 train 5.808125 (lr=3.2428e-05) (hash(x)=48585730)
+21700 val loss 6.1385
+21700 val perplexity 463.3798
+21700 train 6.320489 (lr=4.5205e-05) (hash(x)=61249800)
+19800 val loss 6.1834
+19800 val perplexity 484.6146
+19800 train 6.154061 (lr=6.9767e-05) (hash(x)=56663417)
+21700 val loss 5.9661
+21700 val perplexity 389.9913
+21700 train 6.156963 (lr=3.2289e-05) (hash(x)=61249800)
+21800 val loss 6.1459
+21800 val perplexity 466.7917
+21800 train 5.897917 (lr=4.5010e-05) (hash(x)=49264022)
+19900 val loss 6.1820
+19900 val perplexity 483.9657
+19900 train 6.063060 (lr=6.9498e-05) (hash(x)=41160884)
+21800 val loss 5.9556
+21800 val perplexity 385.9250
+21800 train 5.720652 (lr=3.2150e-05) (hash(x)=49264022)
+21900 val loss 6.1392
+21900 val perplexity 463.6750
+21900 train 6.063102 (lr=4.4816e-05) (hash(x)=47761044)
+20000 val loss 6.1799
+20000 val perplexity 482.9412
+20000 train 6.098162 (lr=6.9229e-05) (hash(x)=53380308)
+21900 val loss 5.9596
+21900 val perplexity 387.4464
+21900 train 5.900072 (lr=3.2011e-05) (hash(x)=47761044)
+22000 val loss 6.1364
+22000 val perplexity 462.3898
+22000 train 6.196565 (lr=4.4621e-05) (hash(x)=49243894)
+20100 val loss 6.1720
+20100 val perplexity 479.1443
+20100 train 6.015300 (lr=6.8960e-05) (hash(x)=46345322)
+22000 val loss 5.9533
+22000 val perplexity 385.0378
+22000 train 6.052825 (lr=3.1872e-05) (hash(x)=49243894)
+22100 val loss 6.1385
+22100 val perplexity 463.3583
+22100 train 5.857987 (lr=4.4426e-05) (hash(x)=45613618)
+22100 val loss 5.9643
+22100 val perplexity 389.2872
+22100 train 5.693257 (lr=3.1733e-05) (hash(x)=45613618)
+20200 val loss 6.1937
+20200 val perplexity 489.6474
+20200 train 5.996865 (lr=6.8690e-05) (hash(x)=45030198)
+22200 val loss 6.1277
+22200 val perplexity 458.3710
+22200 train 6.044831 (lr=4.4231e-05) (hash(x)=47671047)
+22200 val loss 5.9771
+22200 val perplexity 394.2805
+22200 train 5.903796 (lr=3.1593e-05) (hash(x)=47671047)
+20300 val loss 6.1750
+20300 val perplexity 480.5879
+20300 train 5.663015 (lr=6.8419e-05) (hash(x)=36613892)
+22300 val loss 6.1418
+22300 val perplexity 464.9053
+22300 train 6.115434 (lr=4.4035e-05) (hash(x)=55327350)
+22300 val loss 5.9599
+22300 val perplexity 387.5772
+22300 train 5.940328 (lr=3.1454e-05) (hash(x)=55327350)
+20400 val loss 6.1719
+20400 val perplexity 479.1105
+20400 train 5.803234 (lr=6.8148e-05) (hash(x)=52987869)
+22400 val loss 6.1288
+22400 val perplexity 458.8643
+22400 train 6.322518 (lr=4.3839e-05) (hash(x)=49864138)
+22400 val loss 5.9531
+22400 val perplexity 384.9440
+22400 train 6.157360 (lr=3.1314e-05) (hash(x)=49864138)
+20500 val loss 6.1687
+20500 val perplexity 477.5483
+20500 train 6.077594 (lr=6.7876e-05) (hash(x)=51868861)
+22500 val loss 6.1233
+22500 val perplexity 456.3487
+22500 train 6.059837 (lr=4.3643e-05) (hash(x)=48102324)
+22500 val loss 5.9441
+22500 val perplexity 381.5111
+22500 train 5.894808 (lr=3.1174e-05) (hash(x)=48102324)
+22600 val loss 6.1267
+22600 val perplexity 457.9022
+22600 train 6.135751 (lr=4.3447e-05) (hash(x)=49662723)
+20600 val loss 6.1625
+20600 val perplexity 474.6221
+20600 train 6.011073 (lr=6.7604e-05) (hash(x)=48914729)
+22600 val loss 5.9440
+22600 val perplexity 381.4529
+22600 train 6.013496 (lr=3.1034e-05) (hash(x)=49662723)
+22700 val loss 6.1144
+22700 val perplexity 452.3062
+22700 train 6.291257 (lr=4.3251e-05) (hash(x)=51372261)
+20700 val loss 6.1529
+20700 val perplexity 470.0969
+20700 train 6.074298 (lr=6.7331e-05) (hash(x)=52260642)
+22700 val loss 5.9382
+22700 val perplexity 379.2613
+22700 train 6.130818 (lr=3.0894e-05) (hash(x)=51372261)
+22800 val loss 6.1108
+22800 val perplexity 450.6853
+22800 train 6.071638 (lr=4.3054e-05) (hash(x)=48811743)
+20800 val loss 6.1672
+20800 val perplexity 476.8450
+20800 train 6.585125 (lr=6.7058e-05) (hash(x)=67146535)
+22800 val loss 5.9331
+22800 val perplexity 377.3160
+22800 train 5.895883 (lr=3.0753e-05) (hash(x)=48811743)
+22900 val loss 6.1070
+22900 val perplexity 448.9683
+22900 train 5.995105 (lr=4.2858e-05) (hash(x)=51623299)
+20900 val loss 6.1598
+20900 val perplexity 473.3420
+20900 train 6.234706 (lr=6.6784e-05) (hash(x)=54553323)
+22900 val loss 5.9320
+22900 val perplexity 376.9069
+22900 train 5.835777 (lr=3.0613e-05) (hash(x)=51623299)
+23000 val loss 6.1117
+23000 val perplexity 451.1217
+23000 train 6.000389 (lr=4.2661e-05) (hash(x)=47452756)
+21000 val loss 6.1511
+21000 val perplexity 469.2110
+21000 train 6.108371 (lr=6.6510e-05) (hash(x)=48875330)
+23000 val loss 5.9511
+23000 val perplexity 384.1771
+23000 train 5.855799 (lr=3.0472e-05) (hash(x)=47452756)
+23100 val loss 6.1143
+23100 val perplexity 452.2760
+23100 train 5.827789 (lr=4.2464e-05) (hash(x)=43735997)
+21100 val loss 6.1438
+21100 val perplexity 465.8320
+21100 train 6.351503 (lr=6.6235e-05) (hash(x)=51341973)
+23100 val loss 5.9353
+23100 val perplexity 378.1685
+23100 train 5.701593 (lr=3.0331e-05) (hash(x)=43735997)
+23200 val loss 6.1201
+23200 val perplexity 454.9291
+23200 train 5.997576 (lr=4.2267e-05) (hash(x)=49451282)
+21200 val loss 6.1496
+21200 val perplexity 468.5362
+21200 train 6.088956 (lr=6.5960e-05) (hash(x)=49271158)
+23200 val loss 5.9328
+23200 val perplexity 377.2018
+23200 train 5.856523 (lr=3.0190e-05) (hash(x)=49451282)
+23300 val loss 6.1108
+23300 val perplexity 450.7143
+23300 train 6.146278 (lr=4.2069e-05) (hash(x)=56682096)
+23300 val loss 5.9376
+23300 val perplexity 379.0405
+23300 train 5.990792 (lr=3.0049e-05) (hash(x)=56682096)
+21300 val loss 6.1550
+21300 val perplexity 471.0694
+21300 train 5.982675 (lr=6.5684e-05) (hash(x)=45591170)
+23400 val loss 6.1011
+23400 val perplexity 446.3608
+23400 train 5.769182 (lr=4.1872e-05) (hash(x)=44941532)
+23400 val loss 5.9304
+23400 val perplexity 376.3157
+23400 train 5.619764 (lr=2.9908e-05) (hash(x)=44941532)
+21400 val loss 6.1613
+21400 val perplexity 474.0571
+21400 train 6.212630 (lr=6.5408e-05) (hash(x)=53118930)
+23500 val loss 6.1009
+23500 val perplexity 446.2529
+23500 train 6.131494 (lr=4.1674e-05) (hash(x)=49488756)
+23500 val loss 5.9268
+23500 val perplexity 374.9562
+23500 train 5.927511 (lr=2.9767e-05) (hash(x)=49488756)
+23600 val loss 6.1050
+23600 val perplexity 448.1021
+23600 train 5.975768 (lr=4.1476e-05) (hash(x)=49425437)
+21500 val loss 6.1392
+21500 val perplexity 463.6640
+21500 train 5.844633 (lr=6.5132e-05) (hash(x)=52415119)
+23600 val loss 5.9334
+23600 val perplexity 377.4465
+23600 train 5.798845 (lr=2.9626e-05) (hash(x)=49425437)
+23700 val loss 6.1099
+23700 val perplexity 450.2922
+23700 train 6.504116 (lr=4.1278e-05) (hash(x)=56257944)
+21600 val loss 6.1489
+21600 val perplexity 468.1996
+21600 train 6.001342 (lr=6.4855e-05) (hash(x)=48585730)
+23700 val loss 5.9325
+23700 val perplexity 377.0872
+23700 train 6.344593 (lr=2.9485e-05) (hash(x)=56257944)
+23800 val loss 6.1124
+23800 val perplexity 451.4183
+23800 train 6.195869 (lr=4.1080e-05) (hash(x)=51424843)
+21700 val loss 6.1481
+21700 val perplexity 467.8303
+21700 train 6.325537 (lr=6.4578e-05) (hash(x)=61249800)
+23800 val loss 5.9284
+23800 val perplexity 375.5578
+23800 train 6.008825 (lr=2.9343e-05) (hash(x)=51424843)
+23900 val loss 6.0977
+23900 val perplexity 444.8233
+23900 train 6.149929 (lr=4.0882e-05) (hash(x)=51833590)
+21800 val loss 6.1408
+21800 val perplexity 464.4139
+21800 train 5.924423 (lr=6.4300e-05) (hash(x)=49264022)
+23900 val loss 5.9337
+23900 val perplexity 377.5376
+23900 train 6.000364 (lr=2.9202e-05) (hash(x)=51833590)
+24000 val loss 6.0954
+24000 val perplexity 443.8301
+24000 train 5.901597 (lr=4.0684e-05) (hash(x)=48412152)
+21900 val loss 6.1541
+21900 val perplexity 470.6493
+21900 train 6.081319 (lr=6.4023e-05) (hash(x)=47761044)
+24000 val loss 5.9247
+24000 val perplexity 374.1620
+24000 train 5.726462 (lr=2.9060e-05) (hash(x)=48412152)
+24100 val loss 6.1171
+24100 val perplexity 453.5444
+24100 train 6.002603 (lr=4.0486e-05) (hash(x)=48804835)
+22000 val loss 6.1653
+22000 val perplexity 475.9423
+22000 train 6.240574 (lr=6.3744e-05) (hash(x)=49243894)
+24100 val loss 5.9334
+24100 val perplexity 377.4503
+24100 train 5.848604 (lr=2.8918e-05) (hash(x)=48804835)
+24200 val loss 6.0871
+24200 val perplexity 440.1234
+24200 train 6.455055 (lr=4.0287e-05) (hash(x)=56459963)
+22100 val loss 6.1504
+22100 val perplexity 468.8929
+22100 train 5.862613 (lr=6.3466e-05) (hash(x)=45613618)
+24200 val loss 5.9242
+24200 val perplexity 373.9656
+24200 train 6.344940 (lr=2.8777e-05) (hash(x)=56459963)
+24300 val loss 6.0823
+24300 val perplexity 438.0208
+24300 train 5.985527 (lr=4.0089e-05) (hash(x)=46716240)
+24300 val loss 5.9214
+24300 val perplexity 372.9502
+24300 train 5.852433 (lr=2.8635e-05) (hash(x)=46716240)
+22200 val loss 6.1602
+22200 val perplexity 473.5102
+22200 train 6.061065 (lr=6.3187e-05) (hash(x)=47671047)
+24400 val loss 6.0867
+24400 val perplexity 439.9503
+24400 train 6.243721 (lr=3.9891e-05) (hash(x)=53022938)
+24400 val loss 5.9234
+24400 val perplexity 373.6973
+24400 train 6.087222 (lr=2.8493e-05) (hash(x)=53022938)
+24500 val loss 6.0981
+24500 val perplexity 445.0075
+24500 train 6.050263 (lr=3.9692e-05) (hash(x)=49422455)
+22300 val loss 6.1683
+22300 val perplexity 477.3819
+22300 train 6.139494 (lr=6.2907e-05) (hash(x)=55327350)
+24500 val loss 5.9191
+24500 val perplexity 372.0645
+24500 train 5.885635 (lr=2.8351e-05) (hash(x)=49422455)
+24600 val loss 6.0834
+24600 val perplexity 438.4985
+24600 train 6.347855 (lr=3.9493e-05) (hash(x)=46095419)
+22400 val loss 6.1460
+22400 val perplexity 466.8687
+22400 train 6.343654 (lr=6.2628e-05) (hash(x)=49864138)
+24600 val loss 5.9201
+24600 val perplexity 372.4554
+24600 train 6.198368 (lr=2.8210e-05) (hash(x)=46095419)
+24700 val loss 6.0889
+24700 val perplexity 440.9454
+24700 train 6.189737 (lr=3.9295e-05) (hash(x)=52504918)
+22500 val loss 6.1448
+22500 val perplexity 466.2971
+22500 train 6.051379 (lr=6.2348e-05) (hash(x)=48102324)
+24700 val loss 5.9203
+24700 val perplexity 372.5203
+24700 train 6.023808 (lr=2.8068e-05) (hash(x)=52504918)
+24800 val loss 6.0735
+24800 val perplexity 434.2162
+24800 train 6.016776 (lr=3.9096e-05) (hash(x)=53353568)
+22600 val loss 6.1366
+22600 val perplexity 462.4595
+22600 train 6.158983 (lr=6.2068e-05) (hash(x)=49662723)
+24800 val loss 5.9217
+24800 val perplexity 373.0393
+24800 train 5.857777 (lr=2.7926e-05) (hash(x)=53353568)
+24900 val loss 6.0814
+24900 val perplexity 437.6431
+24900 train 5.991043 (lr=3.8897e-05) (hash(x)=47247169)
+22700 val loss 6.1229
+22700 val perplexity 456.1812
+22700 train 6.288065 (lr=6.1787e-05) (hash(x)=51372261)
+24900 val loss 5.9119
+24900 val perplexity 369.3970
+24900 train 5.811796 (lr=2.7784e-05) (hash(x)=47247169)
+25000 val loss 6.0838
+25000 val perplexity 438.7051
+25000 train 5.958038 (lr=3.8699e-05) (hash(x)=50655375)
+22800 val loss 6.1326
+22800 val perplexity 460.6456
+22800 train 6.076201 (lr=6.1506e-05) (hash(x)=48811743)
+25000 val loss 5.9158
+25000 val perplexity 370.8477
+25000 train 5.768218 (lr=2.7642e-05) (hash(x)=50655375)
+25100 val loss 6.0756
+25100 val perplexity 435.0909
+25100 train 5.877374 (lr=3.8500e-05) (hash(x)=50341890)
+22900 val loss 6.1193
+22900 val perplexity 454.5405
+22900 train 6.036701 (lr=6.1225e-05) (hash(x)=51623299)
+25100 val loss 5.9135
+25100 val perplexity 370.0094
+25100 train 5.735393 (lr=2.7500e-05) (hash(x)=50341890)
+25200 val loss 6.0753
+25200 val perplexity 434.9726
+25200 train 6.088803 (lr=3.8301e-05) (hash(x)=52368607)
+23000 val loss 6.1442
+23000 val perplexity 466.0197
+23000 train 6.038305 (lr=6.0944e-05) (hash(x)=47452756)
+25200 val loss 5.9162
+25200 val perplexity 371.0170
+25200 train 5.956788 (lr=2.7358e-05) (hash(x)=52368607)
+25300 val loss 6.0728
+25300 val perplexity 433.8882
+25300 train 5.722450 (lr=3.8103e-05) (hash(x)=46898454)
+23100 val loss 6.1481
+23100 val perplexity 467.8258
+23100 train 5.882354 (lr=6.0663e-05) (hash(x)=43735997)
+25300 val loss 5.9056
+25300 val perplexity 367.0740
+25300 train 5.541602 (lr=2.7216e-05) (hash(x)=46898454)
+25400 val loss 6.0704
+25400 val perplexity 432.8719
+25400 train 5.998699 (lr=3.7904e-05) (hash(x)=50423971)
+25400 val loss 5.9091
+25400 val perplexity 368.3924
+25400 train 5.848545 (lr=2.7074e-05) (hash(x)=50423971)
+23200 val loss 6.1344
+23200 val perplexity 461.4429
+23200 train 6.015235 (lr=6.0381e-05) (hash(x)=49451282)
+25500 val loss 6.0667
+25500 val perplexity 431.2422
+25500 train 5.917558 (lr=3.7705e-05) (hash(x)=46878539)
+25500 val loss 5.9059
+25500 val perplexity 367.2151
+25500 train 5.774507 (lr=2.6932e-05) (hash(x)=46878539)
+25600 val loss 6.0724
+25600 val perplexity 433.7289
+25600 train 6.247941 (lr=3.7507e-05) (hash(x)=54616245)
+23300 val loss 6.1230
+23300 val perplexity 456.2203
+23300 train 6.182223 (lr=6.0099e-05) (hash(x)=56682096)
+25600 val loss 5.9152
+25600 val perplexity 370.6167
+25600 train 6.090686 (lr=2.6790e-05) (hash(x)=54616245)
+25700 val loss 6.0679
+25700 val perplexity 431.7663
+25700 train 5.807947 (lr=3.7308e-05) (hash(x)=47739146)
+23400 val loss 6.1256
+23400 val perplexity 457.4387
+23400 train 5.784054 (lr=5.9817e-05) (hash(x)=44941532)
+25700 val loss 5.9028
+25700 val perplexity 366.0637
+25700 train 5.666933 (lr=2.6649e-05) (hash(x)=47739146)
+25800 val loss 6.0653
+25800 val perplexity 430.6711
+25800 train 5.972924 (lr=3.7109e-05) (hash(x)=49016232)
+23500 val loss 6.1192
+23500 val perplexity 454.4848
+23500 train 6.172567 (lr=5.9534e-05) (hash(x)=49488756)
+25800 val loss 5.9072
+25800 val perplexity 367.6897
+25800 train 5.809902 (lr=2.6507e-05) (hash(x)=49016232)
+25900 val loss 6.0674
+25900 val perplexity 431.5759
+25900 train 5.965425 (lr=3.6911e-05) (hash(x)=49512666)
+23600 val loss 6.1288
+23600 val perplexity 458.8991
+23600 train 5.990823 (lr=5.9252e-05) (hash(x)=49425437)
+25900 val loss 5.9046
+25900 val perplexity 366.7052
+25900 train 5.809525 (lr=2.6365e-05) (hash(x)=49512666)
+26000 val loss 6.0660
+26000 val perplexity 430.9323
+26000 train 5.913073 (lr=3.6713e-05) (hash(x)=44586031)
+23700 val loss 6.1433
+23700 val perplexity 465.5749
+23700 train 6.546266 (lr=5.8969e-05) (hash(x)=56257944)
+26000 val loss 5.8986
+26000 val perplexity 364.5334
+26000 train 5.745259 (lr=2.6223e-05) (hash(x)=44586031)
+26100 val loss 6.0575
+26100 val perplexity 427.3129
+26100 train 6.091120 (lr=3.6514e-05) (hash(x)=47151445)
+23800 val loss 6.1362
+23800 val perplexity 462.3072
+23800 train 6.196434 (lr=5.8686e-05) (hash(x)=51424843)
+26100 val loss 5.8917
+26100 val perplexity 362.0284
+26100 train 5.956306 (lr=2.6082e-05) (hash(x)=47151445)
+26200 val loss 6.0507
+26200 val perplexity 424.4200
+26200 train 6.266461 (lr=3.6316e-05) (hash(x)=54800573)
+23900 val loss 6.1293
+23900 val perplexity 459.1035
+23900 train 6.186757 (lr=5.8403e-05) (hash(x)=51833590)
+26200 val loss 5.8865
+26200 val perplexity 360.1413
+26200 train 6.149418 (lr=2.5940e-05) (hash(x)=54800573)
+26300 val loss 6.0518
+26300 val perplexity 424.8778
+26300 train 5.914206 (lr=3.6118e-05) (hash(x)=51810733)
+24000 val loss 6.1270
+24000 val perplexity 458.0797
+24000 train 5.919013 (lr=5.8120e-05) (hash(x)=48412152)
+26300 val loss 5.8888
+26300 val perplexity 360.9731
+26300 train 5.753789 (lr=2.5798e-05) (hash(x)=51810733)
+26400 val loss 6.0481
+26400 val perplexity 423.3150
+26400 train 6.119697 (lr=3.5920e-05) (hash(x)=51482762)
+24100 val loss 6.1334
+24100 val perplexity 461.0130
+24100 train 6.042610 (lr=5.7837e-05) (hash(x)=48804835)
+26400 val loss 5.8869
+26400 val perplexity 360.2805
+26400 train 5.980966 (lr=2.5657e-05) (hash(x)=51482762)
+26500 val loss 6.0487
+26500 val perplexity 423.5792
+26500 train 5.941554 (lr=3.5722e-05) (hash(x)=50129702)
+26500 val loss 5.8890
+26500 val perplexity 361.0482
+26500 train 5.805998 (lr=2.5515e-05) (hash(x)=50129702)
+24200 val loss 6.1302
+24200 val perplexity 459.5219
+24200 train 6.511795 (lr=5.7554e-05) (hash(x)=56459963)
+26600 val loss 6.0358
+26600 val perplexity 418.1230
+26600 train 6.005397 (lr=3.5524e-05) (hash(x)=51104302)
+26600 val loss 5.8765
+26600 val perplexity 356.5668
+26600 train 5.861450 (lr=2.5374e-05) (hash(x)=51104302)
+26700 val loss 6.0383
+26700 val perplexity 419.1601
+26700 train 5.877936 (lr=3.5326e-05) (hash(x)=48490782)
+24300 val loss 6.1126
+24300 val perplexity 451.5018
+24300 train 6.014219 (lr=5.7270e-05) (hash(x)=46716240)
+26700 val loss 5.8900
+26700 val perplexity 361.4176
+26700 train 5.728050 (lr=2.5233e-05) (hash(x)=48490782)
+26800 val loss 6.0646
+26800 val perplexity 430.3326
+26800 train 6.085222 (lr=3.5128e-05) (hash(x)=50602590)
+24400 val loss 6.1224
+24400 val perplexity 455.9637
+24400 train 6.260394 (lr=5.6987e-05) (hash(x)=53022938)
+26800 val loss 5.8807
+26800 val perplexity 358.0682
+26800 train 5.918839 (lr=2.5092e-05) (hash(x)=50602590)
+26900 val loss 6.0447
+26900 val perplexity 421.8910
+26900 train 5.809762 (lr=3.4931e-05) (hash(x)=45624407)
+24500 val loss 6.1045
+24500 val perplexity 447.8726
+24500 train 6.065207 (lr=5.6703e-05) (hash(x)=49422455)
+26900 val loss 5.8860
+26900 val perplexity 359.9544
+26900 train 5.676947 (lr=2.4951e-05) (hash(x)=45624407)
+27000 val loss 6.0349
+27000 val perplexity 417.7579
+27000 train 5.814979 (lr=3.4733e-05) (hash(x)=46448811)
+24600 val loss 6.1070
+24600 val perplexity 448.9721
+24600 train 6.393726 (lr=5.6419e-05) (hash(x)=46095419)
+27000 val loss 5.8797
+27000 val perplexity 357.6958
+27000 train 5.645614 (lr=2.4810e-05) (hash(x)=46448811)
+27100 val loss 6.0353
+27100 val perplexity 417.9307
+27100 train 5.723890 (lr=3.4536e-05) (hash(x)=42263321)
+24700 val loss 6.1352
+24700 val perplexity 461.8288
+24700 train 6.243340 (lr=5.6135e-05) (hash(x)=52504918)
+27100 val loss 5.8818
+27100 val perplexity 358.4509
+27100 train 5.567906 (lr=2.4669e-05) (hash(x)=42263321)
+27200 val loss 6.0267
+27200 val perplexity 414.3543
+27200 train 5.776744 (lr=3.4339e-05) (hash(x)=45998703)
+24800 val loss 6.0932
+24800 val perplexity 442.8545
+24800 train 6.026612 (lr=5.5852e-05) (hash(x)=53353568)
+27200 val loss 5.8721
+27200 val perplexity 354.9858
+27200 train 5.627224 (lr=2.4528e-05) (hash(x)=45998703)
+27300 val loss 6.0532
+27300 val perplexity 425.4879
+27300 train 6.213803 (lr=3.4142e-05) (hash(x)=58008207)
+24900 val loss 6.0928
+24900 val perplexity 442.6459
+24900 train 6.006665 (lr=5.5568e-05) (hash(x)=47247169)
+27300 val loss 5.8762
+27300 val perplexity 356.4554
+27300 train 5.960763 (lr=2.4387e-05) (hash(x)=58008207)
+27400 val loss 6.0430
+27400 val perplexity 421.1720
+27400 train 6.103308 (lr=3.3946e-05) (hash(x)=58914934)
+25000 val loss 6.1069
+25000 val perplexity 448.9342
+25000 train 5.982193 (lr=5.5284e-05) (hash(x)=50655375)
+27400 val loss 5.8746
+27400 val perplexity 355.8654
+27400 train 5.926643 (lr=2.4247e-05) (hash(x)=58914934)
+27500 val loss 6.0363
+27500 val perplexity 418.3502
+27500 train 5.996116 (lr=3.3749e-05) (hash(x)=54961792)
+25100 val loss 6.0932
+25100 val perplexity 442.8452
+27500 val loss 5.8769
+27500 val perplexity 356.7050
+25100 train 5.905435 (lr=5.5000e-05) (hash(x)=50341890)
+27600 val loss 6.0305
+27600 val perplexity 415.9225
+27600 train 5.722549 (lr=3.3553e-05) (hash(x)=43436743)
+27500 train 5.835189 (lr=2.4106e-05) (hash(x)=54961792)
+27700 val loss 6.0270
+27700 val perplexity 414.4666
+27700 train 5.629759 (lr=3.3357e-05) (hash(x)=43954091)
+27600 val loss 5.8695
+27600 val perplexity 354.0824
+27600 train 5.544859 (lr=2.3966e-05) (hash(x)=43436743)
+25200 val loss 6.0907
+25200 val perplexity 441.7247
+25200 train 6.110599 (lr=5.4716e-05) (hash(x)=52368607)
+27700 val loss 5.8784
+27700 val perplexity 357.2492
+27700 train 5.473822 (lr=2.3826e-05) (hash(x)=43954091)
+27800 val loss 6.0305
+27800 val perplexity 415.9158
+27800 train 5.947229 (lr=3.3161e-05) (hash(x)=49556041)
+25300 val loss 6.0847
+25300 val perplexity 439.0726
+25300 train 5.709676 (lr=5.4432e-05) (hash(x)=46898454)
+27800 val loss 5.8703
+27800 val perplexity 354.3638
+27800 train 5.806100 (lr=2.3686e-05) (hash(x)=49556041)
+27900 val loss 6.0227
+27900 val perplexity 412.6724
+27900 train 5.838119 (lr=3.2965e-05) (hash(x)=44542698)
+25400 val loss 6.0930
+25400 val perplexity 442.7607
+25400 train 6.024213 (lr=5.4148e-05) (hash(x)=50423971)
+27900 val loss 5.8620
+27900 val perplexity 351.4414
+27900 train 5.659078 (lr=2.3546e-05) (hash(x)=44542698)
+28000 val loss 6.0297
+28000 val perplexity 415.5857
+28000 train 5.976749 (lr=3.2769e-05) (hash(x)=51737822)
+25500 val loss 6.0894
+25500 val perplexity 441.1532
+25500 train 5.953987 (lr=5.3865e-05) (hash(x)=46878539)
+28000 val loss 5.8627
+28000 val perplexity 351.6655
+28000 train 5.821701 (lr=2.3407e-05) (hash(x)=51737822)
+28100 val loss 6.0081
+28100 val perplexity 406.7294
+28100 train 6.438169 (lr=3.2574e-05) (hash(x)=69830597)
+25600 val loss 6.0876
+25600 val perplexity 440.3661
+25600 train 6.259648 (lr=5.3581e-05) (hash(x)=54616245)
+28100 val loss 5.8625
+28100 val perplexity 351.5973
+28100 train 6.329861 (lr=2.3267e-05) (hash(x)=69830597)
+28200 val loss 6.0045
+28200 val perplexity 405.2684
+28200 train 6.060979 (lr=3.2379e-05) (hash(x)=51670601)
+25700 val loss 6.0855
+25700 val perplexity 439.4538
+25700 train 5.831913 (lr=5.3297e-05) (hash(x)=47739146)
+28200 val loss 5.8490
+28200 val perplexity 346.9010
+28200 train 5.932371 (lr=2.3128e-05) (hash(x)=51670601)
+28300 val loss 6.0017
+28300 val perplexity 404.0967
+28300 train 6.050955 (lr=3.2184e-05) (hash(x)=51628878)
+25800 val loss 6.0868
+25800 val perplexity 439.9898
+25800 train 5.981336 (lr=5.3013e-05) (hash(x)=49016232)
+28300 val loss 5.8536
+28300 val perplexity 348.4960
+28300 train 5.921436 (lr=2.2989e-05) (hash(x)=51628878)
+28400 val loss 6.0027
+28400 val perplexity 404.5355
+28400 train 6.015206 (lr=3.1990e-05) (hash(x)=50973565)
+28400 val loss 5.8501
+28400 val perplexity 347.2567
+28400 train 5.864284 (lr=2.2850e-05) (hash(x)=50973565)
+25900 val loss 6.0921
+25900 val perplexity 442.3286
+25900 train 5.988946 (lr=5.2730e-05) (hash(x)=49512666)
+28500 val loss 6.0014
+28500 val perplexity 403.9896
+28500 train 6.132781 (lr=3.1795e-05) (hash(x)=63914328)
+28500 val loss 5.8495
+28500 val perplexity 347.0701
+28500 train 5.978743 (lr=2.2711e-05) (hash(x)=63914328)
+28600 val loss 6.0003
+28600 val perplexity 403.5666
+28600 train 6.013580 (lr=3.1601e-05) (hash(x)=56517839)
+26000 val loss 6.0844
+26000 val perplexity 438.9579
+26000 train 5.939546 (lr=5.2446e-05) (hash(x)=44586031)
+28600 val loss 5.8411
+28600 val perplexity 344.1504
+28600 train 5.840915 (lr=2.2572e-05) (hash(x)=56517839)
+28700 val loss 6.0003
+28700 val perplexity 403.5629
+28700 train 6.015846 (lr=3.1408e-05) (hash(x)=47631048)
+26100 val loss 6.0682
+26100 val perplexity 431.9187
+26100 train 6.100973 (lr=5.2163e-05) (hash(x)=47151445)
+28700 val loss 5.8446
+28700 val perplexity 345.3616
+28700 train 5.873164 (lr=2.2434e-05) (hash(x)=47631048)
+28800 val loss 6.0039
+28800 val perplexity 404.9857
+28800 train 5.804829 (lr=3.1214e-05) (hash(x)=44376567)
+26200 val loss 6.0777
+26200 val perplexity 436.0070
+26200 train 6.290683 (lr=5.1880e-05) (hash(x)=54800573)
+28800 val loss 5.8432
+28800 val perplexity 344.8926
+28800 train 5.650768 (lr=2.2296e-05) (hash(x)=44376567)
+28900 val loss 6.0049
+28900 val perplexity 405.3988
+28900 train 5.946053 (lr=3.1021e-05) (hash(x)=50934351)
+26300 val loss 6.0655
+26300 val perplexity 430.7242
+26300 train 5.922496 (lr=5.1597e-05) (hash(x)=51810733)
+28900 val loss 5.8443
+28900 val perplexity 345.2609
+28900 train 5.810985 (lr=2.2158e-05) (hash(x)=50934351)
+29000 val loss 6.0022
+29000 val perplexity 404.3154
+29000 train 6.185178 (lr=3.0828e-05) (hash(x)=51712329)
+26400 val loss 6.0617
+26400 val perplexity 429.0950
+26400 train 6.139353 (lr=5.1314e-05) (hash(x)=51482762)
+29000 val loss 5.8386
+29000 val perplexity 343.2891
+29000 train 6.037989 (lr=2.2020e-05) (hash(x)=51712329)
+29100 val loss 5.9990
+29100 val perplexity 403.0212
+29100 train 5.958769 (lr=3.0635e-05) (hash(x)=51282586)
+26500 val loss 6.0659
+26500 val perplexity 430.8895
+26500 train 5.955387 (lr=5.1031e-05) (hash(x)=50129702)
+29100 val loss 5.8463
+29100 val perplexity 345.9400
+29100 train 5.830448 (lr=2.1882e-05) (hash(x)=51282586)
+29200 val loss 5.9984
+29200 val perplexity 402.7780
+29200 train 5.718521 (lr=3.0443e-05) (hash(x)=44457224)
+26600 val loss 6.0599
+26600 val perplexity 428.3233
+26600 train 6.015794 (lr=5.0748e-05) (hash(x)=51104302)
+29200 val loss 5.8406
+29200 val perplexity 344.0006
+29200 train 5.555852 (lr=2.1745e-05) (hash(x)=44457224)
+29300 val loss 5.9981
+29300 val perplexity 402.6825
+29300 train 5.917011 (lr=3.0251e-05) (hash(x)=47669361)
+26700 val loss 6.0709
+26700 val perplexity 433.0734
+26700 train 5.911325 (lr=5.0466e-05) (hash(x)=48490782)
+29300 val loss 5.8387
+29300 val perplexity 343.3423
+29300 train 5.772800 (lr=2.1608e-05) (hash(x)=47669361)
+29400 val loss 5.9899
+29400 val perplexity 399.3928
+29400 train 6.062664 (lr=3.0060e-05) (hash(x)=49442135)
+26800 val loss 6.0654
+26800 val perplexity 430.7034
+26800 train 6.097916 (lr=5.0183e-05) (hash(x)=50602590)
+29400 val loss 5.8396
+29400 val perplexity 343.6570
+29400 train 5.912821 (lr=2.1471e-05) (hash(x)=49442135)
+29500 val loss 5.9921
+29500 val perplexity 400.2511
+29500 train 5.490143 (lr=2.9868e-05) (hash(x)=39350955)
+26900 val loss 6.0564
+26900 val perplexity 426.8437
+26900 train 5.828437 (lr=4.9901e-05) (hash(x)=45624407)
+29500 val loss 5.8454
+29500 val perplexity 345.6369
+29500 train 5.381187 (lr=2.1335e-05) (hash(x)=39350955)
+29600 val loss 5.9898
+29600 val perplexity 399.3499
+29600 train 6.467952 (lr=2.9677e-05) (hash(x)=53114560)
+27000 val loss 6.0637
+27000 val perplexity 429.9518
+27000 train 5.826560 (lr=4.9619e-05) (hash(x)=46448811)
+29600 val loss 5.8388
+29600 val perplexity 343.3613
+29700 val loss 5.9815
+29700 val perplexity 396.0278
+29600 train 6.333019 (lr=2.1198e-05) (hash(x)=53114560)
+29700 train 5.881025 (lr=2.9487e-05) (hash(x)=45302862)
+27100 val loss 6.0680
+27100 val perplexity 431.8003
+27100 train 5.762912 (lr=4.9337e-05) (hash(x)=42263321)
+29700 val loss 5.8300
+29700 val perplexity 340.3463
+29700 train 5.736992 (lr=2.1062e-05) (hash(x)=45302862)
+29800 val loss 5.9815
+29800 val perplexity 396.0533
+29800 train 5.968771 (lr=2.9297e-05) (hash(x)=51679878)
+27200 val loss 6.0582
+27200 val perplexity 427.5987
+27200 train 5.819321 (lr=4.9056e-05) (hash(x)=45998703)
+29800 val loss 5.8299
+29800 val perplexity 340.3369
+29800 train 5.763031 (lr=2.0926e-05) (hash(x)=51679878)
+29900 val loss 5.9843
+29900 val perplexity 397.1570
+29900 train 5.798481 (lr=2.9107e-05) (hash(x)=47344663)
+27300 val loss 6.0673
+27300 val perplexity 431.5078
+27300 train 6.224754 (lr=4.8775e-05) (hash(x)=58008207)
+29900 val loss 5.8316
+29900 val perplexity 340.9067
+29900 train 5.655300 (lr=2.0791e-05) (hash(x)=47344663)
+30000 val loss 5.9842
+30000 val perplexity 397.1127
+30000 train 5.728186 (lr=2.8917e-05) (hash(x)=46530414)
+27400 val loss 6.0720
+27400 val perplexity 433.5327
+27400 train 6.118867 (lr=4.8494e-05) (hash(x)=58914934)
+30000 val loss 5.8346
+30000 val perplexity 341.9292
+30000 train 5.589412 (lr=2.0655e-05) (hash(x)=46530414)
+30100 val loss 5.9845
+30100 val perplexity 397.2377
+30100 train 5.554594 (lr=2.8728e-05) (hash(x)=43262362)
+27500 val loss 6.0554
+27500 val perplexity 426.4100
+27500 train 5.994745 (lr=4.8213e-05) (hash(x)=54961792)
+30100 val loss 5.8322
+30100 val perplexity 341.1007
+30100 train 5.392095 (lr=2.0520e-05) (hash(x)=43262362)
+30200 val loss 5.9802
+30200 val perplexity 395.5237
+30200 train 5.816063 (lr=2.8539e-05) (hash(x)=47879732)
+30200 val loss 5.8273
+30200 val perplexity 339.4288
+30200 train 5.661341 (lr=2.0385e-05) (hash(x)=47879732)
+27600 val loss 6.0559
+27600 val perplexity 426.6425
+27600 train 5.739508 (lr=4.7932e-05) (hash(x)=43436743)
+30300 val loss 5.9771
+30300 val perplexity 394.2851
+30300 train 5.769890 (lr=2.8351e-05) (hash(x)=47335687)
+30300 val loss 5.8292
+30300 val perplexity 340.0746
+30300 train 5.628904 (lr=2.0251e-05) (hash(x)=47335687)
+27700 val loss 6.0499
+27700 val perplexity 424.0787
+27700 train 5.665303 (lr=4.7652e-05) (hash(x)=43954091)
+30400 val loss 5.9810
+30400 val perplexity 395.8262
+30400 train 5.997238 (lr=2.8163e-05) (hash(x)=49940701)
+30400 val loss 5.8293
+30400 val perplexity 340.1254
+30400 train 5.867912 (lr=2.0117e-05) (hash(x)=49940701)
+30500 val loss 5.9804
+30500 val perplexity 395.6046
+30500 train 5.857673 (lr=2.7976e-05) (hash(x)=49860489)
+27800 val loss 6.0467
+27800 val perplexity 422.7273
+27800 train 5.969273 (lr=4.7372e-05) (hash(x)=49556041)
+30500 val loss 5.8309
+30500 val perplexity 340.6524
+30500 train 5.724407 (lr=1.9983e-05) (hash(x)=49860489)
+30600 val loss 5.9994
+30600 val perplexity 403.2059
+30600 train 5.748838 (lr=2.7789e-05) (hash(x)=44643399)
+27900 val loss 6.0436
+27900 val perplexity 421.3867
+27900 train 5.862562 (lr=4.7093e-05) (hash(x)=44542698)
+30600 val loss 5.8379
+30600 val perplexity 343.0716
+30600 train 5.591486 (lr=1.9849e-05) (hash(x)=44643399)
+30700 val loss 5.9809
+30700 val perplexity 395.7924
+30700 train 5.793119 (lr=2.7602e-05) (hash(x)=49388442)
+28000 val loss 6.0452
+28000 val perplexity 422.0614
+28000 train 5.988935 (lr=4.6813e-05) (hash(x)=51737822)
+30800 val loss 5.9740
+30800 val perplexity 393.0896
+30800 train 5.937857 (lr=2.7416e-05) (hash(x)=56160596)
+30700 val loss 5.8319
+30700 val perplexity 341.0164
+30700 train 5.644805 (lr=1.9716e-05) (hash(x)=49388442)
+28100 val loss 6.0408
+28100 val perplexity 420.2293
+28100 train 6.431351 (lr=4.6534e-05) (hash(x)=69830597)
+30800 val loss 5.8289
+30800 val perplexity 339.9865
+30800 train 5.815920 (lr=1.9583e-05) (hash(x)=56160596)
+30900 val loss 5.9717
+30900 val perplexity 392.1562
+30900 train 5.930646 (lr=2.7230e-05) (hash(x)=49005636)
+28200 val loss 6.0367
+28200 val perplexity 418.5054
+28200 train 6.090349 (lr=4.6256e-05) (hash(x)=51670601)
+30900 val loss 5.8255
+30900 val perplexity 338.8249
+30900 train 5.797294 (lr=1.9450e-05) (hash(x)=49005636)
+31000 val loss 5.9754
+31000 val perplexity 393.6448
+31000 train 6.203425 (lr=2.7045e-05) (hash(x)=54646552)
+28300 val loss 6.0271
+28300 val perplexity 414.5247
+28300 train 6.049088 (lr=4.5977e-05) (hash(x)=51628878)
+31000 val loss 5.8215
+31000 val perplexity 337.4719
+31000 train 6.032523 (lr=1.9318e-05) (hash(x)=54646552)
+31100 val loss 5.9654
+31100 val perplexity 389.6898
+31100 train 5.804157 (lr=2.6860e-05) (hash(x)=44816272)
+28400 val loss 6.0472
+28400 val perplexity 422.9353
+28400 train 6.074213 (lr=4.5700e-05) (hash(x)=50973565)
+31100 val loss 5.8205
+31100 val perplexity 337.1508
+31100 train 5.646246 (lr=1.9186e-05) (hash(x)=44816272)
+31200 val loss 5.9641
+31200 val perplexity 389.2166
+31200 train 6.128108 (lr=2.6675e-05) (hash(x)=53040860)
+28500 val loss 6.0329
+28500 val perplexity 416.9289
+28500 train 6.163313 (lr=4.5422e-05) (hash(x)=63914328)
+31200 val loss 5.8220
+31200 val perplexity 337.6446
+31200 train 6.022907 (lr=1.9054e-05) (hash(x)=53040860)
+31300 val loss 5.9687
+31300 val perplexity 390.9788
+31300 train 6.174263 (lr=2.6491e-05) (hash(x)=52051764)
+28600 val loss 6.0286
+28600 val perplexity 415.1244
+28600 train 6.037053 (lr=4.5145e-05) (hash(x)=56517839)
+31300 val loss 5.8178
+31300 val perplexity 336.2187
+31300 train 6.046365 (lr=1.8922e-05) (hash(x)=52051764)
+31400 val loss 5.9691
+31400 val perplexity 391.1619
+31400 train 5.940002 (lr=2.6308e-05) (hash(x)=51564589)
+28700 val loss 6.0258
+28700 val perplexity 413.9781
+28700 train 6.036065 (lr=4.4868e-05) (hash(x)=47631048)
+31400 val loss 5.8139
+31400 val perplexity 334.9343
+31400 train 5.756424 (lr=1.8791e-05) (hash(x)=51564589)
+31500 val loss 5.9683
+31500 val perplexity 390.8475
+31500 train 5.699306 (lr=2.6125e-05) (hash(x)=49676250)
+28800 val loss 6.0292
+28800 val perplexity 415.3858
+28800 train 5.843014 (lr=4.4592e-05) (hash(x)=44376567)
+31500 val loss 5.8101
+31500 val perplexity 333.6571
+31500 train 5.550847 (lr=1.8661e-05) (hash(x)=49676250)
+31600 val loss 5.9691
+31600 val perplexity 391.1578
+31600 train 5.896336 (lr=2.5942e-05) (hash(x)=49244964)
+31600 val loss 5.8081
+31600 val perplexity 332.9929
+31600 train 5.752309 (lr=1.8530e-05) (hash(x)=49244964)
+28900 val loss 6.0276
+28900 val perplexity 414.7323
+28900 train 5.979816 (lr=4.4316e-05) (hash(x)=50934351)
+31700 val loss 5.9622
+31700 val perplexity 388.4520
+31700 train 6.168909 (lr=2.5760e-05) (hash(x)=49527873)
+31700 val loss 5.8098
+31700 val perplexity 333.5623
+31700 train 6.045442 (lr=1.8400e-05) (hash(x)=49527873)
+31800 val loss 5.9669
+31800 val perplexity 390.3124
+31800 train 5.923289 (lr=2.5579e-05) (hash(x)=49566755)
+29000 val loss 6.0243
+29000 val perplexity 413.3477
+29000 train 6.204475 (lr=4.4040e-05) (hash(x)=51712329)
+31800 val loss 5.8218
+31800 val perplexity 337.5768
+31800 train 5.783108 (lr=1.8271e-05) (hash(x)=49566755)
+31900 val loss 5.9640
+31900 val perplexity 389.1502
+31900 train 5.772141 (lr=2.5398e-05) (hash(x)=45456170)
+29100 val loss 6.0291
+29100 val perplexity 415.3400
+29100 train 5.993180 (lr=4.3765e-05) (hash(x)=51282586)
+31900 val loss 5.8053
+31900 val perplexity 332.0555
+31900 train 5.614435 (lr=1.8141e-05) (hash(x)=45456170)
+32000 val loss 5.9573
+32000 val perplexity 386.5804
+32000 train 6.021672 (lr=2.5218e-05) (hash(x)=49986978)
+29200 val loss 6.0389
+29200 val perplexity 419.4180
+29200 train 5.761024 (lr=4.3490e-05) (hash(x)=44457224)
+32000 val loss 5.8044
+32000 val perplexity 331.7572
+32100 val loss 5.9577
+32100 val perplexity 386.7297
+32000 train 5.883710 (lr=1.8013e-05) (hash(x)=49986978)
+32100 train 5.598077 (lr=2.5038e-05) (hash(x)=50539149)
+29300 val loss 6.0307
+29300 val perplexity 415.9908
+29300 train 5.962490 (lr=4.3216e-05) (hash(x)=47669361)
+32200 val loss 5.9584
+32200 val perplexity 386.9891
+32200 train 5.904149 (lr=2.4858e-05) (hash(x)=50998615)
+32100 val loss 5.8139
+32100 val perplexity 334.9378
+32100 train 5.460774 (lr=1.7884e-05) (hash(x)=50539149)
+29400 val loss 6.0301
+29400 val perplexity 415.7476
+29400 train 6.108774 (lr=4.2942e-05) (hash(x)=49442135)
+32300 val loss 5.9550
+32300 val perplexity 385.6723
+32300 train 5.784231 (lr=2.4679e-05) (hash(x)=44030191)
+32200 val loss 5.8095
+32200 val perplexity 333.4435
+32200 train 5.759093 (lr=1.7756e-05) (hash(x)=50998615)
+29500 val loss 6.0441
+29500 val perplexity 421.6283
+29500 train 5.537832 (lr=4.2669e-05) (hash(x)=39350955)
+32400 val loss 5.9621
+32400 val perplexity 388.4196
+32400 train 5.888499 (lr=2.4501e-05) (hash(x)=47269028)
+32300 val loss 5.8086
+32300 val perplexity 333.1536
+32300 train 5.670200 (lr=1.7628e-05) (hash(x)=44030191)
+29600 val loss 6.0214
+29600 val perplexity 412.1352
+29600 train 6.494508 (lr=4.2396e-05) (hash(x)=53114560)
+32500 val loss 5.9565
+32500 val perplexity 386.2638
+32500 train 5.777795 (lr=2.4323e-05) (hash(x)=50128289)
+32400 val loss 5.8033
+32400 val perplexity 331.4057
+32400 train 5.730560 (lr=1.7501e-05) (hash(x)=47269028)
+29700 val loss 6.0202
+29700 val perplexity 411.6758
+29700 train 5.940580 (lr=4.2124e-05) (hash(x)=45302862)
+32600 val loss 5.9642
+32600 val perplexity 389.2467
+32600 train 5.801872 (lr=2.4146e-05) (hash(x)=42781351)
+32500 val loss 5.8047
+32500 val perplexity 331.8452
+32500 train 5.641759 (lr=1.7374e-05) (hash(x)=50128289)
+29800 val loss 6.0158
+29800 val perplexity 409.8520
+29800 train 6.001534 (lr=4.1852e-05) (hash(x)=51679878)
+32700 val loss 5.9590
+32700 val perplexity 387.2246
+32700 train 6.055271 (lr=2.3970e-05) (hash(x)=51712117)
+32600 val loss 5.8048
+32600 val perplexity 331.8872
+32600 train 5.621361 (lr=1.7247e-05) (hash(x)=42781351)
+29900 val loss 6.0177
+29900 val perplexity 410.6227
+29900 train 5.824810 (lr=4.1581e-05) (hash(x)=47344663)
+32800 val loss 5.9455
+32800 val perplexity 382.0401
+32800 train 6.030862 (lr=2.3794e-05) (hash(x)=52478104)
+32700 val loss 5.8127
+32700 val perplexity 334.5206
+32700 train 5.901848 (lr=1.7121e-05) (hash(x)=51712117)
+30000 val loss 6.0296
+30000 val perplexity 415.5389
+30000 train 5.787665 (lr=4.1310e-05) (hash(x)=46530414)
+32800 val loss 5.7982
+32800 val perplexity 329.7212
+32800 train 5.903448 (lr=1.6995e-05) (hash(x)=52478104)
+32900 val loss 5.9386
+32900 val perplexity 379.4034
+32900 train 5.884224 (lr=2.3618e-05) (hash(x)=53588694)
+32900 val loss 5.7975
+32900 val perplexity 329.4850
+32900 train 5.737524 (lr=1.6870e-05) (hash(x)=53588694)
+33000 val loss 5.9344
+33000 val perplexity 377.8080
+33000 train 5.818412 (lr=2.3443e-05) (hash(x)=50821436)
+30100 val loss 6.0201
+30100 val perplexity 411.6204
+30100 train 5.600601 (lr=4.1040e-05) (hash(x)=43262362)
+33000 val loss 5.7923
+33000 val perplexity 327.7762
+33000 train 5.669199 (lr=1.6745e-05) (hash(x)=50821436)
+33100 val loss 5.9394
+33100 val perplexity 379.7194
+33100 train 5.892412 (lr=2.3269e-05) (hash(x)=50371199)
+30200 val loss 6.0249
+30200 val perplexity 413.5985
+30200 train 5.841463 (lr=4.0771e-05) (hash(x)=47879732)
+33100 val loss 5.7909
+33100 val perplexity 327.3000
+33100 train 5.743008 (lr=1.6621e-05) (hash(x)=50371199)
+33200 val loss 5.9406
+33200 val perplexity 380.1528
+33200 train 6.119138 (lr=2.3095e-05) (hash(x)=54544623)
+30300 val loss 6.0242
+30300 val perplexity 413.3002
+30300 train 5.807525 (lr=4.0502e-05) (hash(x)=47335687)
+33200 val loss 5.7917
+33200 val perplexity 327.5760
+33300 val loss 5.9384
+33300 val perplexity 379.3323
+33200 train 6.008244 (lr=1.6497e-05) (hash(x)=54544623)
+33300 train 5.717372 (lr=2.2922e-05) (hash(x)=55456921)
+30400 val loss 6.0205
+30400 val perplexity 411.7994
+30400 train 6.041892 (lr=4.0233e-05) (hash(x)=49940701)
+33300 val loss 5.7894
+33300 val perplexity 326.8158
+33300 train 5.569286 (lr=1.6373e-05) (hash(x)=55456921)
+33400 val loss 5.9448
+33400 val perplexity 381.7524
+33400 train 5.583863 (lr=2.2750e-05) (hash(x)=37240412)
+30500 val loss 6.0229
+30500 val perplexity 412.7765
+30500 train 5.896065 (lr=3.9965e-05) (hash(x)=49860489)
+33500 val loss 5.9436
+33500 val perplexity 381.3225
+33500 train 6.196152 (lr=2.2578e-05) (hash(x)=56551894)
+33400 val loss 5.7941
+33400 val perplexity 328.3684
+33400 train 5.386881 (lr=1.6250e-05) (hash(x)=37240412)
+30600 val loss 6.0311
+30600 val perplexity 416.1628
+30600 train 5.800098 (lr=3.9698e-05) (hash(x)=44643399)
+33600 val loss 5.9390
+33600 val perplexity 379.5388
+33600 train 5.841659 (lr=2.2407e-05) (hash(x)=47514225)
+33500 val loss 5.7895
+33500 val perplexity 326.8537
+33500 train 6.019800 (lr=1.6127e-05) (hash(x)=56551894)
+30700 val loss 6.0243
+30700 val perplexity 413.3524
+30700 train 5.884185 (lr=3.9431e-05) (hash(x)=49388442)
+33700 val loss 5.9422
+33700 val perplexity 380.7785
+33700 train 5.784485 (lr=2.2237e-05) (hash(x)=51164416)
+33600 val loss 5.7885
+33600 val perplexity 326.5360
+33600 train 5.714428 (lr=1.6005e-05) (hash(x)=47514225)
+30800 val loss 6.0215
+30800 val perplexity 412.1853
+30800 train 5.995923 (lr=3.9165e-05) (hash(x)=56160596)
+33800 val loss 5.9406
+33800 val perplexity 380.1490
+33800 train 5.760533 (lr=2.2067e-05) (hash(x)=50660204)
+33700 val loss 5.7849
+33700 val perplexity 325.3629
+33700 train 5.629024 (lr=1.5883e-05) (hash(x)=51164416)
+30900 val loss 6.0056
+30900 val perplexity 405.6912
+30900 train 5.963522 (lr=3.8900e-05) (hash(x)=49005636)
+33900 val loss 5.9383
+33900 val perplexity 379.2924
+33900 train 5.924343 (lr=2.1898e-05) (hash(x)=48511113)
+33800 val loss 5.7858
+33800 val perplexity 325.6466
+33800 train 5.623178 (lr=1.5762e-05) (hash(x)=50660204)
+31000 val loss 6.0090
+31000 val perplexity 407.0585
+31000 train 6.238300 (lr=3.8635e-05) (hash(x)=54646552)
+34000 val loss 5.9425
+34000 val perplexity 380.8846
+34000 train 5.565374 (lr=2.1729e-05) (hash(x)=41560072)
+33900 val loss 5.7916
+33900 val perplexity 327.5377
+33900 train 5.771410 (lr=1.5641e-05) (hash(x)=48511113)
+31100 val loss 6.0041
+31100 val perplexity 405.0728
+31100 train 5.864131 (lr=3.8371e-05) (hash(x)=44816272)
+34100 val loss 5.9380
+34100 val perplexity 379.1898
+34100 train 5.551881 (lr=2.1561e-05) (hash(x)=50354364)
+34000 val loss 5.7882
+34000 val perplexity 326.4260
+34000 train 5.413061 (lr=1.5521e-05) (hash(x)=41560072)
+31200 val loss 5.9949
+31200 val perplexity 401.3797
+31200 train 6.178637 (lr=3.8108e-05) (hash(x)=53040860)
+34200 val loss 5.9364
+34200 val perplexity 378.5548
+34200 train 6.116266 (lr=2.1394e-05) (hash(x)=54291562)
+34100 val loss 5.7894
+34100 val perplexity 326.8008
+34100 train 5.403695 (lr=1.5401e-05) (hash(x)=50354364)
+31300 val loss 5.9972
+31300 val perplexity 402.2843
+31300 train 6.192145 (lr=3.7845e-05) (hash(x)=52051764)
+34300 val loss 5.9460
+34300 val perplexity 382.2118
+34300 train 6.337032 (lr=2.1227e-05) (hash(x)=53340739)
+34200 val loss 5.7852
+34200 val perplexity 325.4492
+34200 train 5.969238 (lr=1.5281e-05) (hash(x)=54291562)
+34400 val loss 5.9334
+34400 val perplexity 377.4218
+34400 train 5.564046 (lr=2.1062e-05) (hash(x)=44872469)
+31400 val loss 5.9973
+31400 val perplexity 402.3336
+31400 train 5.958656 (lr=3.7583e-05) (hash(x)=51564589)
+34300 val loss 5.7894
+34300 val perplexity 326.8222
+34300 train 6.197208 (lr=1.5162e-05) (hash(x)=53340739)
+34500 val loss 5.9341
+34500 val perplexity 377.7030
+34500 train 5.025874 (lr=2.0896e-05) (hash(x)=23649625)
+31500 val loss 5.9998
+31500 val perplexity 403.3503
+31500 train 5.760409 (lr=3.7321e-05) (hash(x)=49676250)
+34400 val loss 5.7846
+34400 val perplexity 325.2628
+34400 train 5.422194 (lr=1.5044e-05) (hash(x)=44872469)
+34600 val loss 5.9337
+34600 val perplexity 377.5572
+34600 train 5.865894 (lr=2.0732e-05) (hash(x)=46230721)
+31600 val loss 5.9927
+31600 val perplexity 400.5085
+31600 train 5.933610 (lr=3.7061e-05) (hash(x)=49244964)
+34500 val loss 5.7804
+34500 val perplexity 323.8937
+34500 train 4.878257 (lr=1.4926e-05) (hash(x)=23649625)
+34700 val loss 5.9235
+34700 val perplexity 373.7112
+34700 train 5.711586 (lr=2.0568e-05) (hash(x)=44277516)
+31700 val loss 5.9937
+31700 val perplexity 400.9005
+31700 train 6.233249 (lr=3.6801e-05) (hash(x)=49527873)
+34600 val loss 5.7907
+34600 val perplexity 327.2462
+34600 train 5.752843 (lr=1.4809e-05) (hash(x)=46230721)
+34800 val loss 5.9240
+34800 val perplexity 373.9077
+34800 train 6.004957 (lr=2.0405e-05) (hash(x)=50346875)
+31800 val loss 6.0204
+31800 val perplexity 411.7344
+31800 train 6.004154 (lr=3.6541e-05) (hash(x)=49566755)
+34700 val loss 5.7750
+34700 val perplexity 322.1385
+34700 train 5.557886 (lr=1.4692e-05) (hash(x)=44277516)
+34900 val loss 5.9278
+34900 val perplexity 375.3319
+34900 train 5.866704 (lr=2.0243e-05) (hash(x)=45110550)
+31900 val loss 5.9966
+31900 val perplexity 402.0706
+31900 train 5.825653 (lr=3.6283e-05) (hash(x)=45456170)
+34800 val loss 5.7748
+34800 val perplexity 322.0846
+34800 train 5.829043 (lr=1.4575e-05) (hash(x)=50346875)
+35000 val loss 5.9182
+35000 val perplexity 371.7319
+35000 train 5.920014 (lr=2.0081e-05) (hash(x)=46812915)
+34900 val loss 5.7720
+34900 val perplexity 321.1802
+32000 val loss 5.9997
+32000 val perplexity 403.3213
+34900 train 5.727943 (lr=1.4459e-05) (hash(x)=45110550)
+32000 train 6.059169 (lr=3.6025e-05) (hash(x)=49986978)
+35100 val loss 5.9208
+35100 val perplexity 372.7207
+35100 train 6.078620 (lr=1.9921e-05) (hash(x)=54729285)
+35000 val loss 5.7717
+35000 val perplexity 321.0736
+35000 train 5.777222 (lr=1.4344e-05) (hash(x)=46812915)
+32100 val loss 5.9950
+32100 val perplexity 401.3996
+32100 train 5.665666 (lr=3.5768e-05) (hash(x)=50539149)
+35200 val loss 5.9262
+35200 val perplexity 374.7274
+35200 train 5.658880 (lr=1.9760e-05) (hash(x)=49651453)
+35100 val loss 5.7708
+35100 val perplexity 320.7886
+35100 train 5.920694 (lr=1.4229e-05) (hash(x)=54729285)
+32200 val loss 5.9912
+32200 val perplexity 399.8976
+32200 train 5.936154 (lr=3.5512e-05) (hash(x)=50998615)
+35300 val loss 5.9223
+35300 val perplexity 373.2706
+35300 train 5.794049 (lr=1.9601e-05) (hash(x)=51398605)
+35200 val loss 5.7704
+35200 val perplexity 320.6660
+35200 train 5.494881 (lr=1.4115e-05) (hash(x)=49651453)
+32300 val loss 5.9954
+32300 val perplexity 401.5585
+32300 train 5.825207 (lr=3.5256e-05) (hash(x)=44030191)
+35400 val loss 5.9208
+35400 val perplexity 372.6965
+35400 train 5.703063 (lr=1.9442e-05) (hash(x)=52214092)
+35300 val loss 5.7716
+35300 val perplexity 321.0661
+35300 train 5.669948 (lr=1.4001e-05) (hash(x)=51398605)
+32400 val loss 5.9985
+32400 val perplexity 402.8164
+32400 train 5.953857 (lr=3.5002e-05) (hash(x)=47269028)
+35500 val loss 5.9153
+35500 val perplexity 370.6734
+35500 train 6.209591 (lr=1.9285e-05) (hash(x)=60115839)
+35400 val loss 5.7703
+35400 val perplexity 320.6328
+35400 train 5.548427 (lr=1.3887e-05) (hash(x)=52214092)
+35600 val loss 5.9498
+35600 val perplexity 383.6671
+35600 train 6.562247 (lr=1.9128e-05) (hash(x)=53051302)
+32500 val loss 5.9967
+32500 val perplexity 402.1194
+32500 train 5.813130 (lr=3.4748e-05) (hash(x)=50128289)
+35500 val loss 5.7697
+35500 val perplexity 320.4387
+35500 train 6.081394 (lr=1.3775e-05) (hash(x)=60115839)
+35700 val loss 5.9232
+35700 val perplexity 373.5927
+35700 train 5.626495 (lr=1.8971e-05) (hash(x)=47769916)
+32600 val loss 6.0076
+32600 val perplexity 406.5134
+32600 train 5.874361 (lr=3.4495e-05) (hash(x)=42781351)
+35600 val loss 5.7860
+35600 val perplexity 325.7147
+35600 train 6.413506 (lr=1.3663e-05) (hash(x)=53051302)
+35800 val loss 5.9231
+35800 val perplexity 373.5498
+35800 train 5.692234 (lr=1.8816e-05) (hash(x)=48652911)
+32700 val loss 5.9894
+32700 val perplexity 399.1628
+32700 train 6.090833 (lr=3.4242e-05) (hash(x)=51712117)
+35700 val loss 5.7690
+35700 val perplexity 320.2285
+35700 train 5.468635 (lr=1.3551e-05) (hash(x)=47769916)
+35900 val loss 5.9212
+35900 val perplexity 372.8640
+35900 train 5.836629 (lr=1.8661e-05) (hash(x)=50452504)
+32800 val loss 5.9886
+32800 val perplexity 398.8493
+32800 train 6.088406 (lr=3.3991e-05) (hash(x)=52478104)
+35800 val loss 5.7692
+35800 val perplexity 320.2763
+35800 train 5.534987 (lr=1.3440e-05) (hash(x)=48652911)
+36000 val loss 5.9277
+36000 val perplexity 375.2843
+36000 train 5.916679 (lr=1.8507e-05) (hash(x)=44888966)
+32900 val loss 5.9836
+32900 val perplexity 396.8552
+32900 train 5.927821 (lr=3.3740e-05) (hash(x)=53588694)
+35900 val loss 5.7676
+35900 val perplexity 319.7616
+35900 train 5.702505 (lr=1.3329e-05) (hash(x)=50452504)
+36100 val loss 5.9185
+36100 val perplexity 371.8711
+36100 train 5.780041 (lr=1.8354e-05) (hash(x)=48245730)
+33000 val loss 5.9857
+33000 val perplexity 397.7124
+36000 val loss 5.7703
+36000 val perplexity 320.6316
+33000 train 5.878014 (lr=3.3490e-05) (hash(x)=50821436)
+36000 train 5.782542 (lr=1.3219e-05) (hash(x)=44888966)
+36200 val loss 5.9188
+36200 val perplexity 371.9759
+36200 train 5.890483 (lr=1.8201e-05) (hash(x)=51614607)
+36100 val loss 5.7698
+36100 val perplexity 320.4659
+36100 train 5.641719 (lr=1.3110e-05) (hash(x)=48245730)
+33100 val loss 5.9762
+33100 val perplexity 393.9346
+33100 train 5.923144 (lr=3.3242e-05) (hash(x)=50371199)
+36300 val loss 5.9136
+36300 val perplexity 370.0232
+36300 train 5.959064 (lr=1.8050e-05) (hash(x)=43069600)
+36200 val loss 5.7681
+36200 val perplexity 319.9161
+36200 train 5.747694 (lr=1.3001e-05) (hash(x)=51614607)
+36400 val loss 5.9079
+36400 val perplexity 367.9286
+36400 train 5.954860 (lr=1.7899e-05) (hash(x)=52756265)
+33200 val loss 5.9800
+33200 val perplexity 395.4338
+33200 train 6.167976 (lr=3.2993e-05) (hash(x)=54544623)
+36300 val loss 5.7611
+36300 val perplexity 317.7116
+36300 train 5.831945 (lr=1.2893e-05) (hash(x)=43069600)
+36500 val loss 5.9057
+36500 val perplexity 367.1093
+36500 train 5.705001 (lr=1.7749e-05) (hash(x)=47894210)
+33300 val loss 5.9807
+33300 val perplexity 395.7107
+33300 train 5.806248 (lr=3.2746e-05) (hash(x)=55456921)
+36600 val loss 5.9065
+36600 val perplexity 367.4315
+36600 train 5.917275 (lr=1.7600e-05) (hash(x)=46906979)
+36400 val loss 5.7586
+36400 val perplexity 316.9061
+36400 train 5.805235 (lr=1.2785e-05) (hash(x)=52756265)
+33400 val loss 5.9812
+33400 val perplexity 395.9162
+33400 train 5.589169 (lr=3.2500e-05) (hash(x)=37240412)
+36700 val loss 5.9063
+36700 val perplexity 367.3376
+36700 train 6.596786 (lr=1.7452e-05) (hash(x)=45349010)
+36500 val loss 5.7568
+36500 val perplexity 316.3482
+36500 train 5.554217 (lr=1.2678e-05) (hash(x)=47894210)
+33500 val loss 5.9829
+33500 val perplexity 396.5762
+33500 train 6.189414 (lr=3.2255e-05) (hash(x)=56551894)
+36800 val loss 5.9032
+36800 val perplexity 366.2145
+36800 train 5.672704 (lr=1.7305e-05) (hash(x)=40761454)
+36600 val loss 5.7610
+36600 val perplexity 317.6748
+36600 train 5.753194 (lr=1.2572e-05) (hash(x)=46906979)
+33600 val loss 5.9811
+33600 val perplexity 395.8841
+33600 train 5.884028 (lr=3.2010e-05) (hash(x)=47514225)
+36900 val loss 5.9045
+36900 val perplexity 366.6984
+36900 train 6.013577 (lr=1.7158e-05) (hash(x)=52150265)
+36700 val loss 5.7576
+36700 val perplexity 316.6031
+36700 train 6.510787 (lr=1.2466e-05) (hash(x)=45349010)
+33700 val loss 5.9750
+33700 val perplexity 393.4758
+33700 train 5.820536 (lr=3.1767e-05) (hash(x)=51164416)
+37000 val loss 5.9050
+37000 val perplexity 366.8657
+37000 train 5.954855 (lr=1.7012e-05) (hash(x)=48982242)
+36800 val loss 5.7534
+36800 val perplexity 315.2697
+36800 train 5.528115 (lr=1.2360e-05) (hash(x)=40761454)
+33800 val loss 5.9833
+33800 val perplexity 396.7519
+33800 train 5.811645 (lr=3.1524e-05) (hash(x)=50660204)
+37100 val loss 5.9046
+37100 val perplexity 366.7078
+37100 train 5.844863 (lr=1.6867e-05) (hash(x)=52100108)
+36900 val loss 5.7535
+36900 val perplexity 315.2880
+36900 train 5.866702 (lr=1.2256e-05) (hash(x)=52150265)
+37200 val loss 5.9152
+37200 val perplexity 370.6419
+37200 train 5.721406 (lr=1.6723e-05) (hash(x)=47721621)
+33900 val loss 5.9835
+33900 val perplexity 396.8295
+33900 train 5.958007 (lr=3.1282e-05) (hash(x)=48511113)
+37000 val loss 5.7543
+37000 val perplexity 315.5526
+37000 train 5.836509 (lr=1.2152e-05) (hash(x)=48982242)
+37300 val loss 5.9038
+37300 val perplexity 366.4096
+37300 train 5.695620 (lr=1.6580e-05) (hash(x)=49312812)
+34000 val loss 5.9786
+34000 val perplexity 394.8889
+34000 train 5.608625 (lr=3.1041e-05) (hash(x)=41560072)
+37100 val loss 5.7524
+37100 val perplexity 314.9599
+37100 train 5.706089 (lr=1.2048e-05) (hash(x)=52100108)
+37400 val loss 5.9026
+37400 val perplexity 365.9883
+37400 train 5.736526 (lr=1.6438e-05) (hash(x)=48536877)
+37200 val loss 5.7633
+37200 val perplexity 318.3847
+37200 train 5.581213 (lr=1.1945e-05) (hash(x)=47721621)
+34100 val loss 5.9784
+34100 val perplexity 394.8224
+34100 train 5.570742 (lr=3.0802e-05) (hash(x)=50354364)
+37500 val loss 5.9107
+37500 val perplexity 368.9655
+37500 train 5.860770 (lr=1.6297e-05) (hash(x)=50612814)
+37300 val loss 5.7523
+37300 val perplexity 314.9262
+37300 train 5.567525 (lr=1.1843e-05) (hash(x)=49312812)
+34200 val loss 5.9756
+34200 val perplexity 393.7104
+34200 train 6.144897 (lr=3.0563e-05) (hash(x)=54291562)
+37600 val loss 5.9110
+37600 val perplexity 369.0783
+37600 train 6.408382 (lr=1.6156e-05) (hash(x)=60685704)
+37400 val loss 5.7550
+37400 val perplexity 315.7796
+37400 train 5.594305 (lr=1.1741e-05) (hash(x)=48536877)
+34300 val loss 5.9963
+34300 val perplexity 401.9445
+34300 train 6.332589 (lr=3.0325e-05) (hash(x)=53340739)
+37700 val loss 5.9094
+37700 val perplexity 368.5012
+37700 train 5.984560 (lr=1.6016e-05) (hash(x)=56543272)
+37500 val loss 5.7565
+37500 val perplexity 316.2277
+37500 train 5.719148 (lr=1.1640e-05) (hash(x)=50612814)
+34400 val loss 5.9785
+34400 val perplexity 394.8461
+34400 train 5.606893 (lr=3.0088e-05) (hash(x)=44872469)
+37800 val loss 5.9097
+37800 val perplexity 368.5869
+37800 train 6.018356 (lr=1.5878e-05) (hash(x)=56659623)
+37600 val loss 5.7526
+37600 val perplexity 315.0231
+37600 train 6.202497 (lr=1.1540e-05) (hash(x)=60685704)
+34500 val loss 5.9714
+34500 val perplexity 392.0702
+34500 train 5.070857 (lr=2.9852e-05) (hash(x)=23649625)
+37900 val loss 5.9059
+37900 val perplexity 367.2103
+37900 train 5.579789 (lr=1.5740e-05) (hash(x)=48577952)
+37700 val loss 5.7533
+37700 val perplexity 315.2282
+37700 train 5.840503 (lr=1.1440e-05) (hash(x)=56543272)
+34600 val loss 5.9700
+34600 val perplexity 391.4864
+34600 train 5.886979 (lr=2.9617e-05) (hash(x)=46230721)
+38000 val loss 5.9061
+38000 val perplexity 367.2748
+38000 train 5.531864 (lr=1.5603e-05) (hash(x)=46719420)
+37800 val loss 5.7520
+37800 val perplexity 314.8298
+37800 train 5.849054 (lr=1.1341e-05) (hash(x)=56659623)
+34700 val loss 5.9640
+34700 val perplexity 389.1593
+34700 train 5.768269 (lr=2.9383e-05) (hash(x)=44277516)
+38100 val loss 5.9013
+38100 val perplexity 365.5296
+38100 train 5.842225 (lr=1.5467e-05) (hash(x)=56290069)
+37900 val loss 5.7519
+37900 val perplexity 314.7726
+37900 train 5.430374 (lr=1.1243e-05) (hash(x)=48577952)
+34800 val loss 5.9643
+34800 val perplexity 389.2964
+34800 train 6.035602 (lr=2.9150e-05) (hash(x)=50346875)
+38200 val loss 5.8980
+38200 val perplexity 364.3053
+38200 train 5.870751 (lr=1.5332e-05) (hash(x)=50056974)
+38000 val loss 5.7502
+38000 val perplexity 314.2440
+38000 train 5.378278 (lr=1.1145e-05) (hash(x)=46719420)
+38300 val loss 5.8994
+38300 val perplexity 364.8189
+38300 train 5.906262 (lr=1.5198e-05) (hash(x)=48797079)
+34900 val loss 5.9627
+34900 val perplexity 388.6727
+34900 train 5.890477 (lr=2.8918e-05) (hash(x)=45110550)
+38100 val loss 5.7475
+38100 val perplexity 313.4057
+38100 train 5.671755 (lr=1.1048e-05) (hash(x)=56290069)
+38400 val loss 5.8955
+38400 val perplexity 363.4131
+38400 train 5.766184 (lr=1.5064e-05) (hash(x)=49702043)
+35000 val loss 5.9614
+35000 val perplexity 388.1683
+35000 train 5.954320 (lr=2.8688e-05) (hash(x)=46812915)
+38200 val loss 5.7460
+38200 val perplexity 312.9408
+38200 train 5.734985 (lr=1.0951e-05) (hash(x)=50056974)
+38500 val loss 5.8951
+38500 val perplexity 363.2463
+38500 train 5.896713 (lr=1.4932e-05) (hash(x)=53375877)
+35100 val loss 5.9603
+35100 val perplexity 387.7407
+35100 train 6.104949 (lr=2.8458e-05) (hash(x)=54729285)
+38300 val loss 5.7506
+38300 val perplexity 314.3904
+38300 train 5.793857 (lr=1.0855e-05) (hash(x)=48797079)
+38600 val loss 5.8907
+38600 val perplexity 361.6438
+38600 train 5.924218 (lr=1.4801e-05) (hash(x)=52757407)
+35200 val loss 5.9641
+35200 val perplexity 389.2074
+35200 train 5.708729 (lr=2.8229e-05) (hash(x)=49651453)
+38400 val loss 5.7444
+38400 val perplexity 312.4273
+38400 train 5.612573 (lr=1.0760e-05) (hash(x)=49702043)
+38700 val loss 5.8897
+38700 val perplexity 361.3105
+38700 train 5.980316 (lr=1.4670e-05) (hash(x)=51335882)
+38500 val loss 5.7501
+38500 val perplexity 314.2229
+38500 train 5.758672 (lr=1.0666e-05) (hash(x)=53375877)
+35300 val loss 5.9593
+35300 val perplexity 387.3341
+35300 train 5.837429 (lr=2.8002e-05) (hash(x)=51398605)
+38800 val loss 5.8886
+38800 val perplexity 360.8871
+38800 train 5.843975 (lr=1.4541e-05) (hash(x)=46048181)
+38600 val loss 5.7405
+38600 val perplexity 311.2345
+38600 train 5.766521 (lr=1.0572e-05) (hash(x)=52757407)
+35400 val loss 5.9641
+35400 val perplexity 389.2109
+35400 train 5.720802 (lr=2.7775e-05) (hash(x)=52214092)
+38900 val loss 5.8925
+38900 val perplexity 362.2949
+38900 train 5.894951 (lr=1.4412e-05) (hash(x)=50536450)
+38700 val loss 5.7394
+38700 val perplexity 310.8754
+38700 train 5.840066 (lr=1.0479e-05) (hash(x)=51335882)
+35500 val loss 5.9579
+35500 val perplexity 386.8046
+35500 train 6.262779 (lr=2.7549e-05) (hash(x)=60115839)
+39000 val loss 5.8909
+39000 val perplexity 361.7365
+39000 train 5.733507 (lr=1.4285e-05) (hash(x)=49201278)
+38800 val loss 5.7400
+38800 val perplexity 311.0584
+38800 train 5.707039 (lr=1.0386e-05) (hash(x)=46048181)
+35600 val loss 5.9726
+35600 val perplexity 392.5206
+35600 train 6.608598 (lr=2.7325e-05) (hash(x)=53051302)
+39100 val loss 5.8942
+39100 val perplexity 362.9412
+39100 train 5.979485 (lr=1.4158e-05) (hash(x)=52669740)
+38900 val loss 5.7431
+38900 val perplexity 312.0286
+38900 train 5.794366 (lr=1.0294e-05) (hash(x)=50536450)
+35700 val loss 5.9619
+35700 val perplexity 388.3659
+35700 train 5.651849 (lr=2.7102e-05) (hash(x)=47769916)
+39200 val loss 5.8856
+39200 val perplexity 359.8025
+39200 train 5.918985 (lr=1.4032e-05) (hash(x)=45566341)
+39000 val loss 5.7360
+39000 val perplexity 309.8333
+39000 train 5.598691 (lr=1.0203e-05) (hash(x)=49201278)
+35800 val loss 5.9652
+35800 val perplexity 389.6476
+35800 train 5.729043 (lr=2.6880e-05) (hash(x)=48652911)
+39300 val loss 5.8898
+39300 val perplexity 361.3280
+39300 train 6.474769 (lr=1.3908e-05) (hash(x)=50582342)
+39100 val loss 5.7421
+39100 val perplexity 311.7166
+39100 train 5.855915 (lr=1.0113e-05) (hash(x)=52669740)
+39400 val loss 5.8845
+39400 val perplexity 359.4338
+39400 train 5.644921 (lr=1.3784e-05) (hash(x)=43591296)
+35900 val loss 5.9642
+35900 val perplexity 389.2486
+35900 train 5.887765 (lr=2.6659e-05) (hash(x)=50452504)
+39200 val loss 5.7334
+39200 val perplexity 309.0333
+39200 train 5.745511 (lr=1.0023e-05) (hash(x)=45566341)
+39500 val loss 5.8870
+39500 val perplexity 360.3081
+39500 train 5.672634 (lr=1.3661e-05) (hash(x)=49119520)
+36000 val loss 5.9668
+36000 val perplexity 390.2736
+36000 train 5.965039 (lr=2.6439e-05) (hash(x)=44888966)
+39300 val loss 5.7350
+39300 val perplexity 309.5198
+39300 train 6.364397 (lr=9.9341e-06) (hash(x)=50582342)
+39600 val loss 5.8894
+39600 val perplexity 361.2062
+39600 train 6.152529 (lr=1.3540e-05) (hash(x)=50278799)
+36100 val loss 5.9600
+36100 val perplexity 387.6286
+36100 train 5.837125 (lr=2.6220e-05) (hash(x)=48245730)
+39400 val loss 5.7312
+39400 val perplexity 308.3344
+39400 train 5.493511 (lr=9.8458e-06) (hash(x)=43591296)
+39700 val loss 5.8907
+39700 val perplexity 361.6723
+39700 train 5.793178 (lr=1.3419e-05) (hash(x)=54799713)
+36200 val loss 5.9556
+36200 val perplexity 385.8949
+36200 train 5.919549 (lr=2.6002e-05) (hash(x)=51614607)
+39500 val loss 5.7377
+39500 val perplexity 310.3366
+39500 train 5.531897 (lr=9.7581e-06) (hash(x)=49119520)
+39800 val loss 5.8902
+39800 val perplexity 361.4893
+39800 train 5.968296 (lr=1.3299e-05) (hash(x)=46290010)
+36300 val loss 5.9589
+36300 val perplexity 387.2019
+39600 val loss 5.7333
+39600 val perplexity 308.9752
+36300 train 6.025589 (lr=2.5786e-05) (hash(x)=43069600)
+39600 train 5.973525 (lr=9.6712e-06) (hash(x)=50278799)
+39900 val loss 5.8848
+39900 val perplexity 359.5159
+39900 train 5.575523 (lr=1.3180e-05) (hash(x)=41503413)
+39700 val loss 5.7371
+39700 val perplexity 310.1647
+39700 train 5.645143 (lr=9.5849e-06) (hash(x)=54799713)
+36400 val loss 5.9505
+36400 val perplexity 383.9391
+36400 train 5.992536 (lr=2.5570e-05) (hash(x)=52756265)
+40000 val loss 5.8812
+40000 val perplexity 358.2535
+40000 train 5.890353 (lr=1.3063e-05) (hash(x)=44949087)
+39800 val loss 5.7374
+39800 val perplexity 310.2665
+39800 train 5.844509 (lr=9.4994e-06) (hash(x)=46290010)
+36500 val loss 5.9535
+36500 val perplexity 385.0833
+36500 train 5.743821 (lr=2.5356e-05) (hash(x)=47894210)
+40100 val loss 5.8842
+40100 val perplexity 359.3134
+40100 train 5.445794 (lr=1.2946e-05) (hash(x)=45005676)
+39900 val loss 5.7342
+39900 val perplexity 309.2613
+39900 train 5.424328 (lr=9.4146e-06) (hash(x)=41503413)
+36600 val loss 5.9613
+36600 val perplexity 388.1315
+36600 train 5.988231 (lr=2.5143e-05) (hash(x)=46906979)
+40200 val loss 5.8873
+40200 val perplexity 360.4409
+40200 train 6.014713 (lr=1.2830e-05) (hash(x)=51341494)
+40000 val loss 5.7298
+40000 val perplexity 307.9146
+40000 train 5.780132 (lr=9.3305e-06) (hash(x)=44949087)
+36700 val loss 5.9497
+36700 val perplexity 383.6526
+36700 train 6.647394 (lr=2.4931e-05) (hash(x)=45349010)
+40300 val loss 5.8895
+40300 val perplexity 361.2133
+40300 train 5.618561 (lr=1.2716e-05) (hash(x)=47686820)
+40100 val loss 5.7352
+40100 val perplexity 309.5796
+40100 train 5.289546 (lr=9.2472e-06) (hash(x)=45005676)
+36800 val loss 5.9443
+36800 val perplexity 381.5670
+40400 val loss 5.8878
+40400 val perplexity 360.6049
+36800 train 5.698735 (lr=2.4721e-05) (hash(x)=40761454)
+40400 train 5.754405 (lr=1.2602e-05) (hash(x)=47142962)
+40200 val loss 5.7344
+40200 val perplexity 309.3215
+40200 train 5.891187 (lr=9.1646e-06) (hash(x)=51341494)
+40500 val loss 5.8892
+40500 val perplexity 361.1151
+40500 train 5.630478 (lr=1.2489e-05) (hash(x)=47585616)
+36900 val loss 5.9474
+36900 val perplexity 382.7688
+36900 train 6.036154 (lr=2.4511e-05) (hash(x)=52150265)
+40300 val loss 5.7342
+40300 val perplexity 309.2594
+40300 train 5.450536 (lr=9.0827e-06) (hash(x)=47686820)
+40600 val loss 5.8905
+40600 val perplexity 361.5762
+40600 train 5.703403 (lr=1.2378e-05) (hash(x)=50075832)
+37000 val loss 5.9514
+37000 val perplexity 384.2935
+37000 train 6.014409 (lr=2.4303e-05) (hash(x)=48982242)
+40400 val loss 5.7329
+40400 val perplexity 308.8696
+40400 train 5.605383 (lr=9.0015e-06) (hash(x)=47142962)
+40700 val loss 5.8813
+40700 val perplexity 358.2859
+40700 train 5.902246 (lr=1.2267e-05) (hash(x)=52764040)
+37100 val loss 5.9486
+37100 val perplexity 383.2138
+37100 train 5.875566 (lr=2.4096e-05) (hash(x)=52100108)
+40500 val loss 5.7341
+40500 val perplexity 309.2464
+40500 train 5.492334 (lr=8.9211e-06) (hash(x)=47585616)
+40800 val loss 5.8754
+40800 val perplexity 356.1506
+40800 train 6.216906 (lr=1.2158e-05) (hash(x)=56727171)
+37200 val loss 5.9553
+37200 val perplexity 385.7828
+37200 train 5.766670 (lr=2.3890e-05) (hash(x)=47721621)
+40600 val loss 5.7348
+40600 val perplexity 309.4454
+40600 train 5.560070 (lr=8.8414e-06) (hash(x)=50075832)
+40900 val loss 5.8740
+40900 val perplexity 355.6850
+40900 train 5.906456 (lr=1.2049e-05) (hash(x)=53111899)
+37300 val loss 5.9476
+37300 val perplexity 382.8522
+37300 train 5.734115 (lr=2.3686e-05) (hash(x)=49312812)
+40700 val loss 5.7274
+40700 val perplexity 307.1826
+40700 train 5.776053 (lr=8.7624e-06) (hash(x)=52764040)
+41000 val loss 5.8728
+41000 val perplexity 355.2520
+41000 train 5.748232 (lr=1.1942e-05) (hash(x)=45951630)
+37400 val loss 5.9480
+37400 val perplexity 382.9915
+37400 train 5.784721 (lr=2.3483e-05) (hash(x)=48536877)
+40800 val loss 5.7236
+40800 val perplexity 305.9948
+40800 train 6.051008 (lr=8.6842e-06) (hash(x)=56727171)
+41100 val loss 5.8737
+41100 val perplexity 355.5605
+41100 train 6.203441 (lr=1.1836e-05) (hash(x)=59664516)
+37500 val loss 5.9521
+37500 val perplexity 384.5589
+40900 val loss 5.7222
+40900 val perplexity 305.5659
+40900 train 5.765467 (lr=8.6068e-06) (hash(x)=53111899)
+37500 train 5.895791 (lr=2.3281e-05) (hash(x)=50612814)
+41200 val loss 5.8818
+41200 val perplexity 358.4606
+41200 train 5.769986 (lr=1.1730e-05) (hash(x)=50414471)
+41000 val loss 5.7185
+41000 val perplexity 304.4630
+41000 train 5.610645 (lr=8.5301e-06) (hash(x)=45951630)
+37600 val loss 5.9558
+37600 val perplexity 385.9926
+37600 train 6.439344 (lr=2.3080e-05) (hash(x)=60685704)
+41300 val loss 5.8697
+41300 val perplexity 354.1432
+41300 train 5.165312 (lr=1.1626e-05) (hash(x)=33823056)
+41100 val loss 5.7223
+41100 val perplexity 305.5957
+41100 train 6.084666 (lr=8.4541e-06) (hash(x)=59664516)
+37700 val loss 5.9558
+37700 val perplexity 385.9683
+37700 train 6.046243 (lr=2.2881e-05) (hash(x)=56543272)
+41400 val loss 5.8664
+41400 val perplexity 352.9914
+41400 train 5.976327 (lr=1.1523e-05) (hash(x)=52911940)
+41200 val loss 5.7251
+41200 val perplexity 306.4495
+41200 train 5.620420 (lr=8.3789e-06) (hash(x)=50414471)
+37800 val loss 5.9506
+37800 val perplexity 383.9885
+37800 train 6.052326 (lr=2.2682e-05) (hash(x)=56659623)
+41500 val loss 5.8748
+41500 val perplexity 355.9481
+41500 train 5.430792 (lr=1.1421e-05) (hash(x)=40005218)
+41300 val loss 5.7186
+41300 val perplexity 304.4719
+41300 train 5.012256 (lr=8.3045e-06) (hash(x)=33823056)
+37900 val loss 5.9526
+37900 val perplexity 384.7452
+37900 train 5.634081 (lr=2.2485e-05) (hash(x)=48577952)
+41600 val loss 5.8644
+41600 val perplexity 352.2685
+41600 train 6.291268 (lr=1.1320e-05) (hash(x)=55223673)
+41400 val loss 5.7181
+41400 val perplexity 304.3268
+41400 train 5.824917 (lr=8.2308e-06) (hash(x)=52911940)
+41700 val loss 5.8624
+41700 val perplexity 351.5634
+41700 train 5.768177 (lr=1.1220e-05) (hash(x)=49603559)
+38000 val loss 5.9556
+38000 val perplexity 385.9087
+38000 train 5.585528 (lr=2.2290e-05) (hash(x)=46719420)
+41500 val loss 5.7258
+41500 val perplexity 306.6671
+41500 train 5.284519 (lr=8.1579e-06) (hash(x)=40005218)
+41800 val loss 5.8629
+41800 val perplexity 351.7599
+41800 train 5.875134 (lr=1.1121e-05) (hash(x)=51692850)
+38100 val loss 5.9439
+38100 val perplexity 381.4094
+38100 train 5.886159 (lr=2.2096e-05) (hash(x)=56290069)
+41600 val loss 5.7169
+41600 val perplexity 303.9648
+41600 train 6.142447 (lr=8.0858e-06) (hash(x)=55223673)
+41900 val loss 5.8626
+41900 val perplexity 351.6398
+41900 train 5.640895 (lr=1.1024e-05) (hash(x)=51210799)
+38200 val loss 5.9461
+38200 val perplexity 382.2668
+38200 train 5.929902 (lr=2.1903e-05) (hash(x)=50056974)
+41700 val loss 5.7138
+41700 val perplexity 303.0121
+41700 train 5.615955 (lr=8.0144e-06) (hash(x)=49603559)
+42000 val loss 5.8647
+42000 val perplexity 352.3722
+42000 train 6.372183 (lr=1.0927e-05) (hash(x)=60592263)
+38300 val loss 5.9449
+38300 val perplexity 381.7824
+38300 train 5.941628 (lr=2.1711e-05) (hash(x)=48797079)
+41800 val loss 5.7159
+41800 val perplexity 303.6550
+41800 train 5.753294 (lr=7.9438e-06) (hash(x)=51692850)
+42100 val loss 5.8651
+42100 val perplexity 352.5169
+42100 train 5.739551 (lr=1.0831e-05) (hash(x)=48159119)
+38400 val loss 5.9387
+38400 val perplexity 379.4237
+38400 train 5.805596 (lr=2.1521e-05) (hash(x)=49702043)
+41900 val loss 5.7153
+41900 val perplexity 303.4893
+41900 train 5.475625 (lr=7.8740e-06) (hash(x)=51210799)
+42200 val loss 5.8667
+42200 val perplexity 353.0951
+42200 train 5.752390 (lr=1.0737e-05) (hash(x)=48619712)
+38500 val loss 5.9378
+38500 val perplexity 379.0855
+38500 train 5.958379 (lr=2.1332e-05) (hash(x)=53375877)
+42000 val loss 5.7153
+42000 val perplexity 303.4699
+42000 train 6.237113 (lr=7.8050e-06) (hash(x)=60592263)
+42300 val loss 5.8667
+42300 val perplexity 353.0851
+42300 train 5.594993 (lr=1.0644e-05) (hash(x)=56574312)
+38600 val loss 5.9333
+38600 val perplexity 377.3937
+42100 val loss 5.7131
+42100 val perplexity 302.8004
+42100 train 5.584966 (lr=7.7368e-06) (hash(x)=48159119)
+38600 train 5.949863 (lr=2.1144e-05) (hash(x)=52757407)
+42400 val loss 5.8681
+42400 val perplexity 353.5847
+42400 train 5.969612 (lr=1.0552e-05) (hash(x)=47020520)
+42200 val loss 5.7152
+42200 val perplexity 303.4401
+42200 train 5.610136 (lr=7.6693e-06) (hash(x)=48619712)
+38700 val loss 5.9322
+38700 val perplexity 377.0004
+38700 train 6.038456 (lr=2.0957e-05) (hash(x)=51335882)
+42500 val loss 5.8668
+42500 val perplexity 353.1131
+42500 train 5.754909 (lr=1.0460e-05) (hash(x)=48787571)
+42300 val loss 5.7129
+42300 val perplexity 302.7350
+42300 train 5.438562 (lr=7.6027e-06) (hash(x)=56574312)
+38800 val loss 5.9379
+38800 val perplexity 379.1240
+38800 train 5.899038 (lr=2.0773e-05) (hash(x)=46048181)
+42600 val loss 5.8668
+42600 val perplexity 353.1163
+42600 train 5.903190 (lr=1.0370e-05) (hash(x)=53734398)
+42400 val loss 5.7153
+42400 val perplexity 303.4768
+42400 train 5.836233 (lr=7.5368e-06) (hash(x)=47020520)
+38900 val loss 5.9450
+38900 val perplexity 381.8209
+38900 train 5.936474 (lr=2.0589e-05) (hash(x)=50536450)
+42700 val loss 5.8699
+42700 val perplexity 354.2037
+42700 train 5.664432 (lr=1.0282e-05) (hash(x)=52179072)
+42500 val loss 5.7147
+42500 val perplexity 303.2787
+42500 train 5.601105 (lr=7.4717e-06) (hash(x)=48787571)
+39000 val loss 5.9299
+39000 val perplexity 376.1078
+39000 train 5.762484 (lr=2.0407e-05) (hash(x)=49201278)
+42800 val loss 5.8727
+42800 val perplexity 355.1973
+42800 train 5.587573 (lr=1.0194e-05) (hash(x)=45303025)
+42600 val loss 5.7174
+42600 val perplexity 304.1232
+42600 train 5.799463 (lr=7.4074e-06) (hash(x)=53734398)
+42900 val loss 5.8688
+42900 val perplexity 353.8340
+42900 train 6.018352 (lr=1.0107e-05) (hash(x)=53649191)
+39100 val loss 5.9353
+39100 val perplexity 378.1472
+39100 train 6.017318 (lr=2.0226e-05) (hash(x)=52669740)
+42700 val loss 5.7176
+42700 val perplexity 304.1840
+42700 train 5.492084 (lr=7.3440e-06) (hash(x)=52179072)
+43000 val loss 5.8701
+43000 val perplexity 354.2768
+43000 train 6.291034 (lr=1.0022e-05) (hash(x)=61985066)
+39200 val loss 5.9277
+39200 val perplexity 375.3090
+39200 train 5.960580 (lr=2.0046e-05) (hash(x)=45566341)
+42800 val loss 5.7190
+42800 val perplexity 304.5946
+42800 train 5.429344 (lr=7.2813e-06) (hash(x)=45303025)
+43100 val loss 5.8730
+43100 val perplexity 355.3002
+43100 train 5.842031 (lr=9.9373e-06) (hash(x)=48533037)
+39300 val loss 5.9307
+39300 val perplexity 376.3995
+39300 train 6.510407 (lr=1.9868e-05) (hash(x)=50582342)
+42900 val loss 5.7162
+42900 val perplexity 303.7518
+42900 train 5.863437 (lr=7.2194e-06) (hash(x)=53649191)
+43200 val loss 5.8654
+43200 val perplexity 352.6376
+43200 train 5.989261 (lr=9.8541e-06) (hash(x)=50630327)
+39400 val loss 5.9261
+39400 val perplexity 374.6799
+39400 train 5.687747 (lr=1.9692e-05) (hash(x)=43591296)
+43000 val loss 5.7161
+43000 val perplexity 303.7293
+43000 train 6.172473 (lr=7.1583e-06) (hash(x)=61985066)
+43300 val loss 5.8752
+43300 val perplexity 356.0905
+43300 train 5.952151 (lr=9.7720e-06) (hash(x)=54845917)
+39500 val loss 5.9277
+39500 val perplexity 375.2769
+39500 train 5.709194 (lr=1.9516e-05) (hash(x)=49119520)
+43100 val loss 5.7156
+43100 val perplexity 303.5709
+43100 train 5.700378 (lr=7.0981e-06) (hash(x)=48533037)
+43400 val loss 5.8592
+43400 val perplexity 350.4557
+43400 train 5.792507 (lr=9.6911e-06) (hash(x)=52858739)
+39600 val loss 5.9286
+39600 val perplexity 375.6375
+39600 train 6.205977 (lr=1.9342e-05) (hash(x)=50278799)
+43200 val loss 5.7116
+43200 val perplexity 302.3573
+43200 train 5.828331 (lr=7.0386e-06) (hash(x)=50630327)
+43500 val loss 5.8588
+43500 val perplexity 350.3079
+43500 train 5.923429 (lr=9.6113e-06) (hash(x)=49761319)
+39700 val loss 5.9306
+39700 val perplexity 376.3667
+39700 train 5.834247 (lr=1.9170e-05) (hash(x)=54799713)
+43600 val loss 5.8551
+43600 val perplexity 349.0164
+43600 train 6.121321 (lr=9.5326e-06) (hash(x)=64879186)
+43300 val loss 5.7172
+43300 val perplexity 304.0533
+43300 train 5.803127 (lr=6.9800e-06) (hash(x)=54845917)
+39800 val loss 5.9308
+39800 val perplexity 376.4626
+39800 train 6.007029 (lr=1.8999e-05) (hash(x)=46290010)
+43700 val loss 5.8557
+43700 val perplexity 349.2031
+43700 train 6.082420 (lr=9.4552e-06) (hash(x)=55246427)
+43400 val loss 5.7106
+43400 val perplexity 302.0658
+43400 train 5.642049 (lr=6.9222e-06) (hash(x)=52858739)
+43800 val loss 5.8551
+43800 val perplexity 348.9951
+43800 train 5.781734 (lr=9.3788e-06) (hash(x)=51238313)
+39900 val loss 5.9276
+39900 val perplexity 375.2378
+39900 train 5.615714 (lr=1.8829e-05) (hash(x)=41503413)
+43500 val loss 5.7057
+43500 val perplexity 300.5862
+43500 train 5.784788 (lr=6.8652e-06) (hash(x)=49761319)
+43900 val loss 5.8544
+43900 val perplexity 348.7738
+43900 train 5.884986 (lr=9.3036e-06) (hash(x)=51664512)
+40000 val loss 5.9289
+40000 val perplexity 375.7552
+40000 train 5.930650 (lr=1.8661e-05) (hash(x)=44949087)
+43600 val loss 5.7053
+43600 val perplexity 300.4639
+43600 train 5.983207 (lr=6.8090e-06) (hash(x)=64879186)
+44000 val loss 5.8898
+44000 val perplexity 361.3231
+44000 train 6.113993 (lr=9.2296e-06) (hash(x)=64172296)
+43700 val loss 5.7048
+43700 val perplexity 300.3196
+43700 train 5.929050 (lr=6.7537e-06) (hash(x)=55246427)
+40100 val loss 5.9307
+40100 val perplexity 376.4187
+40100 train 5.513899 (lr=1.8494e-05) (hash(x)=45005676)
+44100 val loss 5.8512
+44100 val perplexity 347.6626
+44100 train 5.742202 (lr=9.1568e-06) (hash(x)=50205960)
+43800 val loss 5.7055
+43800 val perplexity 300.5147
+43800 train 5.646482 (lr=6.6992e-06) (hash(x)=51238313)
+40200 val loss 5.9346
+40200 val perplexity 377.8765
+40200 train 6.103844 (lr=1.8329e-05) (hash(x)=51341494)
+44200 val loss 5.8520
+44200 val perplexity 347.9336
+44200 train 5.883790 (lr=9.0851e-06) (hash(x)=51065522)
+43900 val loss 5.7033
+43900 val perplexity 299.8510
+43900 train 5.757020 (lr=6.6455e-06) (hash(x)=51664512)
+40300 val loss 5.9347
+40300 val perplexity 377.9381
+40300 train 5.685302 (lr=1.8165e-05) (hash(x)=47686820)
+44300 val loss 5.8538
+44300 val perplexity 348.5608
+44300 train 5.986463 (lr=9.0146e-06) (hash(x)=56163183)
+44000 val loss 5.7167
+44000 val perplexity 303.9045
+44000 train 6.008539 (lr=6.5926e-06) (hash(x)=64172296)
+40400 val loss 5.9359
+40400 val perplexity 378.3888
+40400 train 5.806218 (lr=1.8003e-05) (hash(x)=47142962)
+44400 val loss 5.8563
+44400 val perplexity 349.4405
+44400 train 5.900486 (lr=8.9453e-06) (hash(x)=50818020)
+44100 val loss 5.7018
+44100 val perplexity 299.4078
+44100 train 5.627808 (lr=6.5406e-06) (hash(x)=50205960)
+40500 val loss 5.9411
+40500 val perplexity 380.3375
+40500 train 5.667961 (lr=1.7842e-05) (hash(x)=47585616)
+44500 val loss 5.8550
+44500 val perplexity 348.9717
+44500 train 5.986432 (lr=8.8771e-06) (hash(x)=52855559)
+44200 val loss 5.7011
+44200 val perplexity 299.1825
+44200 train 5.744282 (lr=6.4894e-06) (hash(x)=51065522)
+40600 val loss 5.9388
+40600 val perplexity 379.4912
+40600 train 5.744999 (lr=1.7683e-05) (hash(x)=50075832)
+44600 val loss 5.8543
+44600 val perplexity 348.7167
+44600 train 5.761628 (lr=8.8101e-06) (hash(x)=44014306)
+44300 val loss 5.7041
+44300 val perplexity 300.0847
+44300 train 5.803062 (lr=6.4390e-06) (hash(x)=56163183)
+40700 val loss 5.9241
+40700 val perplexity 373.9476
+40700 train 5.948530 (lr=1.7525e-05) (hash(x)=52764040)
+44700 val loss 5.8532
+44700 val perplexity 348.3551
+44700 train 6.016201 (lr=8.7443e-06) (hash(x)=53648768)
+44400 val loss 5.7043
+44400 val perplexity 300.1653
+44400 train 5.741302 (lr=6.3895e-06) (hash(x)=50818020)
+40800 val loss 5.9185
+40800 val perplexity 371.8546
+40800 train 6.221800 (lr=1.7368e-05) (hash(x)=56727171)
+44800 val loss 5.8530
+44800 val perplexity 348.2740
+44800 train 5.767185 (lr=8.6797e-06) (hash(x)=42436477)
+44500 val loss 5.7031
+44500 val perplexity 299.7960
+44500 train 5.825619 (lr=6.3408e-06) (hash(x)=52855559)
+40900 val loss 5.9173
+40900 val perplexity 371.4223
+40900 train 5.929842 (lr=1.7214e-05) (hash(x)=53111899)
+44900 val loss 5.8536
+44900 val perplexity 348.4800
+44900 train 5.579908 (lr=8.6163e-06) (hash(x)=48409149)
+44600 val loss 5.7044
+44600 val perplexity 300.1988
+44600 train 5.614123 (lr=6.2929e-06) (hash(x)=44014306)
+45000 val loss 5.8550
+45000 val perplexity 348.9794
+41000 val loss 5.9133
+41000 val perplexity 369.9288
+41000 train 5.789283 (lr=1.7060e-05) (hash(x)=45951630)
+45000 train 5.766385 (lr=8.5540e-06) (hash(x)=53084805)
+44700 val loss 5.7035
+44700 val perplexity 299.9299
+44700 train 5.847026 (lr=6.2459e-06) (hash(x)=53648768)
+45100 val loss 5.8559
+45100 val perplexity 349.2931
+45100 train 5.713241 (lr=8.4930e-06) (hash(x)=56386155)
+41100 val loss 5.9142
+41100 val perplexity 370.2535
+41100 train 6.271932 (lr=1.6908e-05) (hash(x)=59664516)
+44800 val loss 5.7036
+44800 val perplexity 299.9361
+44800 train 5.658469 (lr=6.1998e-06) (hash(x)=42436477)
+45200 val loss 5.8540
+45200 val perplexity 348.6206
+45200 train 5.865363 (lr=8.4331e-06) (hash(x)=53389864)
+41200 val loss 5.9227
+41200 val perplexity 373.4152
+41200 train 5.812696 (lr=1.6758e-05) (hash(x)=50414471)
+44900 val loss 5.7056
+44900 val perplexity 300.5589
+44900 train 5.430418 (lr=6.1545e-06) (hash(x)=48409149)
+45300 val loss 5.8582
+45300 val perplexity 350.1080
+45300 train 5.890651 (lr=8.3745e-06) (hash(x)=50817456)
+45000 val loss 5.7020
+45000 val perplexity 299.4788
+45000 train 5.613445 (lr=6.1100e-06) (hash(x)=53084805)
+41300 val loss 5.9105
+41300 val perplexity 368.9034
+41300 train 5.231784 (lr=1.6609e-05) (hash(x)=33823056)
+45400 val loss 5.8611
+45400 val perplexity 351.1030
+45400 train 5.894712 (lr=8.3170e-06) (hash(x)=50081660)
+45100 val loss 5.7031
+45100 val perplexity 299.8070
+45100 train 5.568305 (lr=6.0664e-06) (hash(x)=56386155)
+41400 val loss 5.9102
+41400 val perplexity 368.7863
+41400 train 6.018604 (lr=1.6462e-05) (hash(x)=52911940)
+45500 val loss 5.8623
+45500 val perplexity 351.5262
+45500 train 5.613654 (lr=8.2607e-06) (hash(x)=55567121)
+45200 val loss 5.7030
+45200 val perplexity 299.7704
+45200 train 5.727294 (lr=6.0237e-06) (hash(x)=53389864)
+41500 val loss 5.9153
+41500 val perplexity 370.6497
+41500 train 5.463718 (lr=1.6316e-05) (hash(x)=40005218)
+45600 val loss 5.8531
+45600 val perplexity 348.3091
+45600 train 5.933183 (lr=8.2057e-06) (hash(x)=48902854)
+45300 val loss 5.7054
+45300 val perplexity 300.4919
+45300 train 5.737086 (lr=5.9818e-06) (hash(x)=50817456)
+41600 val loss 5.9086
+41600 val perplexity 368.1989
+41600 train 6.333064 (lr=1.6172e-05) (hash(x)=55223673)
+45700 val loss 5.8465
+45700 val perplexity 346.0248
+45700 train 6.066646 (lr=8.1518e-06) (hash(x)=53803301)
+45400 val loss 5.7068
+45400 val perplexity 300.9168
+45400 train 5.736176 (lr=5.9407e-06) (hash(x)=50081660)
+41700 val loss 5.9105
+41700 val perplexity 368.8865
+41700 train 5.812880 (lr=1.6029e-05) (hash(x)=49603559)
+45800 val loss 5.8459
+45800 val perplexity 345.8038
+45800 train 5.908492 (lr=8.0992e-06) (hash(x)=47576606)
+45500 val loss 5.7068
+45500 val perplexity 300.9174
+45500 train 5.440077 (lr=5.9005e-06) (hash(x)=55567121)
+41800 val loss 5.9084
+41800 val perplexity 368.1216
+41800 train 5.941528 (lr=1.5888e-05) (hash(x)=51692850)
+45900 val loss 5.8443
+45900 val perplexity 345.2653
+45900 train 5.754907 (lr=8.0478e-06) (hash(x)=56276981)
+45600 val loss 5.7003
+45600 val perplexity 298.9519
+45600 train 5.764146 (lr=5.8612e-06) (hash(x)=48902854)
+41900 val loss 5.9105
+41900 val perplexity 368.9075
+41900 train 5.696601 (lr=1.5748e-05) (hash(x)=51210799)
+46000 val loss 5.8472
+46000 val perplexity 346.2800
+46000 train 5.789150 (lr=7.9976e-06) (hash(x)=52174902)
+45700 val loss 5.6972
+45700 val perplexity 298.0337
+45700 train 5.943999 (lr=5.8227e-06) (hash(x)=53803301)
+42000 val loss 5.9139
+42000 val perplexity 370.1391
+42000 train 6.437364 (lr=1.5610e-05) (hash(x)=60592263)
+46100 val loss 5.8457
+46100 val perplexity 345.7564
+46100 train 5.966497 (lr=7.9485e-06) (hash(x)=52155490)
+45800 val loss 5.6954
+45800 val perplexity 297.5070
+45800 train 5.750324 (lr=5.7851e-06) (hash(x)=47576606)
+46200 val loss 5.8436
+46200 val perplexity 345.0224
+42100 val loss 5.9121
+42100 val perplexity 369.4916
+46200 train 5.706237 (lr=7.9008e-06) (hash(x)=46489984)
+42100 train 5.783825 (lr=1.5474e-05) (hash(x)=48159119)
+45900 val loss 5.6947
+45900 val perplexity 297.2770
+45900 train 5.579243 (lr=5.7484e-06) (hash(x)=56276981)
+46300 val loss 5.8429
+46300 val perplexity 344.7708
+46300 train 5.817455 (lr=7.8542e-06) (hash(x)=56603976)
+42200 val loss 5.9136
+42200 val perplexity 370.0398
+42200 train 5.816652 (lr=1.5339e-05) (hash(x)=48619712)
+46000 val loss 5.6938
+46000 val perplexity 297.0068
+46000 train 5.646109 (lr=5.7125e-06) (hash(x)=52174902)
+46400 val loss 5.8429
+46400 val perplexity 344.7834
+46400 train 5.657983 (lr=7.8088e-06) (hash(x)=51228412)
+46100 val loss 5.6935
+46100 val perplexity 296.9176
+46100 train 5.822449 (lr=5.6775e-06) (hash(x)=52155490)
+42300 val loss 5.9119
+42300 val perplexity 369.4113
+42300 train 5.644996 (lr=1.5205e-05) (hash(x)=56574312)
+46500 val loss 5.8424
+46500 val perplexity 344.6148
+46500 train 5.694198 (lr=7.7647e-06) (hash(x)=48511186)
+46200 val loss 5.6927
+46200 val perplexity 296.6807
+46200 train 5.518863 (lr=5.6434e-06) (hash(x)=46489984)
+42400 val loss 5.9162
+42400 val perplexity 370.9913
+42400 train 5.995991 (lr=1.5074e-05) (hash(x)=47020520)
+46600 val loss 5.8406
+46600 val perplexity 343.9942
+46600 train 5.857633 (lr=7.7218e-06) (hash(x)=48832039)
+46300 val loss 5.6930
+46300 val perplexity 296.7935
+46300 train 5.699072 (lr=5.6101e-06) (hash(x)=56603976)
+42500 val loss 5.9106
+42500 val perplexity 368.9411
+42500 train 5.795770 (lr=1.4943e-05) (hash(x)=48787571)
+46700 val loss 5.8418
+46700 val perplexity 344.3911
+46700 train 6.526180 (lr=7.6801e-06) (hash(x)=65821271)
+46400 val loss 5.6924
+46400 val perplexity 296.5966
+46400 train 5.499797 (lr=5.5777e-06) (hash(x)=51228412)
+42600 val loss 5.9168
+42600 val perplexity 371.2081
+42600 train 5.945784 (lr=1.4815e-05) (hash(x)=53734398)
+46800 val loss 5.8418
+46800 val perplexity 344.3912
+46800 train 5.541548 (lr=7.6397e-06) (hash(x)=48005593)
+46500 val loss 5.6914
+46500 val perplexity 296.3133
+46500 train 5.526657 (lr=5.5462e-06) (hash(x)=48511186)
+42700 val loss 5.9162
+42700 val perplexity 370.9931
+42700 train 5.741552 (lr=1.4688e-05) (hash(x)=52179072)
+46900 val loss 5.8433
+46900 val perplexity 344.9227
+46900 train 6.121469 (lr=7.6004e-06) (hash(x)=53182453)
+46600 val loss 5.6906
+46600 val perplexity 296.0728
+46600 train 5.720109 (lr=5.5156e-06) (hash(x)=48832039)
+42800 val loss 5.9187
+42800 val perplexity 371.9264
+42800 train 5.617833 (lr=1.4563e-05) (hash(x)=45303025)
+47000 val loss 5.8440
+47000 val perplexity 345.1522
+47000 train 5.734328 (lr=7.5624e-06) (hash(x)=49385983)
+46700 val loss 5.6914
+46700 val perplexity 296.2935
+46700 train 6.375406 (lr=5.4858e-06) (hash(x)=65821271)
+47100 val loss 5.8438
+47100 val perplexity 345.0918
+47100 train 5.823684 (lr=7.5257e-06) (hash(x)=48801622)
+42900 val loss 5.9158
+42900 val perplexity 370.8459
+42900 train 6.020298 (lr=1.4439e-05) (hash(x)=53649191)
+46800 val loss 5.6914
+46800 val perplexity 296.3135
+46800 train 5.400339 (lr=5.4569e-06) (hash(x)=48005593)
+47200 val loss 5.8434
+47200 val perplexity 344.9362
+47200 train 5.573141 (lr=7.4901e-06) (hash(x)=45216690)
+43000 val loss 5.9187
+43000 val perplexity 371.9230
+43000 train 6.334957 (lr=1.4317e-05) (hash(x)=61985066)
+46900 val loss 5.6906
+46900 val perplexity 296.0840
+46900 train 5.951502 (lr=5.4289e-06) (hash(x)=53182453)
+47300 val loss 5.8429
+47300 val perplexity 344.7614
+47300 train 5.664711 (lr=7.4558e-06) (hash(x)=51185601)
+43100 val loss 5.9161
+43100 val perplexity 370.9705
+43100 train 5.899559 (lr=1.4196e-05) (hash(x)=48533037)
+47000 val loss 5.6906
+47000 val perplexity 296.0736
+47000 train 5.593194 (lr=5.4017e-06) (hash(x)=49385983)
+47400 val loss 5.8412
+47400 val perplexity 344.1809
+47400 train 5.535730 (lr=7.4228e-06) (hash(x)=47743508)
+47100 val loss 5.6921
+47100 val perplexity 296.5078
+47100 train 5.685345 (lr=5.3755e-06) (hash(x)=48801622)
+43200 val loss 5.9092
+43200 val perplexity 368.4084
+43200 train 6.050465 (lr=1.4077e-05) (hash(x)=50630327)
+47500 val loss 5.8437
+47500 val perplexity 345.0433
+47500 train 5.638478 (lr=7.3909e-06) (hash(x)=49318833)
+47200 val loss 5.6920
+47200 val perplexity 296.4918
+47200 train 5.443398 (lr=5.3501e-06) (hash(x)=45216690)
+43300 val loss 5.9195
+43300 val perplexity 372.2409
+43300 train 5.995340 (lr=1.3960e-05) (hash(x)=54845917)
+47600 val loss 5.8437
+47600 val perplexity 345.0602
+47600 train 5.822316 (lr=7.3603e-06) (hash(x)=50430826)
+47300 val loss 5.6924
+47300 val perplexity 296.6056
+47300 train 5.513380 (lr=5.3256e-06) (hash(x)=51185601)
+43400 val loss 5.9105
+43400 val perplexity 368.9018
+43400 train 5.850887 (lr=1.3844e-05) (hash(x)=52858739)
+47700 val loss 5.8444
+47700 val perplexity 345.2808
+47700 train 5.769873 (lr=7.3310e-06) (hash(x)=51966931)
+47400 val loss 5.6915
+47400 val perplexity 296.3458
+47400 train 5.389669 (lr=5.3020e-06) (hash(x)=47743508)
+43500 val loss 5.9053
+43500 val perplexity 366.9831
+43500 train 5.984348 (lr=1.3730e-05) (hash(x)=49761319)
+47800 val loss 5.8448
+47800 val perplexity 345.4438
+47800 train 5.589402 (lr=7.3029e-06) (hash(x)=49517209)
+47500 val loss 5.6937
+47500 val perplexity 296.9857
+47500 train 5.459077 (lr=5.2792e-06) (hash(x)=49318833)
+43600 val loss 5.9012
+43600 val perplexity 365.4760
+43600 train 6.143999 (lr=1.3618e-05) (hash(x)=64879186)
+47900 val loss 5.8478
+47900 val perplexity 346.4768
+47900 train 5.651646 (lr=7.2760e-06) (hash(x)=49749702)
+47600 val loss 5.6958
+47600 val perplexity 297.6037
+47600 train 5.656210 (lr=5.2574e-06) (hash(x)=50430826)
+43700 val loss 5.9007
+43700 val perplexity 365.3025
+43700 train 6.098925 (lr=1.3507e-05) (hash(x)=55246427)
+48000 val loss 5.8472
+48000 val perplexity 346.2651
+48000 train 5.738358 (lr=7.2504e-06) (hash(x)=52535448)
+47700 val loss 5.6924
+47700 val perplexity 296.5990
+47700 train 5.622097 (lr=5.2364e-06) (hash(x)=51966931)
+43800 val loss 5.9012
+43800 val perplexity 365.4811
+43800 train 5.845619 (lr=1.3398e-05) (hash(x)=51238313)
+48100 val loss 5.8393
+48100 val perplexity 343.5267
+48100 train 6.504196 (lr=7.2260e-06) (hash(x)=53739181)
+47800 val loss 5.6929
+47800 val perplexity 296.7492
+47800 train 5.451366 (lr=5.2163e-06) (hash(x)=49517209)
+48200 val loss 5.8382
+48200 val perplexity 343.1531
+48200 train 5.818700 (lr=7.2029e-06) (hash(x)=48086710)
+43900 val loss 5.8992
+43900 val perplexity 364.7461
+43900 train 5.942101 (lr=1.3291e-05) (hash(x)=51664512)
+47900 val loss 5.6976
+47900 val perplexity 298.1566
+47900 train 5.511366 (lr=5.1972e-06) (hash(x)=49749702)
+48300 val loss 5.8370
+48300 val perplexity 342.7467
+48300 train 5.848775 (lr=7.1810e-06) (hash(x)=43991942)
+44000 val loss 5.9200
+44000 val perplexity 372.3936
+44000 train 6.194808 (lr=1.3185e-05) (hash(x)=64172296)
+48000 val loss 5.6964
+48000 val perplexity 297.7842
+48000 train 5.567257 (lr=5.1788e-06) (hash(x)=52535448)
+48400 val loss 5.8352
+48400 val perplexity 342.1305
+48400 train 5.733413 (lr=7.1603e-06) (hash(x)=50801906)
+48100 val loss 5.6909
+48100 val perplexity 296.1606
+48100 train 6.392676 (lr=5.1614e-06) (hash(x)=53739181)
+44100 val loss 5.8970
+44100 val perplexity 363.9280
+44100 train 5.794561 (lr=1.3081e-05) (hash(x)=50205960)
+48500 val loss 5.8337
+48500 val perplexity 341.6209
+48500 train 5.901326 (lr=7.1409e-06) (hash(x)=48915599)
+48200 val loss 5.6897
+48200 val perplexity 295.7967
+48200 train 5.674656 (lr=5.1449e-06) (hash(x)=48086710)
+44200 val loss 5.8998
+44200 val perplexity 364.9543
+44200 train 5.935834 (lr=1.2979e-05) (hash(x)=51065522)
+48600 val loss 5.8336
+48600 val perplexity 341.6013
+48600 train 5.971982 (lr=7.1228e-06) (hash(x)=54450172)
+48300 val loss 5.6892
+48300 val perplexity 295.6688
+48300 train 5.667455 (lr=5.1293e-06) (hash(x)=43991942)
+44300 val loss 5.8949
+44300 val perplexity 363.1784
+44300 train 5.992972 (lr=1.2878e-05) (hash(x)=56163183)
+48700 val loss 5.8338
+48700 val perplexity 341.6383
+48700 train 5.711917 (lr=7.1059e-06) (hash(x)=48755509)
+48400 val loss 5.6864
+48400 val perplexity 294.8432
+48400 train 5.564478 (lr=5.1145e-06) (hash(x)=50801906)
+44400 val loss 5.8974
+44400 val perplexity 364.0868
+44400 train 5.945497 (lr=1.2779e-05) (hash(x)=50818020)
+48800 val loss 5.8392
+48800 val perplexity 343.5165
+48800 train 5.730361 (lr=7.0902e-06) (hash(x)=51380096)
+48500 val loss 5.6856
+48500 val perplexity 294.5995
+48500 train 5.748676 (lr=5.1007e-06) (hash(x)=48915599)
+44500 val loss 5.8986
+44500 val perplexity 364.5092
+44500 train 6.022935 (lr=1.2682e-05) (hash(x)=52855559)
+48900 val loss 5.8356
+48900 val perplexity 342.2801
+48900 train 5.701429 (lr=7.0758e-06) (hash(x)=43404218)
+48600 val loss 5.6854
+48600 val perplexity 294.5244
+48600 train 5.860590 (lr=5.0877e-06) (hash(x)=54450172)
+44600 val loss 5.8999
+44600 val perplexity 364.9890
+44600 train 5.808752 (lr=1.2586e-05) (hash(x)=44014306)
+49000 val loss 5.8345
+49000 val perplexity 341.8860
+49000 train 5.789010 (lr=7.0627e-06) (hash(x)=55059739)
+48700 val loss 5.6858
+48700 val perplexity 294.6566
+48700 train 5.584652 (lr=5.0756e-06) (hash(x)=48755509)
+49100 val loss 5.8366
+49100 val perplexity 342.6208
+49100 train 6.049414 (lr=7.0508e-06) (hash(x)=47889309)
+44700 val loss 5.9001
+44700 val perplexity 365.0735
+44700 train 6.067726 (lr=1.2492e-05) (hash(x)=53648768)
+48800 val loss 5.6946
+48800 val perplexity 297.2485
+48800 train 5.575311 (lr=5.0644e-06) (hash(x)=51380096)
+49200 val loss 5.8351
+49200 val perplexity 342.0996
+49200 train 5.590982 (lr=7.0401e-06) (hash(x)=46202589)
+44800 val loss 5.8991
+44800 val perplexity 364.7090
+44800 train 5.813978 (lr=1.2400e-05) (hash(x)=42436477)
+48900 val loss 5.6870
+48900 val perplexity 295.0107
+48900 train 5.563878 (lr=5.0542e-06) (hash(x)=43404218)
+49300 val loss 5.8364
+49300 val perplexity 342.5402
+49300 train 5.780857 (lr=7.0307e-06) (hash(x)=49161813)
+49000 val loss 5.6857
+49000 val perplexity 294.6233
+49000 train 5.617337 (lr=5.0448e-06) (hash(x)=55059739)
+44900 val loss 5.9009
+44900 val perplexity 365.3680
+44900 train 5.626163 (lr=1.2309e-05) (hash(x)=48409149)
+49400 val loss 5.8356
+49400 val perplexity 342.2803
+49400 train 5.664212 (lr=7.0226e-06) (hash(x)=49505044)
+49100 val loss 5.6893
+49100 val perplexity 295.6787
+49100 train 5.932079 (lr=5.0363e-06) (hash(x)=47889309)
+49500 val loss 5.8357
+49500 val perplexity 342.3185
+49500 train 5.609909 (lr=7.0157e-06) (hash(x)=46610327)
+45000 val loss 5.9028
+45000 val perplexity 366.0562
+45000 train 5.809747 (lr=1.2220e-05) (hash(x)=53084805)
+49600 val loss 5.8377
+49600 val perplexity 342.9799
+49600 train 6.178606 (lr=7.0100e-06) (hash(x)=51188240)
+49200 val loss 5.6863
+49200 val perplexity 294.8148
+49200 train 5.418609 (lr=5.0286e-06) (hash(x)=46202589)
+45100 val loss 5.9046
+45100 val perplexity 366.7288
+45100 train 5.750527 (lr=1.2133e-05) (hash(x)=56386155)
+49700 val loss 5.8351
+49700 val perplexity 342.0913
+49700 train 5.623811 (lr=7.0056e-06) (hash(x)=41492016)
+49300 val loss 5.6890
+49300 val perplexity 295.5885
+49300 train 5.639915 (lr=5.0219e-06) (hash(x)=49161813)
+45200 val loss 5.9095
+45200 val perplexity 368.5215
+45200 train 5.916084 (lr=1.2047e-05) (hash(x)=53389864)
+49800 val loss 5.8371
+49800 val perplexity 342.7840
+49800 train 5.768189 (lr=7.0025e-06) (hash(x)=53488833)
+49400 val loss 5.6843
+49400 val perplexity 294.2065
+49400 train 5.520928 (lr=5.0161e-06) (hash(x)=49505044)
+49900 val loss 5.8337
+49900 val perplexity 341.6134
+45300 val loss 5.9062
+45300 val perplexity 367.2926
+49900 train 6.150034 (lr=7.0006e-06) (hash(x)=60260935)
+45300 train 5.941709 (lr=1.1964e-05) (hash(x)=50817456)
+49500 val loss 5.6846
+49500 val perplexity 294.3100
+49500 train 5.475948 (lr=5.0112e-06) (hash(x)=46610327)
+49999 val loss 5.8301
+49999 val perplexity 340.3875
+45400 val loss 5.9098
+45400 val perplexity 368.6502
+45400 train 5.938283 (lr=1.1881e-05) (hash(x)=50081660)
+49600 val loss 5.6854
+49600 val perplexity 294.5216
+49600 train 6.014633 (lr=5.0072e-06) (hash(x)=51188240)
+45500 val loss 5.9098
+45500 val perplexity 368.6245
+45500 train 5.674471 (lr=1.1801e-05) (hash(x)=55567121)
+49700 val loss 5.6844
+49700 val perplexity 294.2508
+49700 train 5.471631 (lr=5.0040e-06) (hash(x)=41492016)
+45600 val loss 5.9013
+45600 val perplexity 365.5045
+45600 train 6.000695 (lr=1.1722e-05) (hash(x)=48902854)
+49800 val loss 5.6848
+49800 val perplexity 294.3614
+49800 train 5.601730 (lr=5.0018e-06) (hash(x)=53488833)
+45700 val loss 5.8965
+45700 val perplexity 363.7566
+45700 train 6.110654 (lr=1.1645e-05) (hash(x)=53803301)
+49900 val loss 5.6843
+49900 val perplexity 294.2244
+49900 train 6.015575 (lr=5.0004e-06) (hash(x)=60260935)
+49999 val loss 5.6840
+49999 val perplexity 294.1094
+45800 val loss 5.8942
+45800 val perplexity 362.9217
+45800 train 5.937474 (lr=1.1570e-05) (hash(x)=47576606)
+45900 val loss 5.8893
+45900 val perplexity 361.1489
+45900 train 5.784353 (lr=1.1497e-05) (hash(x)=56276981)
+46000 val loss 5.8933
+46000 val perplexity 362.6179
+46000 train 5.853423 (lr=1.1425e-05) (hash(x)=52174902)
+46100 val loss 5.8956
+46100 val perplexity 363.4301
+46100 train 6.022799 (lr=1.1355e-05) (hash(x)=52155490)
+46200 val loss 5.8915
+46200 val perplexity 361.9449
+46200 train 5.770359 (lr=1.1287e-05) (hash(x)=46489984)
+46300 val loss 5.8893
+46300 val perplexity 361.1601
+46300 train 5.820558 (lr=1.1220e-05) (hash(x)=56603976)
+46400 val loss 5.8879
+46400 val perplexity 360.6450
+46400 train 5.693027 (lr=1.1155e-05) (hash(x)=51228412)
diff --git a/attention_kindselective_n_heads2_seed1341/model_02500.pt b/attention_kindselective_n_heads2_seed1341/model_02500.pt
index 9217d7d7c508a271c2ef26f11c1af80bb6dd2fc6..e128fb5f0e5060c751f21dca34ad0abed31c66b4 100644
--- a/attention_kindselective_n_heads2_seed1341/model_02500.pt
+++ b/attention_kindselective_n_heads2_seed1341/model_02500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f81c5305e28766ffe17b41d892c4195cc93a4d18969824c3a0f91c8db6cf9c61
+oid sha256:df2dbff4078c4b902e52c1b7e5066b645d7034b614be5e59574471f8960ecc8d
 size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_05000.pt b/attention_kindselective_n_heads2_seed1341/model_05000.pt
index c4919a9e9b85ef98ccf99b69d9c1ec4881bf90b0..2dbf6f890a8b65027cefce698f03da91fbe6321b 100644
--- a/attention_kindselective_n_heads2_seed1341/model_05000.pt
+++ b/attention_kindselective_n_heads2_seed1341/model_05000.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1981d296a42527ba71384319070268006aa51ed1ce3f34d492953367c7d791d
+oid sha256:2b8b2db67768b1b7c4f36a2f59bdbf2ce5cdb8709c2f1a8aab106c8f26270d14
 size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_07500.pt b/attention_kindselective_n_heads2_seed1341/model_07500.pt
index dd7276075cfe3e851160cff547d1bb9fa5e07893..3d3ed034710983a56c5018413de7b42c2c4ceebf 100644
--- a/attention_kindselective_n_heads2_seed1341/model_07500.pt
+++ b/attention_kindselective_n_heads2_seed1341/model_07500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1a38e1ba00e798ca0b0a8d8e9524e6b25782fda42839c6c43fc8de932eaaa2f8
+oid sha256:364f24f243b3307cbb71c22973ac7657870fb5d5ab6db86c6b6ce16187242859
 size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_10000.pt b/attention_kindselective_n_heads2_seed1341/model_10000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7768e426e2ce40d20d6c25b4a17533867d73742c
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_10000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39597fb5f366edff6ba2340cd9532f6013379941b9f4fa74758f88d9c97ae137
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_12500.pt b/attention_kindselective_n_heads2_seed1341/model_12500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f2d8055562d9eae03be94eeb78dbd00c35967ce1
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_12500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d31a60ee89698642b0ebc621b75f58a692275ef26044e594d7268780509fd84
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_15000.pt b/attention_kindselective_n_heads2_seed1341/model_15000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..6ddb1112a848772aff3eebcfaf07b576169a74ef
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_15000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1af66647de9c5fc855d02d637ba618c8144799d5efe8ddc623aa3b3de6f726c
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_17500.pt b/attention_kindselective_n_heads2_seed1341/model_17500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..07e657cd7cec48be9dd11fe49587e843245ea6f8
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_17500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e82ddeade8b883e1bec9ae78a0caacae82fc16e39d9c5e40a814ee4b400df35
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_20000.pt b/attention_kindselective_n_heads2_seed1341/model_20000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9d63abf174444af1582fc82f9409bfa6a80b137d
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_20000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be0d57d46256da655ab0efb88dce7b9010af64ff833c4c08ae94bcb3b7138cc8
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_22500.pt b/attention_kindselective_n_heads2_seed1341/model_22500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e625862ed12a6752eec8f48f9349472216417a32
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_22500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e93ca9b8574d2d714f7cae55c176cb3e64cb31d5ab43da6c34b0b03b0cbead07
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_25000.pt b/attention_kindselective_n_heads2_seed1341/model_25000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..041aac67fe2d152db5a666f8d0ae9b4de05c716c
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_25000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:454c04312ec7cba780576b7269421be5981dbee4e8a4d91baec0e8013e07cc02
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_27500.pt b/attention_kindselective_n_heads2_seed1341/model_27500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..94bebb488ffe11ca341559213409e460db771fa3
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_27500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8961199b8c739f6e36b00526c87a6b5246765f4d1c0dbe1093faede9f3344450
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_30000.pt b/attention_kindselective_n_heads2_seed1341/model_30000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a804ad2f601d5cb2173b101e7ed3c0ba428be30f
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_30000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:197046029793284598988dd04e5df15c623eaef6f155707df97b86154436a148
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_32500.pt b/attention_kindselective_n_heads2_seed1341/model_32500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e8451cdf53bfcf69a02c55d700c713289b262a35
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_32500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96771b15cd62ac259acfeaa8c3bea3c6a0311b32a69fc5f4cbdf396812aaa6f2
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_35000.pt b/attention_kindselective_n_heads2_seed1341/model_35000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2a8eddeb3043eba9e375d56da7a4f3e745c734d3
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_35000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64439ba3a2d2759349369c61c46122f5f8d43d494d81def5c4c784414d39443e
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_37500.pt b/attention_kindselective_n_heads2_seed1341/model_37500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..848f780b62b294916b918f86e004fb05f221b4e5
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_37500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:974dfee447c5c353591f60ad35df9d68980e9c6f37f94c1beca686db8af86a80
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_40000.pt b/attention_kindselective_n_heads2_seed1341/model_40000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..39cbdaf62cdec03962771bed5603c27a182774a9
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_40000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92e9a07a5d7a3a5bbe8777892da0e16dbe8d52cfe28c66d9263c00ba04a7a708
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_42500.pt b/attention_kindselective_n_heads2_seed1341/model_42500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f79511e67d3cc46e9186650ae50ddfc42b2b8eed
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_42500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff5074eedcb82cb39bd1c3cc5b79ff90173c0a1ed1cc59315acda50a131de0f9
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_45000.pt b/attention_kindselective_n_heads2_seed1341/model_45000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..37bf3b37beb0e770a42a976b2fcb94e3efaccccb
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_45000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd0be6b5f6053bea40260d382a66627f2572e0d23f93145bd41b03959b43fffb
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_47500.pt b/attention_kindselective_n_heads2_seed1341/model_47500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..cd22e4be43f645b4f6ab551ade7482587c37e375
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_47500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f10489eeeefc32b2aec0615aaaedae12a844b5c9ea48c80427d8873fd45e4bb
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/model_49999.pt b/attention_kindselective_n_heads2_seed1341/model_49999.pt
new file mode 100644
index 0000000000000000000000000000000000000000..0b55e017ce52f1d3d7c34b56104554815bfa743f
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/model_49999.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa2f3cec42d9036e2f87260a9e0795ade9b256a480f4f1261f7cffdf84d202db
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_02500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_02500.pt
index b50205beb5b70fc7eb0d7625b50f231b650a8000..96eff12e37a0283c681f965d61364100e0e3efcf 100644
--- a/attention_kindselective_n_heads2_seed1341/optimizer_02500.pt
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_02500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fbe352394c98f7aeb470becf95bc3d2fadc1b26eaace9a300ecb094670cdc944
+oid sha256:08f1a057ee80dea9056a043af4613f7d138e6c121717d5b995f40421c2185b5c
 size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_05000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_05000.pt
index 2c8d4350129d0bccf61e075c171c80b509385673..829295d728805c89ee7a9cc5df6097d9dee4c887 100644
--- a/attention_kindselective_n_heads2_seed1341/optimizer_05000.pt
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_05000.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1b8de7fcbdc412e17dba8f4280f46014e7504b0ac36b8318f82ce0046bc68161
+oid sha256:04707c9dc1a915ce8844444f88086b021a84ba76d2a51ff88b231c3a89315907
 size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_07500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_07500.pt
index 6c5f1b755a29075693b50215f564dfd504dc2c73..dd7b8999ff3012874a12316c62404aa5107d42c0 100644
--- a/attention_kindselective_n_heads2_seed1341/optimizer_07500.pt
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_07500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d1fdfd4cbdb5969528ced4c2cc8e4fab4c83f4ad557140cab2cdaecb053a8fc
+oid sha256:8bde31ee1de5e0b28258329d8e1bfc745047816434565ddeb118534f3321b382
 size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_10000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_10000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9021a85d0299df657d4d0a8a76af7e677d8b6e07
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_10000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e5e3a84f6d97d753668237ae05b7fa519625213870b1b982e72889351306dd8
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_12500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_12500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..95f42668d7b872639a13688bfce41b3bb68088d5
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_12500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7c7772956f413b64e0b2580b4b54fa679916524a74e477187123032eff1ae02
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_15000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_15000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..c94d881724e1311858f309163a2433d66cee4d12
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_15000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17b316437951257f7981be6620775537819e53492b0b26786967acf06aaba24f
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_17500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_17500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f5c6d3cfafaeacad4ba747c70a468c6e0a732a70
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_17500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37b5c575b4a0eceaad61e429a71cacabbb2b7157606d6b02c3a150f792a4a780
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_20000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_20000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a27632d8a3d6d1239703531df898c01a7caf49d1
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_20000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35fcaaa4f7c9226e4469c3c8c567557d0558d617c0ee65e7084bc28b23fcdb08
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_22500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_22500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..25e561a7781a4aa4f8e461915b3b083ab131330d
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_22500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01a74d2065d62b5f96d30821d64a6cb8d83df3adc4a127d9141f548214c01d10
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_25000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_25000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..12832baa649bba8703cabd7e23dbb435a9e53046
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_25000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f69c4fa5bd5d85abce331c063b8f8661ea0716d83655b11b5c59e07943ea8a7a
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_27500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_27500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f6e131ef25884c46ff005355c3e0c9baa413121a
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_27500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b93749c7b464e5db7708c3a19920f5e874f533484b74a8caea9b3b2c5f1598b
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_30000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_30000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..cab2a88b7932b35fdd041d9ff750a22bf9d2816a
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_30000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:392b205b5156664bfcdad140f817d7ee6bf4b37f656b83cb5ed96c9c2306f7c7
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_32500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_32500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..fe2ec2b89b15054b9ee2fbd6116f2670a7a61fb2
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_32500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fb0c0ea688e2d87f36438404ab30d956ddfb8f310e01cc4809cbc8ed4bcedd3
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_35000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_35000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..02f07cdb3f08403ec4357e7f1e740fe8dde67b87
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_35000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecb017bdba3cd06c70679a5a079b90587239f0df90789e9bd2aae6dd8ce5e1be
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_37500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_37500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..aa353573ed1f0ae28832570e692a65e069763bb1
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_37500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:610caea9a22bd843ba3bbc9a9b0c98cc8865d9e51f81e576865ccd2ee860315c
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_40000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_40000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9f27945f0686852a4fdcee9fada2a09bf3bd1afa
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_40000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c9fc6c8996145b248f831d6a856c99c569135e92fdadb4a9f8c234776bbb338
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_42500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_42500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..debe7c68f9f9a3eddb6c911073c09e70aa3caf85
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_42500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ce5b637f0f687de66c681ad9bf5b28775497b12beed5490ae4e2f3ef9fa2239
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_45000.pt b/attention_kindselective_n_heads2_seed1341/optimizer_45000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a5caa6ccc404af48d0c4f4bba571ac4011ca6635
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_45000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e219bed0d1290d31e6fe1c5c0024faf85f4d71699c8592408b56e2059cfb5f7
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_47500.pt b/attention_kindselective_n_heads2_seed1341/optimizer_47500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..72d11234bc5444037b42f96133505ed2410d4200
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_47500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40db19e0e2056eef8cf8b772f3c21365474b8d9d8725dee8528b75cd13754308
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1341/optimizer_49999.pt b/attention_kindselective_n_heads2_seed1341/optimizer_49999.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b4361050de0851b51e9f4018f710248459d78076
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1341/optimizer_49999.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbd88d144afa2cebd143ee7b83af343275bd2fb3add7cc3461fd7b3376010628
+size 70895430