diff --git a/attention_kindselective_n_heads2_seed1340/args.json b/attention_kindselective_n_heads2_seed1340/args.json
index bedd11d4fc6aa6d3ca8a5a66f58869ea46e39631..486439fdcb6d55a60fded08488769c0112891b42 100644
--- a/attention_kindselective_n_heads2_seed1340/args.json
+++ b/attention_kindselective_n_heads2_seed1340/args.json
@@ -1 +1 @@
-{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_6/attention_kindselective_n_heads2_seed1340", "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": 1340, "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_1340", "n_embd": 128}
\ No newline at end of file
+{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads2_seed1340", "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": 1340, "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_1340", "n_embd": 128}
\ No newline at end of file
diff --git a/attention_kindselective_n_heads2_seed1340/dataloader_10000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_10000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e1452fd777edd0ccb65d6b47710c50208c14b312
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_12500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_12500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f63806424a0177a7f2d678c2c63138219ed021f3
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_15000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_15000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..0ea0f5a1bfab75667c4ebb0ca01b358cdc836a54
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_17500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_17500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..eb9392348b1209c827e3e376b05eeda80e779aa8
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_20000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_20000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..ba32fca059aec40e4f758de96bdac6df23b9d9f5
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_22500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_22500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..498dc444f528d893090328a5bd1e2f37da46dc12
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_25000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_25000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b657bc134192d0ea956f984c289d0c682979a1f4
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_27500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_27500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d92d43d89390714f43db4f0782e49af0145b4a90
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_30000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_30000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..06856f53253b6c8cbefd7d595d9b9b7266b22621
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_32500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_32500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..dc0b4b40e57f41aa1046b3bd2697256635160c09
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_35000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_35000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..62647b540fa7925361626b9f8dfa3959eebb7608
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_37500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_37500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..475d8b538138a8e39b76a4cf04c8eaeac074d295
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_40000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_40000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..50ec00ef6be330bbdb4cdf88e9a1097345da0d4d
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_42500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_42500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..74e133f73a7293d5f4d6407784703c91f705d6e3
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_45000.pt b/attention_kindselective_n_heads2_seed1340/dataloader_45000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..718bc149d695b1e9498bdd0693053d7417207818
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_47500.pt b/attention_kindselective_n_heads2_seed1340/dataloader_47500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..3eab97c7c6d4b86d90405bf1c4f3435727495da4
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/dataloader_49999.pt b/attention_kindselective_n_heads2_seed1340/dataloader_49999.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d87f88b62a343a49411f8a6feee8f527879fcd1f
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/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_seed1340/log2.txt b/attention_kindselective_n_heads2_seed1340/log2.txt
index 480b785bf77977a380b294b46e4dc53ece89b256..67909bd01780a3fac2463f518e7d2703a81257c4 100644
--- a/attention_kindselective_n_heads2_seed1340/log2.txt
+++ b/attention_kindselective_n_heads2_seed1340/log2.txt
@@ -1,515 +1,4469 @@
-max_steps: 10000
-3000 val loss 6.8497
-3000 val perplexity 943.5969
-3000 train 6.557367 (lr=8.3057e-05) (hash(x)=44336167)
+max_steps: 50000
 0 val loss 11.8201
 0 val perplexity 135955.4688
-3100 val loss 6.8108
-3100 val perplexity 907.6118
-3100 train 7.108453 (lr=8.1915e-05) (hash(x)=44479330)
-3200 val loss 6.8168
-3200 val perplexity 913.0430
-3200 train 6.861060 (lr=8.0745e-05) (hash(x)=54593096)
-3300 val loss 6.7666
-3300 val perplexity 868.3719
-3300 train 6.689936 (lr=7.9549e-05) (hash(x)=45347643)
-0 train 11.815376 (lr=7.5000e-07) (hash(x)=57791809)
-3400 val loss 6.7480
-3400 val perplexity 852.3517
-3400 train 6.868625 (lr=7.8328e-05) (hash(x)=47797247)
-100 val loss 9.4103
-100 val perplexity 12214.1318
-100 train 9.435302 (lr=7.5750e-05) (hash(x)=48211824)
-3500 val loss 6.7277
-3500 val perplexity 835.2067
-3500 train 6.586505 (lr=7.7082e-05) (hash(x)=46115683)
-200 val loss 7.8489
-200 val perplexity 2562.9167
-200 train 7.782820 (lr=1.5000e-04) (hash(x)=50375849)
-3600 val loss 6.7276
-3600 val perplexity 835.1287
-3600 train 6.583769 (lr=7.5814e-05) (hash(x)=44502074)
-300 val loss 7.7186
-300 val perplexity 2249.8511
-300 train 8.037295 (lr=1.4997e-04) (hash(x)=57250808)
-3700 val loss 6.6958
-3700 val perplexity 808.9891
-3700 train 6.847288 (lr=7.4525e-05) (hash(x)=55388443)
-3800 val loss 6.6831
-3800 val perplexity 798.8195
-3800 train 6.478742 (lr=7.3215e-05) (hash(x)=43790341)
-400 val loss 7.6896
-400 val perplexity 2185.4397
-400 train 8.281481 (lr=1.4986e-04) (hash(x)=62519858)
-3900 val loss 6.6852
-3900 val perplexity 800.4603
-3900 train 6.680488 (lr=7.1887e-05) (hash(x)=50013318)
-500 val loss 7.6498
-500 val perplexity 2100.2070
-500 train 7.555541 (lr=1.4969e-04) (hash(x)=47226806)
-600 val loss 7.6013
-600 val perplexity 2000.8110
-600 train 7.611244 (lr=1.4945e-04) (hash(x)=51149322)
-4000 val loss 6.6633
-4000 val perplexity 783.1157
-4000 train 6.724937 (lr=7.0541e-05) (hash(x)=51704787)
-700 val loss 7.5505
-700 val perplexity 1901.7231
-700 train 7.570621 (lr=1.4913e-04) (hash(x)=51564551)
-4100 val loss 6.6502
-4100 val perplexity 772.9760
-4100 train 6.690805 (lr=6.9180e-05) (hash(x)=50821964)
-800 val loss 7.4868
-800 val perplexity 1784.4075
-800 train 7.224887 (lr=1.4876e-04) (hash(x)=45093459)
-4200 val loss 6.6401
-4200 val perplexity 765.1970
-4200 train 6.676926 (lr=6.7804e-05) (hash(x)=49675080)
-900 val loss 7.4101
-900 val perplexity 1652.5939
-900 train 7.676486 (lr=1.4831e-04) (hash(x)=54988361)
-4300 val loss 6.6300
-4300 val perplexity 757.5169
-4300 train 6.297640 (lr=6.6414e-05) (hash(x)=43239281)
-1000 val loss 7.3711
-1000 val perplexity 1589.3545
-1000 train 7.263645 (lr=1.4779e-04) (hash(x)=47588648)
-4400 val loss 6.6187
-4400 val perplexity 748.9337
-4400 train 6.297700 (lr=6.5013e-05) (hash(x)=45076737)
-1100 val loss 7.3459
-1100 val perplexity 1549.8197
-1100 train 6.974550 (lr=1.4721e-04) (hash(x)=37984588)
-4500 val loss 6.6069
-4500 val perplexity 740.1642
-4500 train 6.723946 (lr=6.3602e-05) (hash(x)=57930262)
-1200 val loss 7.2713
-1200 val perplexity 1438.3854
-1200 train 7.423686 (lr=1.4656e-04) (hash(x)=56333817)
-4600 val loss 6.5956
-4600 val perplexity 731.8596
-4600 train 6.346147 (lr=6.2182e-05) (hash(x)=46721614)
-1300 val loss 7.2478
-1300 val perplexity 1405.0525
-1300 train 7.408738 (lr=1.4585e-04) (hash(x)=53454056)
-4700 val loss 6.5840
-4700 val perplexity 723.4270
-4700 train 6.413177 (lr=6.0754e-05) (hash(x)=49837920)
-1400 val loss 7.1724
-1400 val perplexity 1302.9978
-1400 train 7.307054 (lr=1.4507e-04) (hash(x)=55284163)
-4800 val loss 6.5830
-4800 val perplexity 722.7350
-4800 train 6.747669 (lr=5.9321e-05) (hash(x)=48380045)
-1500 val loss 7.1053
-1500 val perplexity 1218.3500
-1500 train 6.960086 (lr=1.4422e-04) (hash(x)=48162598)
-4900 val loss 6.5705
-4900 val perplexity 713.7289
-4900 train 6.455478 (lr=5.7883e-05) (hash(x)=44202577)
-1600 val loss 7.0936
-1600 val perplexity 1204.2488
-1600 train 7.176346 (lr=1.4332e-04) (hash(x)=54214535)
-5000 val loss 6.5548
-5000 val perplexity 702.6354
-5000 train 6.531763 (lr=5.6442e-05) (hash(x)=52038024)
-1700 val loss 7.0477
-1700 val perplexity 1150.2354
-1700 train 7.288129 (lr=1.4235e-04) (hash(x)=53525003)
-5100 val loss 6.5472
-5100 val perplexity 697.2672
-5100 train 6.682254 (lr=5.5000e-05) (hash(x)=53700038)
-1800 val loss 6.9886
-1800 val perplexity 1084.1967
-1800 train 7.055763 (lr=1.4131e-04) (hash(x)=51848994)
-5200 val loss 6.5449
-5200 val perplexity 695.6867
-5200 train 6.514779 (lr=5.3558e-05) (hash(x)=48137625)
-1900 val loss 6.9741
-1900 val perplexity 1068.6400
-1900 train 6.808705 (lr=1.4022e-04) (hash(x)=48405987)
-5300 val loss 6.5362
-5300 val perplexity 689.6800
-5300 train 6.367496 (lr=5.2117e-05) (hash(x)=43161573)
-2000 val loss 6.9172
-2000 val perplexity 1009.4669
-2000 train 7.280006 (lr=1.3907e-04) (hash(x)=58592291)
-2100 val loss 6.9460
-2100 val perplexity 1038.9731
-2100 train 7.032774 (lr=1.3786e-04) (hash(x)=51167081)
-5400 val loss 6.5369
-5400 val perplexity 690.1711
-5400 train 6.705731 (lr=5.0679e-05) (hash(x)=56673322)
-2200 val loss 6.8788
-2200 val perplexity 971.4705
-2200 train 6.855141 (lr=1.3660e-04) (hash(x)=47994988)
-5500 val loss 6.5188
-5500 val perplexity 677.7745
-5500 train 6.729006 (lr=4.9246e-05) (hash(x)=53468295)
-2300 val loss 6.9196
-2300 val perplexity 1011.8920
-2300 train 6.912843 (lr=1.3527e-04) (hash(x)=47377604)
-5600 val loss 6.5010
-5600 val perplexity 665.8105
-5600 train 6.796459 (lr=4.7818e-05) (hash(x)=59287280)
-2400 val loss 6.8704
-2400 val perplexity 963.3668
-2400 train 6.928603 (lr=1.3390e-04) (hash(x)=53554323)
-5700 val loss 6.5016
-5700 val perplexity 666.2183
-5700 train 6.710692 (lr=4.6398e-05) (hash(x)=57575806)
-2500 val loss 6.8641
-2500 val perplexity 957.2761
-2500 train 6.931783 (lr=1.3247e-04) (hash(x)=50780417)
-5800 val loss 6.4968
-5800 val perplexity 663.0387
-5800 train 6.534732 (lr=4.4987e-05) (hash(x)=46897279)
-2600 val loss 6.8701
-2600 val perplexity 963.0816
-2600 train 6.758279 (lr=1.3099e-04) (hash(x)=46453562)
-5900 val loss 6.4894
-5900 val perplexity 658.1490
-5900 train 6.375528 (lr=4.3586e-05) (hash(x)=47565679)
-2700 val loss 6.8107
-2700 val perplexity 907.5153
-2700 train 6.781026 (lr=1.2946e-04) (hash(x)=54404221)
-6000 val loss 6.4841
-6000 val perplexity 654.6342
-6000 train 6.333650 (lr=4.2196e-05) (hash(x)=51590090)
-2800 val loss 6.7924
-2800 val perplexity 891.0690
-2800 train 7.439435 (lr=1.2788e-04) (hash(x)=59318895)
-6100 val loss 6.4780
-6100 val perplexity 650.6495
-6100 train 6.915969 (lr=4.0820e-05) (hash(x)=59732271)
-2900 val loss 6.7875
-2900 val perplexity 886.6631
-2900 train 6.780449 (lr=1.2626e-04) (hash(x)=47845760)
-6200 val loss 6.4670
-6200 val perplexity 643.5385
-6200 train 6.530252 (lr=3.9459e-05) (hash(x)=46394422)
-3000 val loss 6.7641
-3000 val perplexity 866.1562
-3000 train 6.461198 (lr=1.2459e-04) (hash(x)=44336167)
-6300 val loss 6.4697
-6300 val perplexity 645.2667
-6300 train 6.481519 (lr=3.8113e-05) (hash(x)=53748145)
-3100 val loss 6.7478
-3100 val perplexity 852.1684
-3100 train 7.055303 (lr=1.2287e-04) (hash(x)=44479330)
-6400 val loss 6.4581
-6400 val perplexity 637.8579
-6400 train 6.312898 (lr=3.6785e-05) (hash(x)=46054751)
-3200 val loss 6.7569
-3200 val perplexity 859.9979
-3200 train 6.806598 (lr=1.2112e-04) (hash(x)=54593096)
-6500 val loss 6.4519
-6500 val perplexity 633.8904
-6500 train 6.736262 (lr=3.5475e-05) (hash(x)=51816809)
-3300 val loss 6.7278
-3300 val perplexity 835.3075
-3300 train 6.684790 (lr=1.1932e-04) (hash(x)=45347643)
-6600 val loss 6.4588
-6600 val perplexity 638.2906
-6600 train 6.274892 (lr=3.4186e-05) (hash(x)=52453336)
-3400 val loss 6.7000
-3400 val perplexity 812.3824
-3400 train 6.819838 (lr=1.1749e-04) (hash(x)=47797247)
-6700 val loss 6.4472
-6700 val perplexity 630.9039
-6700 train 6.439846 (lr=3.2918e-05) (hash(x)=49108775)
-3500 val loss 6.6888
-3500 val perplexity 803.3629
-3500 train 6.543953 (lr=1.1562e-04) (hash(x)=46115683)
-6800 val loss 6.4345
-6800 val perplexity 622.9869
-6800 train 6.340245 (lr=3.1672e-05) (hash(x)=46745396)
-3600 val loss 6.6714
-3600 val perplexity 789.5205
-3600 train 6.545329 (lr=1.1372e-04) (hash(x)=44502074)
-6900 val loss 6.4267
-6900 val perplexity 618.1205
-6900 train 6.602197 (lr=3.0451e-05) (hash(x)=46534986)
-3700 val loss 6.6690
-3700 val perplexity 787.6163
-3700 train 6.809144 (lr=1.1179e-04) (hash(x)=55388443)
-7000 val loss 6.4191
-7000 val perplexity 613.4761
-7000 train 6.785058 (lr=2.9255e-05) (hash(x)=49317888)
-3800 val loss 6.6656
-3800 val perplexity 784.8956
-3800 train 6.453290 (lr=1.0982e-04) (hash(x)=43790341)
-7100 val loss 6.4132
-7100 val perplexity 609.8173
-7100 train 6.441988 (lr=2.8085e-05) (hash(x)=50360484)
-3900 val loss 6.6528
-3900 val perplexity 774.9131
-3900 train 6.637487 (lr=1.0783e-04) (hash(x)=50013318)
-7200 val loss 6.4056
-7200 val perplexity 605.1969
-7200 train 6.250135 (lr=2.6943e-05) (hash(x)=49515094)
-4000 val loss 6.6511
-4000 val perplexity 773.6136
-4000 train 6.725606 (lr=1.0581e-04) (hash(x)=51704787)
-7300 val loss 6.4160
-7300 val perplexity 611.5779
-7300 train 6.397116 (lr=2.5830e-05) (hash(x)=51546861)
-4100 val loss 6.6422
-4100 val perplexity 766.7709
-4100 train 6.687906 (lr=1.0377e-04) (hash(x)=50821964)
-7400 val loss 6.4147
-7400 val perplexity 610.7546
-7400 train 6.393087 (lr=2.4746e-05) (hash(x)=48320948)
-4200 val loss 6.6313
-4200 val perplexity 758.4469
-4200 train 6.676696 (lr=1.0171e-04) (hash(x)=49675080)
-7500 val loss 6.4144
-7500 val perplexity 610.5697
-7500 train 6.344778 (lr=2.3694e-05) (hash(x)=40167457)
-4300 val loss 6.6157
-4300 val perplexity 746.7346
-4300 train 6.275900 (lr=9.9622e-05) (hash(x)=43239281)
-7600 val loss 6.3956
-7600 val perplexity 599.2078
-7600 train 6.340053 (lr=2.2674e-05) (hash(x)=49942165)
-4400 val loss 6.6233
-4400 val perplexity 752.4334
-4400 train 6.317001 (lr=9.7520e-05) (hash(x)=45076737)
-7700 val loss 6.3912
-7700 val perplexity 596.5556
-7700 train 6.109773 (lr=2.1687e-05) (hash(x)=48853311)
-4500 val loss 6.6173
-4500 val perplexity 747.8881
-4500 train 6.730391 (lr=9.5403e-05) (hash(x)=57930262)
-7800 val loss 6.3869
-7800 val perplexity 594.0394
-7800 train 6.283017 (lr=2.0735e-05) (hash(x)=48510117)
-4600 val loss 6.5940
-4600 val perplexity 730.6702
-4600 train 6.354017 (lr=9.3273e-05) (hash(x)=46721614)
-7900 val loss 6.3907
-7900 val perplexity 596.3019
-7900 train 6.387419 (lr=1.9818e-05) (hash(x)=48339781)
-4700 val loss 6.5861
-4700 val perplexity 724.9485
-4700 train 6.410762 (lr=9.1132e-05) (hash(x)=49837920)
-8000 val loss 6.3852
-8000 val perplexity 593.0266
-8000 train 6.477322 (lr=1.8936e-05) (hash(x)=54927320)
-4800 val loss 6.5926
-4800 val perplexity 729.6981
-4800 train 6.733562 (lr=8.8982e-05) (hash(x)=48380045)
-8100 val loss 6.3803
-8100 val perplexity 590.0946
-8100 train 6.103539 (lr=1.8092e-05) (hash(x)=46461786)
-4900 val loss 6.5911
-4900 val perplexity 728.6015
-4900 train 6.470277 (lr=8.6825e-05) (hash(x)=44202577)
-8200 val loss 6.3766
-8200 val perplexity 587.9359
-8200 train 6.291387 (lr=1.7286e-05) (hash(x)=51536260)
-5000 val loss 6.5647
-5000 val perplexity 709.6234
-5000 train 6.568296 (lr=8.4663e-05) (hash(x)=52038024)
-8300 val loss 6.3755
-8300 val perplexity 587.3049
-8300 train 6.118826 (lr=1.6519e-05) (hash(x)=44770722)
-5100 val loss 6.5647
-5100 val perplexity 709.6227
-5100 train 6.695176 (lr=8.2500e-05) (hash(x)=53700038)
-8400 val loss 6.3739
-8400 val perplexity 586.3613
-8400 train 6.378792 (lr=1.5791e-05) (hash(x)=50104957)
-5200 val loss 6.5524
-5200 val perplexity 700.9124
-5200 train 6.519544 (lr=8.0337e-05) (hash(x)=48137625)
-8500 val loss 6.3685
-8500 val perplexity 583.1942
-8500 train 6.459326 (lr=1.5103e-05) (hash(x)=50132971)
-5300 val loss 6.5438
-5300 val perplexity 694.8887
-5300 train 6.366642 (lr=7.8175e-05) (hash(x)=43161573)
-8600 val loss 6.3651
-8600 val perplexity 581.2270
-8600 train 6.323270 (lr=1.4456e-05) (hash(x)=52193699)
-5400 val loss 6.5453
-5400 val perplexity 695.9648
-5400 train 6.720807 (lr=7.6018e-05) (hash(x)=56673322)
-8700 val loss 6.3602
-8700 val perplexity 578.3399
-8700 train 6.344991 (lr=1.3851e-05) (hash(x)=47902319)
-5500 val loss 6.5354
-5500 val perplexity 689.1136
-5500 train 6.723743 (lr=7.3868e-05) (hash(x)=53468295)
-8800 val loss 6.3528
-8800 val perplexity 574.0699
-8800 train 6.642093 (lr=1.3289e-05) (hash(x)=54904230)
-5600 val loss 6.5390
-5600 val perplexity 691.6022
-5600 train 6.854773 (lr=7.1727e-05) (hash(x)=59287280)
-8900 val loss 6.3538
-8900 val perplexity 574.6966
-8900 train 6.230078 (lr=1.2769e-05) (hash(x)=46311615)
-5700 val loss 6.5331
-5700 val perplexity 687.5286
-5700 train 6.746007 (lr=6.9597e-05) (hash(x)=57575806)
-9000 val loss 6.3527
-9000 val perplexity 574.0275
-9000 train 6.200503 (lr=1.2292e-05) (hash(x)=48535188)
-5800 val loss 6.5352
-5800 val perplexity 688.9601
-5800 train 6.542769 (lr=6.7480e-05) (hash(x)=46897279)
-9100 val loss 6.3477
-9100 val perplexity 571.1558
-9100 train 6.375886 (lr=1.1860e-05) (hash(x)=51757372)
-5900 val loss 6.5230
-5900 val perplexity 680.6183
-5900 train 6.415868 (lr=6.5378e-05) (hash(x)=47565679)
-9200 val loss 6.3460
-9200 val perplexity 570.2044
-9200 train 6.207469 (lr=1.1472e-05) (hash(x)=51131708)
-6000 val loss 6.5250
-6000 val perplexity 682.0081
-6000 train 6.372841 (lr=6.3294e-05) (hash(x)=51590090)
-9300 val loss 6.3423
-9300 val perplexity 568.0868
-9300 train 6.299280 (lr=1.1128e-05) (hash(x)=44784276)
-6100 val loss 6.5130
-6100 val perplexity 673.8539
-6100 train 6.930476 (lr=6.1230e-05) (hash(x)=59732271)
-9400 val loss 6.3432
-9400 val perplexity 568.6069
-9400 train 6.450517 (lr=1.0830e-05) (hash(x)=51981169)
-6200 val loss 6.5090
-6200 val perplexity 671.1452
-9500 val loss 6.3407
-9500 val perplexity 567.1857
-6200 train 6.571561 (lr=5.9188e-05) (hash(x)=46394422)
-9500 train 6.312754 (lr=1.0577e-05) (hash(x)=47232936)
-9600 val loss 6.3410
-9600 val perplexity 567.3699
-9600 train 6.355495 (lr=1.0369e-05) (hash(x)=53800450)
-6300 val loss 6.5214
-6300 val perplexity 679.5251
-6300 train 6.526929 (lr=5.7169e-05) (hash(x)=53748145)
-9700 val loss 6.3406
-9700 val perplexity 567.1387
-9700 train 6.445551 (lr=1.0208e-05) (hash(x)=55768123)
-6400 val loss 6.5170
-6400 val perplexity 676.5420
-6400 train 6.375550 (lr=5.5177e-05) (hash(x)=46054751)
-9800 val loss 6.3420
-9800 val perplexity 567.9086
-9800 train 6.281539 (lr=1.0092e-05) (hash(x)=47745177)
-6500 val loss 6.5060
-6500 val perplexity 669.1499
-6500 train 6.733078 (lr=5.3213e-05) (hash(x)=51816809)
-9900 val loss 6.3431
-9900 val perplexity 568.5347
-9900 train 6.613006 (lr=1.0023e-05) (hash(x)=56592246)
-6600 val loss 6.4993
-6600 val perplexity 664.6857
-6600 train 6.381551 (lr=5.1279e-05) (hash(x)=52453336)
-9999 val loss 6.3466
-9999 val perplexity 570.5632
-6700 val loss 6.5058
-6700 val perplexity 669.0009
-6700 train 6.497029 (lr=4.9377e-05) (hash(x)=49108775)
-6800 val loss 6.4920
-6800 val perplexity 659.8412
-6800 train 6.398888 (lr=4.7509e-05) (hash(x)=46745396)
-6900 val loss 6.4899
-6900 val perplexity 658.4554
-6900 train 6.663884 (lr=4.5676e-05) (hash(x)=46534986)
-7000 val loss 6.4768
-7000 val perplexity 649.8740
-7000 train 6.840258 (lr=4.3882e-05) (hash(x)=49317888)
-7100 val loss 6.4738
-7100 val perplexity 647.9160
-7100 train 6.493526 (lr=4.2128e-05) (hash(x)=50360484)
-7200 val loss 6.4698
-7200 val perplexity 645.3282
-7200 train 6.339218 (lr=4.0414e-05) (hash(x)=49515094)
-7300 val loss 6.4730
-7300 val perplexity 647.4099
-7300 train 6.455276 (lr=3.8745e-05) (hash(x)=51546861)
-7400 val loss 6.4622
-7400 val perplexity 640.4369
-7400 train 6.428791 (lr=3.7120e-05) (hash(x)=48320948)
-7500 val loss 6.4678
-7500 val perplexity 644.0776
-7500 train 6.414234 (lr=3.5541e-05) (hash(x)=40167457)
-7600 val loss 6.4528
-7600 val perplexity 634.4956
-7600 train 6.400621 (lr=3.4011e-05) (hash(x)=49942165)
-7700 val loss 6.4515
-7700 val perplexity 633.6224
-7700 train 6.178792 (lr=3.2531e-05) (hash(x)=48853311)
-7800 val loss 6.4469
-7800 val perplexity 630.7529
-7800 train 6.332662 (lr=3.1102e-05) (hash(x)=48510117)
-7900 val loss 6.4452
-7900 val perplexity 629.6969
-7900 train 6.454893 (lr=2.9726e-05) (hash(x)=48339781)
-8000 val loss 6.4485
-8000 val perplexity 631.7830
-8000 train 6.561584 (lr=2.8405e-05) (hash(x)=54927320)
-8100 val loss 6.4404
-8100 val perplexity 626.6378
-8100 train 6.167403 (lr=2.7138e-05) (hash(x)=46461786)
-8200 val loss 6.4388
-8200 val perplexity 625.6752
-8200 train 6.347755 (lr=2.5929e-05) (hash(x)=51536260)
-8300 val loss 6.4365
-8300 val perplexity 624.2048
-8300 train 6.190253 (lr=2.4778e-05) (hash(x)=44770722)
-8400 val loss 6.4290
-8400 val perplexity 619.5564
-8400 train 6.434593 (lr=2.3686e-05) (hash(x)=50104957)
-8500 val loss 6.4245
-8500 val perplexity 616.7615
-8500 train 6.533162 (lr=2.2655e-05) (hash(x)=50132971)
-8600 val loss 6.4212
-8600 val perplexity 614.7573
-8600 train 6.378406 (lr=2.1685e-05) (hash(x)=52193699)
-8700 val loss 6.4183
-8700 val perplexity 612.9755
-8700 train 6.427931 (lr=2.0777e-05) (hash(x)=47902319)
-8800 val loss 6.4139
-8800 val perplexity 610.2626
-8800 train 6.709031 (lr=1.9933e-05) (hash(x)=54904230)
-8900 val loss 6.4051
-8900 val perplexity 604.9390
-8900 train 6.288378 (lr=1.9153e-05) (hash(x)=46311615)
-9000 val loss 6.4113
-9000 val perplexity 608.7029
-9000 train 6.270270 (lr=1.8439e-05) (hash(x)=48535188)
-9100 val loss 6.4026
-9100 val perplexity 603.3902
-9100 train 6.419470 (lr=1.7790e-05) (hash(x)=51757372)
-9200 val loss 6.4024
-9200 val perplexity 603.2843
-9200 train 6.288178 (lr=1.7208e-05) (hash(x)=51131708)
-9300 val loss 6.4021
-9300 val perplexity 603.1321
-9300 train 6.361725 (lr=1.6692e-05) (hash(x)=44784276)
-9400 val loss 6.4006
-9400 val perplexity 602.2206
-9400 train 6.529187 (lr=1.6245e-05) (hash(x)=51981169)
-9500 val loss 6.3956
-9500 val perplexity 599.1955
-9500 train 6.358074 (lr=1.5865e-05) (hash(x)=47232936)
-9600 val loss 6.3954
-9600 val perplexity 599.1069
-9600 train 6.410747 (lr=1.5554e-05) (hash(x)=53800450)
-9700 val loss 6.3937
-9700 val perplexity 598.0549
-9700 train 6.500488 (lr=1.5312e-05) (hash(x)=55768123)
-9800 val loss 6.3977
-9800 val perplexity 600.4852
-9800 train 6.321241 (lr=1.5139e-05) (hash(x)=47745177)
-9900 val loss 6.3977
-9900 val perplexity 600.4746
-9900 train 6.664391 (lr=1.5035e-05) (hash(x)=56592246)
-9999 val loss 6.4009
-9999 val perplexity 602.3967
+0 val loss 11.8201
+0 val perplexity 135955.4688
+0 val loss 11.8201
+0 val perplexity 135955.4688
+0 train 11.815376 (lr=3.5000e-07) (hash(x)=57791809)
+0 train 11.815376 (lr=2.5000e-07) (hash(x)=57791809)
+0 train 11.815376 (lr=5.0000e-07) (hash(x)=57791809)
+100 val loss 10.0061
+100 val perplexity 22160.5547
+100 train 10.031336 (lr=3.5350e-05) (hash(x)=48211824)
+100 val loss 10.1395
+100 val perplexity 25324.7344
+100 train 10.163652 (lr=2.5250e-05) (hash(x)=48211824)
+100 val loss 9.7609
+100 val perplexity 17342.8867
+100 train 9.788915 (lr=5.0500e-05) (hash(x)=48211824)
+200 val loss 8.7280
+200 val perplexity 6173.5259
+200 train 8.680338 (lr=7.0000e-05) (hash(x)=50375849)
+200 val loss 8.9196
+200 val perplexity 7477.1943
+200 train 8.881001 (lr=5.0000e-05) (hash(x)=50375849)
+200 val loss 7.9945
+200 val perplexity 2964.6406
+200 train 7.936388 (lr=1.0000e-04) (hash(x)=50375849)
+300 val loss 7.9588
+300 val perplexity 2860.6704
+300 train 8.242803 (lr=6.9999e-05) (hash(x)=57250808)
+300 val loss 8.0427
+300 val perplexity 3111.0901
+300 train 8.321117 (lr=5.0000e-05) (hash(x)=57250808)
+300 val loss 7.6873
+300 val perplexity 2180.5630
+300 train 8.001595 (lr=9.9999e-05) (hash(x)=57250808)
+400 val loss 7.7226
+400 val perplexity 2258.8765
+400 train 8.318707 (lr=6.9997e-05) (hash(x)=62519858)
+400 val loss 7.7574
+400 val perplexity 2338.7456
+400 train 8.329361 (lr=4.9998e-05) (hash(x)=62519858)
+400 val loss 7.6750
+400 val perplexity 2153.8877
+400 train 8.293846 (lr=9.9996e-05) (hash(x)=62519858)
+500 val loss 7.6695
+500 val perplexity 2142.0701
+500 train 7.568311 (lr=6.9994e-05) (hash(x)=47226806)
+500 val loss 7.6585
+500 val perplexity 2118.6331
+500 train 7.559607 (lr=4.9996e-05) (hash(x)=47226806)
+500 val loss 7.6635
+500 val perplexity 2129.2957
+500 train 7.589068 (lr=9.9992e-05) (hash(x)=47226806)
+600 val loss 7.6428
+600 val perplexity 2085.4954
+600 train 7.671701 (lr=6.9990e-05) (hash(x)=51149322)
+600 val loss 7.6290
+600 val perplexity 2057.0649
+600 train 7.657920 (lr=4.9993e-05) (hash(x)=51149322)
+600 val loss 7.6835
+600 val perplexity 2172.2947
+600 train 7.712746 (lr=9.9986e-05) (hash(x)=51149322)
+700 val loss 7.6430
+700 val perplexity 2085.8882
+700 train 7.653198 (lr=6.9984e-05) (hash(x)=51564551)
+700 val loss 7.6192
+700 val perplexity 2036.8857
+700 train 7.630872 (lr=4.9989e-05) (hash(x)=51564551)
+700 val loss 7.6578
+700 val perplexity 2117.1748
+700 train 7.676233 (lr=9.9978e-05) (hash(x)=51564551)
+800 val loss 7.6202
+800 val perplexity 2039.0431
+800 train 7.373461 (lr=6.9977e-05) (hash(x)=45093459)
+800 val loss 7.6100
+800 val perplexity 2018.2389
+800 train 7.380928 (lr=4.9984e-05) (hash(x)=45093459)
+800 val loss 7.6409
+800 val perplexity 2081.5652
+800 train 7.411017 (lr=9.9968e-05) (hash(x)=45093459)
+900 val loss 7.6041
+900 val perplexity 2006.4135
+900 train 7.915620 (lr=6.9969e-05) (hash(x)=54988361)
+900 val loss 7.5990
+900 val perplexity 1996.2872
+900 train 7.901796 (lr=4.9978e-05) (hash(x)=54988361)
+900 val loss 7.6248
+900 val perplexity 2048.3447
+900 train 7.913666 (lr=9.9956e-05) (hash(x)=54988361)
+1000 val loss 7.6032
+1000 val perplexity 2004.5659
+1000 train 7.487446 (lr=6.9960e-05) (hash(x)=47588648)
+1000 val loss 7.6414
+1000 val perplexity 2082.6899
+1000 train 7.527693 (lr=9.9943e-05) (hash(x)=47588648)
+1000 val loss 7.5895
+1000 val perplexity 1977.3726
+1000 train 7.466284 (lr=4.9971e-05) (hash(x)=47588648)
+1100 val loss 7.5861
+1100 val perplexity 1970.5438
+1100 train 7.207222 (lr=6.9949e-05) (hash(x)=37984588)
+1100 val loss 7.6216
+1100 val perplexity 2041.8978
+1100 train 7.275620 (lr=9.9927e-05) (hash(x)=37984588)
+1100 val loss 7.5907
+1100 val perplexity 1979.6897
+1100 train 7.223131 (lr=4.9964e-05) (hash(x)=37984588)
+1200 val loss 7.5671
+1200 val perplexity 1933.5526
+1200 train 7.731050 (lr=6.9937e-05) (hash(x)=56333817)
+1200 val loss 7.5809
+1200 val perplexity 1960.4640
+1200 train 7.750664 (lr=9.9910e-05) (hash(x)=56333817)
+1200 val loss 7.5857
+1200 val perplexity 1969.8665
+1200 train 7.749602 (lr=4.9955e-05) (hash(x)=56333817)
+1300 val loss 7.5572
+1300 val perplexity 1914.3943
+1300 train 7.702303 (lr=6.9924e-05) (hash(x)=53454056)
+1300 val loss 7.5620
+1300 val perplexity 1923.5961
+1300 train 7.703107 (lr=9.9892e-05) (hash(x)=53454056)
+1400 val loss 7.5422
+1400 val perplexity 1885.9550
+1400 train 7.643173 (lr=6.9910e-05) (hash(x)=55284163)
+1300 val loss 7.5975
+1300 val perplexity 1993.1512
+1300 train 7.718252 (lr=4.9946e-05) (hash(x)=53454056)
+1400 val loss 7.5226
+1400 val perplexity 1849.4447
+1400 train 7.625996 (lr=9.9871e-05) (hash(x)=55284163)
+1500 val loss 7.6245
+1500 val perplexity 2047.8584
+1500 train 7.548000 (lr=6.9894e-05) (hash(x)=48162598)
+1400 val loss 7.5680
+1400 val perplexity 1935.2434
+1400 train 7.663744 (lr=4.9936e-05) (hash(x)=55284163)
+1600 val loss 7.5211
+1600 val perplexity 1846.5297
+1600 train 7.567415 (lr=6.9877e-05) (hash(x)=54214535)
+1500 val loss 7.4830
+1500 val perplexity 1777.5719
+1500 train 7.413181 (lr=9.9849e-05) (hash(x)=48162598)
+1500 val loss 7.5403
+1500 val perplexity 1882.3864
+1500 train 7.457651 (lr=4.9924e-05) (hash(x)=48162598)
+1700 val loss 7.4984
+1700 val perplexity 1805.1927
+1700 train 7.708467 (lr=6.9859e-05) (hash(x)=53525003)
+1600 val loss 7.4618
+1600 val perplexity 1740.1932
+1600 train 7.523397 (lr=9.9825e-05) (hash(x)=54214535)
+1600 val loss 7.5234
+1600 val perplexity 1850.8969
+1600 train 7.582265 (lr=4.9912e-05) (hash(x)=54214535)
+1800 val loss 7.4766
+1800 val perplexity 1766.2341
+1800 train 7.561733 (lr=6.9840e-05) (hash(x)=51848994)
+1700 val loss 7.4384
+1700 val perplexity 1700.0876
+1700 train 7.670763 (lr=9.9799e-05) (hash(x)=53525003)
+1700 val loss 7.5229
+1700 val perplexity 1849.8566
+1700 train 7.723702 (lr=4.9899e-05) (hash(x)=53525003)
+1900 val loss 7.4697
+1900 val perplexity 1754.1061
+1900 train 7.271915 (lr=6.9819e-05) (hash(x)=48405987)
+1800 val loss 7.4013
+1800 val perplexity 1638.1406
+1800 train 7.474629 (lr=9.9771e-05) (hash(x)=51848994)
+1800 val loss 7.5197
+1800 val perplexity 1843.9948
+1800 train 7.608754 (lr=4.9885e-05) (hash(x)=51848994)
+2000 val loss 7.4516
+2000 val perplexity 1722.6387
+2000 train 7.840192 (lr=6.9797e-05) (hash(x)=58592291)
+1900 val loss 7.3939
+1900 val perplexity 1626.0569
+1900 train 7.207140 (lr=9.9741e-05) (hash(x)=48405987)
+1900 val loss 7.5030
+1900 val perplexity 1813.5634
+1900 train 7.312572 (lr=4.9871e-05) (hash(x)=48405987)
+2100 val loss 7.4566
+2100 val perplexity 1731.1819
+2100 train 7.538568 (lr=6.9774e-05) (hash(x)=51167081)
+2000 val loss 7.3727
+2000 val perplexity 1591.9402
+2000 train 7.721725 (lr=9.9710e-05) (hash(x)=58592291)
+2000 val loss 7.5025
+2000 val perplexity 1812.5009
+2000 train 7.861420 (lr=4.9855e-05) (hash(x)=58592291)
+2200 val loss 7.4264
+2200 val perplexity 1679.8331
+2200 train 7.429912 (lr=6.9750e-05) (hash(x)=47994988)
+2100 val loss 7.3837
+2100 val perplexity 1609.5760
+2100 train 7.464706 (lr=9.9677e-05) (hash(x)=51167081)
+2100 val loss 7.4977
+2100 val perplexity 1803.8229
+2100 train 7.601411 (lr=4.9839e-05) (hash(x)=51167081)
+2300 val loss 7.4027
+2300 val perplexity 1640.4669
+2300 train 7.390494 (lr=6.9724e-05) (hash(x)=47377604)
+2200 val loss 7.4287
+2200 val perplexity 1683.6567
+2200 train 7.427091 (lr=9.9642e-05) (hash(x)=47994988)
+2400 val loss 7.3508
+2400 val perplexity 1557.4562
+2400 train 7.401909 (lr=6.9697e-05) (hash(x)=53554323)
+2200 val loss 7.4798
+2200 val perplexity 1771.9459
+2200 train 7.475770 (lr=4.9821e-05) (hash(x)=47994988)
+2300 val loss 7.3702
+2300 val perplexity 1587.9591
+2300 train 7.350887 (lr=9.9606e-05) (hash(x)=47377604)
+2500 val loss 7.3187
+2500 val perplexity 1508.2208
+2500 train 7.323335 (lr=6.9669e-05) (hash(x)=50780417)
+2300 val loss 7.4838
+2300 val perplexity 1778.9948
+2300 train 7.455084 (lr=4.9803e-05) (hash(x)=47377604)
+2400 val loss 7.3521
+2400 val perplexity 1559.5281
+2400 train 7.410884 (lr=9.9567e-05) (hash(x)=53554323)
+2600 val loss 7.2853
+2600 val perplexity 1458.7247
+2600 train 7.181382 (lr=6.9640e-05) (hash(x)=46453562)
+2400 val loss 7.4971
+2400 val perplexity 1802.8408
+2400 train 7.552158 (lr=4.9784e-05) (hash(x)=53554323)
+2500 val loss 7.3333
+2500 val perplexity 1530.3920
+2500 train 7.320897 (lr=9.9527e-05) (hash(x)=50780417)
+2700 val loss 7.2758
+2700 val perplexity 1444.8698
+2700 train 7.258449 (lr=6.9609e-05) (hash(x)=54404221)
+2500 val loss 7.4722
+2500 val perplexity 1758.5254
+2500 train 7.460775 (lr=4.9764e-05) (hash(x)=50780417)
+2600 val loss 7.3253
+2600 val perplexity 1518.1760
+2600 train 7.212368 (lr=9.9485e-05) (hash(x)=46453562)
+2800 val loss 7.2339
+2800 val perplexity 1385.6417
+2800 train 7.944000 (lr=6.9577e-05) (hash(x)=59318895)
+2600 val loss 7.4753
+2600 val perplexity 1763.9120
+2600 train 7.385237 (lr=4.9743e-05) (hash(x)=46453562)
+2900 val loss 7.2038
+2900 val perplexity 1344.5312
+2900 train 7.183782 (lr=6.9544e-05) (hash(x)=47845760)
+2700 val loss 7.3421
+2700 val perplexity 1543.9019
+2700 train 7.332971 (lr=9.9442e-05) (hash(x)=54404221)
+2700 val loss 7.4581
+2700 val perplexity 1733.8107
+2700 train 7.486804 (lr=4.9721e-05) (hash(x)=54404221)
+3000 val loss 7.1747
+3000 val perplexity 1306.0240
+3000 train 6.879607 (lr=6.9510e-05) (hash(x)=44336167)
+2800 val loss 7.3176
+2800 val perplexity 1506.5914
+2800 train 8.045705 (lr=9.9396e-05) (hash(x)=59318895)
+2800 val loss 7.4447
+2800 val perplexity 1710.8124
+2800 train 8.284585 (lr=4.9698e-05) (hash(x)=59318895)
+3100 val loss 7.1436
+3100 val perplexity 1265.9867
+3100 train 7.444237 (lr=6.9474e-05) (hash(x)=44479330)
+2900 val loss 7.2990
+2900 val perplexity 1478.8462
+2900 train 7.271435 (lr=9.9349e-05) (hash(x)=47845760)
+2900 val loss 7.4506
+2900 val perplexity 1720.9794
+2900 train 7.431831 (lr=4.9674e-05) (hash(x)=47845760)
+3200 val loss 7.1320
+3200 val perplexity 1251.3336
+3200 train 7.191150 (lr=6.9438e-05) (hash(x)=54593096)
+3000 val loss 7.2992
+3000 val perplexity 1479.0916
+3000 train 6.994728 (lr=9.9300e-05) (hash(x)=44336167)
+3300 val loss 7.0997
+3300 val perplexity 1211.6255
+3300 train 7.011525 (lr=6.9400e-05) (hash(x)=45347643)
+3000 val loss 7.4237
+3000 val perplexity 1675.1554
+3000 train 7.138683 (lr=4.9650e-05) (hash(x)=44336167)
+3100 val loss 7.2724
+3100 val perplexity 1440.0378
+3100 train 7.530640 (lr=9.9249e-05) (hash(x)=44479330)
+3400 val loss 7.0938
+3400 val perplexity 1204.4785
+3400 train 7.215239 (lr=6.9360e-05) (hash(x)=47797247)
+3100 val loss 7.4045
+3100 val perplexity 1643.3512
+3100 train 7.655307 (lr=4.9625e-05) (hash(x)=44479330)
+3200 val loss 7.2596
+3200 val perplexity 1421.6920
+3200 train 7.307819 (lr=9.9197e-05) (hash(x)=54593096)
+3500 val loss 7.0682
+3500 val perplexity 1174.0056
+3500 train 6.962046 (lr=6.9320e-05) (hash(x)=46115683)
+3200 val loss 7.4067
+3200 val perplexity 1647.0352
+3200 train 7.450519 (lr=4.9598e-05) (hash(x)=54593096)
+3300 val loss 7.2790
+3300 val perplexity 1449.4850
+3300 train 7.129585 (lr=9.9142e-05) (hash(x)=45347643)
+3600 val loss 7.0492
+3600 val perplexity 1151.9397
+3600 train 6.904495 (lr=6.9278e-05) (hash(x)=44502074)
+3300 val loss 7.3963
+3300 val perplexity 1629.9904
+3300 train 7.239784 (lr=4.9571e-05) (hash(x)=45347643)
+3400 val loss 7.2493
+3400 val perplexity 1407.1310
+3400 train 7.375264 (lr=9.9086e-05) (hash(x)=47797247)
+3700 val loss 7.0325
+3700 val perplexity 1132.8109
+3700 train 7.175766 (lr=6.9235e-05) (hash(x)=55388443)
+3400 val loss 7.3565
+3400 val perplexity 1566.4130
+3400 train 7.464308 (lr=4.9543e-05) (hash(x)=47797247)
+3500 val loss 7.2425
+3500 val perplexity 1397.5653
+3500 train 7.154598 (lr=9.9028e-05) (hash(x)=46115683)
+3800 val loss 7.0013
+3800 val perplexity 1098.0376
+3800 train 6.810500 (lr=6.9191e-05) (hash(x)=43790341)
+3500 val loss 7.3431
+3500 val perplexity 1545.4553
+3500 train 7.265201 (lr=4.9514e-05) (hash(x)=46115683)
+3600 val loss 7.2292
+3600 val perplexity 1379.1554
+3600 train 7.053658 (lr=9.8969e-05) (hash(x)=44502074)
+3900 val loss 6.9691
+3900 val perplexity 1063.2583
+3900 train 6.973560 (lr=6.9146e-05) (hash(x)=50013318)
+3600 val loss 7.3240
+3600 val perplexity 1516.3240
+3600 train 7.149223 (lr=4.9484e-05) (hash(x)=44502074)
+3700 val loss 7.2192
+3700 val perplexity 1365.4557
+3700 train 7.358182 (lr=9.8908e-05) (hash(x)=55388443)
+4000 val loss 6.9317
+4000 val perplexity 1024.2280
+4000 train 7.021192 (lr=6.9099e-05) (hash(x)=51704787)
+3700 val loss 7.3146
+3700 val perplexity 1502.0967
+3700 train 7.437807 (lr=4.9454e-05) (hash(x)=55388443)
+3800 val loss 7.2244
+3800 val perplexity 1372.5228
+3800 train 7.005445 (lr=9.8845e-05) (hash(x)=43790341)
+4100 val loss 6.9325
+4100 val perplexity 1025.0312
+4100 train 7.015868 (lr=6.9051e-05) (hash(x)=50821964)
+3800 val loss 7.2936
+3800 val perplexity 1470.9061
+3800 train 7.100006 (lr=4.9422e-05) (hash(x)=43790341)
+3900 val loss 7.1979
+3900 val perplexity 1336.5901
+3900 train 7.187948 (lr=9.8780e-05) (hash(x)=50013318)
+4200 val loss 6.9185
+4200 val perplexity 1010.7939
+4200 train 6.944781 (lr=6.9002e-05) (hash(x)=49675080)
+3900 val loss 7.2722
+3900 val perplexity 1439.6870
+3900 train 7.297838 (lr=4.9390e-05) (hash(x)=50013318)
+4000 val loss 7.1747
+4000 val perplexity 1305.9094
+4000 train 7.256429 (lr=9.8713e-05) (hash(x)=51704787)
+4300 val loss 6.9167
+4300 val perplexity 1008.9371
+4300 train 6.592006 (lr=6.8952e-05) (hash(x)=43239281)
+4400 val loss 6.8773
+4400 val perplexity 969.9754
+4400 train 6.568204 (lr=6.8901e-05) (hash(x)=45076737)
+4000 val loss 7.2549
+4000 val perplexity 1414.9628
+4000 train 7.339958 (lr=4.9357e-05) (hash(x)=51704787)
+4100 val loss 7.1546
+4100 val perplexity 1280.0354
+4100 train 7.218249 (lr=9.8645e-05) (hash(x)=50821964)
+4500 val loss 6.8609
+4500 val perplexity 954.2654
+4500 train 6.980698 (lr=6.8848e-05) (hash(x)=57930262)
+4200 val loss 7.1594
+4200 val perplexity 1286.0985
+4200 train 7.182823 (lr=9.8575e-05) (hash(x)=49675080)
+4100 val loss 7.2445
+4100 val perplexity 1400.3203
+4100 train 7.328505 (lr=4.9322e-05) (hash(x)=50821964)
+4600 val loss 6.8439
+4600 val perplexity 938.1647
+4600 train 6.625656 (lr=6.8794e-05) (hash(x)=46721614)
+4300 val loss 7.1313
+4300 val perplexity 1250.5415
+4300 train 6.837461 (lr=9.8503e-05) (hash(x)=43239281)
+4200 val loss 7.2199
+4200 val perplexity 1366.3474
+4200 train 7.257685 (lr=4.9287e-05) (hash(x)=49675080)
+4400 val loss 7.1284
+4400 val perplexity 1246.8333
+4400 train 6.826270 (lr=9.8430e-05) (hash(x)=45076737)
+4700 val loss 6.8368
+4700 val perplexity 931.5139
+4700 train 6.672508 (lr=6.8739e-05) (hash(x)=49837920)
+4300 val loss 7.2085
+4300 val perplexity 1350.9222
+4300 train 6.875511 (lr=4.9252e-05) (hash(x)=43239281)
+4500 val loss 7.1382
+4500 val perplexity 1259.2046
+4500 train 7.242622 (lr=9.8355e-05) (hash(x)=57930262)
+4800 val loss 6.8120
+4800 val perplexity 908.7191
+4800 train 6.964200 (lr=6.8683e-05) (hash(x)=48380045)
+4400 val loss 7.1996
+4400 val perplexity 1338.9216
+4400 train 6.902917 (lr=4.9215e-05) (hash(x)=45076737)
+4600 val loss 7.1429
+4600 val perplexity 1265.0376
+4600 train 6.952327 (lr=9.8278e-05) (hash(x)=46721614)
+4900 val loss 6.7899
+4900 val perplexity 888.8326
+4900 train 6.666987 (lr=6.8626e-05) (hash(x)=44202577)
+4500 val loss 7.1818
+4500 val perplexity 1315.2502
+4500 train 7.273689 (lr=4.9177e-05) (hash(x)=57930262)
+5000 val loss 6.7547
+5000 val perplexity 858.0734
+5000 train 6.760885 (lr=6.8567e-05) (hash(x)=52038024)
+4700 val loss 7.1145
+4700 val perplexity 1229.6857
+4700 train 6.983832 (lr=9.8199e-05) (hash(x)=49837920)
+4600 val loss 7.1764
+4600 val perplexity 1308.2018
+4600 train 7.008025 (lr=4.9139e-05) (hash(x)=46721614)
+5100 val loss 6.7286
+5100 val perplexity 835.9825
+5100 train 6.895877 (lr=6.8507e-05) (hash(x)=53700038)
+4800 val loss 7.0932
+4800 val perplexity 1203.7006
+4800 train 7.200828 (lr=9.8119e-05) (hash(x)=48380045)
+4700 val loss 7.1608
+4700 val perplexity 1287.9175
+4700 train 7.031097 (lr=4.9099e-05) (hash(x)=49837920)
+5200 val loss 6.7278
+5200 val perplexity 835.3342
+5200 train 6.687095 (lr=6.8446e-05) (hash(x)=48137625)
+4900 val loss 7.1155
+4900 val perplexity 1230.9294
+4900 train 6.997181 (lr=9.8036e-05) (hash(x)=44202577)
+4800 val loss 7.1430
+4800 val perplexity 1265.2040
+4800 train 7.249168 (lr=4.9059e-05) (hash(x)=48380045)
+5300 val loss 6.7074
+5300 val perplexity 818.4043
+5300 train 6.550315 (lr=6.8384e-05) (hash(x)=43161573)
+5000 val loss 7.1163
+5000 val perplexity 1231.9117
+5000 train 7.120258 (lr=9.7953e-05) (hash(x)=52038024)
+4900 val loss 7.1146
+4900 val perplexity 1229.7953
+4900 train 6.998655 (lr=4.9018e-05) (hash(x)=44202577)
+5400 val loss 6.7170
+5400 val perplexity 826.3092
+5400 train 6.898795 (lr=6.8320e-05) (hash(x)=56673322)
+5100 val loss 7.0771
+5100 val perplexity 1184.5468
+5100 train 7.241188 (lr=9.7867e-05) (hash(x)=53700038)
+5000 val loss 7.0868
+5000 val perplexity 1196.0487
+5000 train 7.105583 (lr=4.8976e-05) (hash(x)=52038024)
+5500 val loss 6.6953
+5500 val perplexity 808.5854
+5500 train 6.883098 (lr=6.8256e-05) (hash(x)=53468295)
+5200 val loss 7.0188
+5200 val perplexity 1117.4941
+5200 train 6.979750 (lr=9.7780e-05) (hash(x)=48137625)
+5100 val loss 7.0759
+5100 val perplexity 1183.0616
+5100 train 7.257845 (lr=4.8934e-05) (hash(x)=53700038)
+5600 val loss 6.6755
+5600 val perplexity 792.7690
+5600 train 6.960256 (lr=6.8190e-05) (hash(x)=59287280)
+5300 val loss 6.9958
+5300 val perplexity 1092.0437
+5300 train 6.847336 (lr=9.7691e-05) (hash(x)=43161573)
+5200 val loss 7.0647
+5200 val perplexity 1169.9166
+5200 train 7.022595 (lr=4.8890e-05) (hash(x)=48137625)
+5700 val loss 6.6592
+5700 val perplexity 779.9288
+5700 train 6.861547 (lr=6.8123e-05) (hash(x)=57575806)
+5400 val loss 7.0182
+5400 val perplexity 1116.8011
+5400 train 7.199425 (lr=9.7600e-05) (hash(x)=56673322)
+5300 val loss 7.0409
+5300 val perplexity 1142.4619
+5300 train 6.914318 (lr=4.8846e-05) (hash(x)=43161573)
+5800 val loss 6.6521
+5800 val perplexity 774.4315
+5800 train 6.634932 (lr=6.8055e-05) (hash(x)=46897279)
+5500 val loss 7.0327
+5500 val perplexity 1133.1058
+5500 train 7.204011 (lr=9.7508e-05) (hash(x)=53468295)
+5400 val loss 7.0017
+5400 val perplexity 1098.5042
+5400 train 7.188062 (lr=4.8800e-05) (hash(x)=56673322)
+5900 val loss 6.6694
+5900 val perplexity 787.9207
+5900 train 6.553205 (lr=6.7985e-05) (hash(x)=47565679)
+5600 val loss 6.9888
+5600 val perplexity 1084.4437
+5600 train 7.272110 (lr=9.7414e-05) (hash(x)=59287280)
+5500 val loss 6.9708
+5500 val perplexity 1065.1246
+5500 train 7.154151 (lr=4.8754e-05) (hash(x)=53468295)
+6000 val loss 6.6493
+6000 val perplexity 772.2234
+6000 train 6.505241 (lr=6.7915e-05) (hash(x)=51590090)
+5700 val loss 6.9836
+5700 val perplexity 1078.7675
+5700 train 7.219868 (lr=9.7318e-05) (hash(x)=57575806)
+5600 val loss 6.9361
+5600 val perplexity 1028.7036
+5600 train 7.260452 (lr=4.8707e-05) (hash(x)=59287280)
+6100 val loss 6.6288
+6100 val perplexity 756.5986
+6100 train 7.037403 (lr=6.7843e-05) (hash(x)=59732271)
+5800 val loss 6.9847
+5800 val perplexity 1079.9718
+5800 train 6.926145 (lr=9.7221e-05) (hash(x)=46897279)
+5700 val loss 6.9153
+5700 val perplexity 1007.5467
+5700 train 7.155314 (lr=4.8659e-05) (hash(x)=57575806)
+6200 val loss 6.6045
+6200 val perplexity 738.4030
+6200 train 6.632282 (lr=6.7770e-05) (hash(x)=46394422)
+5900 val loss 6.9803
+5900 val perplexity 1075.2649
+5900 train 6.846642 (lr=9.7122e-05) (hash(x)=47565679)
+6300 val loss 6.6123
+6300 val perplexity 744.2033
+6300 train 6.649691 (lr=6.7696e-05) (hash(x)=53748145)
+5800 val loss 6.8961
+5800 val perplexity 988.3776
+5800 train 6.821358 (lr=4.8611e-05) (hash(x)=46897279)
+6000 val loss 7.0038
+6000 val perplexity 1100.8066
+6000 train 6.874036 (lr=9.7021e-05) (hash(x)=51590090)
+6400 val loss 6.5895
+6400 val perplexity 727.4420
+6400 train 6.455684 (lr=6.7621e-05) (hash(x)=46054751)
+6100 val loss 6.9770
+6100 val perplexity 1071.6895
+5900 val loss 6.8721
+5900 val perplexity 964.9728
+6100 train 7.369457 (lr=9.6919e-05) (hash(x)=59732271)
+5900 train 6.744565 (lr=4.8561e-05) (hash(x)=47565679)
+6500 val loss 6.5751
+6500 val perplexity 716.9923
+6500 train 6.855925 (lr=6.7545e-05) (hash(x)=51816809)
+6200 val loss 6.9668
+6200 val perplexity 1060.8143
+6200 train 6.925545 (lr=9.6815e-05) (hash(x)=46394422)
+6000 val loss 6.8457
+6000 val perplexity 939.8616
+6000 train 6.720833 (lr=4.8511e-05) (hash(x)=51590090)
+6600 val loss 6.5757
+6600 val perplexity 717.4735
+6600 train 6.422357 (lr=6.7467e-05) (hash(x)=52453336)
+6300 val loss 6.9703
+6300 val perplexity 1064.5681
+6300 train 6.989490 (lr=9.6709e-05) (hash(x)=53748145)
+6100 val loss 6.8210
+6100 val perplexity 916.9286
+6100 train 7.224485 (lr=4.8459e-05) (hash(x)=59732271)
+6700 val loss 6.5693
+6700 val perplexity 712.8660
+6700 train 6.581459 (lr=6.7389e-05) (hash(x)=49108775)
+6400 val loss 6.9472
+6400 val perplexity 1040.2695
+6400 train 6.806530 (lr=9.6602e-05) (hash(x)=46054751)
+6200 val loss 6.8051
+6200 val perplexity 902.4440
+6200 train 6.774230 (lr=4.8407e-05) (hash(x)=46394422)
+6800 val loss 6.5485
+6800 val perplexity 698.1708
+6800 train 6.429181 (lr=6.7309e-05) (hash(x)=46745396)
+6500 val loss 6.9401
+6500 val perplexity 1032.8649
+6500 train 7.142802 (lr=9.6493e-05) (hash(x)=51816809)
+6300 val loss 6.7819
+6300 val perplexity 881.7365
+6300 train 6.804878 (lr=4.8355e-05) (hash(x)=53748145)
+6900 val loss 6.5446
+6900 val perplexity 695.5093
+6900 train 6.687532 (lr=6.7228e-05) (hash(x)=46534986)
+6600 val loss 6.9536
+6600 val perplexity 1046.8857
+6600 train 6.907654 (lr=9.6382e-05) (hash(x)=52453336)
+6400 val loss 6.7609
+6400 val perplexity 863.4091
+6400 train 6.635312 (lr=4.8301e-05) (hash(x)=46054751)
+7000 val loss 6.5358
+7000 val perplexity 689.4045
+7000 train 6.884590 (lr=6.7146e-05) (hash(x)=49317888)
+6700 val loss 6.9180
+6700 val perplexity 1010.2961
+6700 train 6.927352 (lr=9.6270e-05) (hash(x)=49108775)
+6500 val loss 6.7379
+6500 val perplexity 843.8121
+6500 train 6.975233 (lr=4.8246e-05) (hash(x)=51816809)
+6800 val loss 6.9378
+6800 val perplexity 1030.5094
+6800 train 6.820474 (lr=9.6156e-05) (hash(x)=46745396)
+7100 val loss 6.5310
+7100 val perplexity 686.0746
+7100 train 6.542171 (lr=6.7063e-05) (hash(x)=50360484)
+6600 val loss 6.7466
+6600 val perplexity 851.1754
+6600 train 6.636971 (lr=4.8191e-05) (hash(x)=52453336)
+7200 val loss 6.5334
+7200 val perplexity 687.7217
+7200 train 6.403098 (lr=6.6978e-05) (hash(x)=49515094)
+6900 val loss 6.9012
+6900 val perplexity 993.4229
+6900 train 7.024705 (lr=9.6040e-05) (hash(x)=46534986)
+6700 val loss 6.7190
+6700 val perplexity 828.0261
+6700 train 6.735272 (lr=4.8135e-05) (hash(x)=49108775)
+7300 val loss 6.5433
+7300 val perplexity 694.5876
+7300 train 6.540457 (lr=6.6893e-05) (hash(x)=51546861)
+7000 val loss 6.8997
+7000 val perplexity 991.9900
+7000 train 7.183020 (lr=9.5923e-05) (hash(x)=49317888)
+6800 val loss 6.7086
+6800 val perplexity 819.4630
+6800 train 6.608204 (lr=4.8078e-05) (hash(x)=46745396)
+7100 val loss 6.8746
+7100 val perplexity 967.3943
+7400 val loss 6.5318
+7400 val perplexity 686.6266
+7100 train 6.887483 (lr=9.5804e-05) (hash(x)=50360484)
+7400 train 6.516788 (lr=6.6806e-05) (hash(x)=48320948)
+6900 val loss 6.6809
+6900 val perplexity 797.0172
+6900 train 6.816370 (lr=4.8020e-05) (hash(x)=46534986)
+7200 val loss 6.8591
+7200 val perplexity 952.5170
+7200 train 6.759748 (lr=9.5683e-05) (hash(x)=49515094)
+7500 val loss 6.5409
+7500 val perplexity 692.9114
+7500 train 6.417346 (lr=6.6718e-05) (hash(x)=40167457)
+7300 val loss 6.8558
+7300 val perplexity 949.3570
+7300 train 6.855737 (lr=9.5561e-05) (hash(x)=51546861)
+7000 val loss 6.6739
+7000 val perplexity 791.4651
+7000 train 7.034169 (lr=4.7961e-05) (hash(x)=49317888)
+7600 val loss 6.5199
+7600 val perplexity 678.4926
+7600 train 6.449652 (lr=6.6630e-05) (hash(x)=49942165)
+7400 val loss 6.8292
+7400 val perplexity 924.4711
+7400 train 6.817543 (lr=9.5437e-05) (hash(x)=48320948)
+7100 val loss 6.6571
+7100 val perplexity 778.2993
+7100 train 6.667387 (lr=4.7902e-05) (hash(x)=50360484)
+7700 val loss 6.5375
+7700 val perplexity 690.5319
+7700 train 6.272810 (lr=6.6540e-05) (hash(x)=48853311)
+7500 val loss 6.8651
+7500 val perplexity 958.2407
+7200 val loss 6.6463
+7200 val perplexity 769.8997
+7500 train 6.620294 (lr=9.5312e-05) (hash(x)=40167457)
+7200 train 6.539092 (lr=4.7842e-05) (hash(x)=49515094)
+7800 val loss 6.5229
+7800 val perplexity 680.5576
+7800 train 6.419036 (lr=6.6448e-05) (hash(x)=48510117)
+7600 val loss 6.7919
+7600 val perplexity 890.5652
+7600 train 6.704808 (lr=9.5185e-05) (hash(x)=49942165)
+7300 val loss 6.6627
+7300 val perplexity 782.6371
+7300 train 6.658652 (lr=4.7781e-05) (hash(x)=51546861)
+7900 val loss 6.5182
+7900 val perplexity 677.3441
+7900 train 6.517859 (lr=6.6356e-05) (hash(x)=48339781)
+7700 val loss 6.7696
+7700 val perplexity 870.9767
+7700 train 6.502357 (lr=9.5057e-05) (hash(x)=48853311)
+7400 val loss 6.6366
+7400 val perplexity 762.4875
+7400 train 6.618514 (lr=4.7719e-05) (hash(x)=48320948)
+8000 val loss 6.5207
+8000 val perplexity 679.0684
+8000 train 6.626083 (lr=6.6263e-05) (hash(x)=54927320)
+7800 val loss 6.7588
+7800 val perplexity 861.5703
+7800 train 6.644112 (lr=9.4926e-05) (hash(x)=48510117)
+7500 val loss 6.6470
+7500 val perplexity 770.4976
+7500 train 6.465994 (lr=4.7656e-05) (hash(x)=40167457)
+8100 val loss 6.5098
+8100 val perplexity 671.6619
+8100 train 6.218547 (lr=6.6169e-05) (hash(x)=46461786)
+7900 val loss 6.7503
+7900 val perplexity 854.2759
+7900 train 6.755628 (lr=9.4795e-05) (hash(x)=48339781)
+7600 val loss 6.6118
+7600 val perplexity 743.8340
+7600 train 6.546916 (lr=4.7593e-05) (hash(x)=49942165)
+8200 val loss 6.4979
+8200 val perplexity 663.7219
+8200 train 6.427443 (lr=6.6073e-05) (hash(x)=51536260)
+8000 val loss 6.7511
+8000 val perplexity 855.0164
+8000 train 6.862554 (lr=9.4661e-05) (hash(x)=54927320)
+7700 val loss 6.5963
+7700 val perplexity 732.3832
+7700 train 6.328202 (lr=4.7528e-05) (hash(x)=48853311)
+8300 val loss 6.5058
+8300 val perplexity 669.0105
+8300 train 6.244885 (lr=6.5976e-05) (hash(x)=44770722)
+8100 val loss 6.7385
+8100 val perplexity 844.2999
+8100 train 6.479411 (lr=9.4526e-05) (hash(x)=46461786)
+7800 val loss 6.5929
+7800 val perplexity 729.8585
+7800 train 6.481411 (lr=4.7463e-05) (hash(x)=48510117)
+8400 val loss 6.4937
+8400 val perplexity 660.9323
+8400 train 6.498044 (lr=6.5879e-05) (hash(x)=50104957)
+8200 val loss 6.7219
+8200 val perplexity 830.3803
+8200 train 6.637498 (lr=9.4390e-05) (hash(x)=51536260)
+7900 val loss 6.5861
+7900 val perplexity 724.9585
+7900 train 6.601991 (lr=4.7397e-05) (hash(x)=48339781)
+8500 val loss 6.4829
+8500 val perplexity 653.8620
+8500 train 6.593787 (lr=6.5780e-05) (hash(x)=50132971)
+8300 val loss 6.7466
+8300 val perplexity 851.1957
+8300 train 6.485762 (lr=9.4252e-05) (hash(x)=44770722)
+8000 val loss 6.5696
+8000 val perplexity 713.0958
+8000 train 6.676756 (lr=4.7331e-05) (hash(x)=54927320)
+8600 val loss 6.4764
+8600 val perplexity 649.6110
+8600 train 6.432298 (lr=6.5680e-05) (hash(x)=52193699)
+8400 val loss 6.7239
+8400 val perplexity 832.0628
+8400 train 6.711411 (lr=9.4112e-05) (hash(x)=50104957)
+8100 val loss 6.5660
+8100 val perplexity 710.5417
+8100 train 6.273369 (lr=4.7263e-05) (hash(x)=46461786)
+8700 val loss 6.4616
+8700 val perplexity 640.1130
+8700 train 6.460426 (lr=6.5579e-05) (hash(x)=47902319)
+8500 val loss 6.7088
+8500 val perplexity 819.5598
+8500 train 6.815518 (lr=9.3971e-05) (hash(x)=50132971)
+8800 val loss 6.4731
+8800 val perplexity 647.5099
+8800 train 6.770748 (lr=6.5477e-05) (hash(x)=54904230)
+8200 val loss 6.5484
+8200 val perplexity 698.1472
+8200 train 6.469982 (lr=4.7195e-05) (hash(x)=51536260)
+8600 val loss 6.6916
+8600 val perplexity 805.6112
+8600 train 6.626624 (lr=9.3828e-05) (hash(x)=52193699)
+8900 val loss 6.4588
+8900 val perplexity 638.2678
+8900 train 6.358573 (lr=6.5374e-05) (hash(x)=46311615)
+8300 val loss 6.5529
+8300 val perplexity 701.2932
+8300 train 6.286871 (lr=4.7126e-05) (hash(x)=44770722)
+8700 val loss 6.6874
+8700 val perplexity 802.2026
+8700 train 6.685539 (lr=9.3684e-05) (hash(x)=47902319)
+9000 val loss 6.4628
+9000 val perplexity 640.8239
+9000 train 6.315720 (lr=6.5270e-05) (hash(x)=48535188)
+8400 val loss 6.5408
+8400 val perplexity 692.8417
+8400 train 6.534705 (lr=4.7056e-05) (hash(x)=50104957)
+8800 val loss 6.6797
+8800 val perplexity 796.0612
+8800 train 6.978171 (lr=9.3538e-05) (hash(x)=54904230)
+9100 val loss 6.4444
+9100 val perplexity 629.1714
+9100 train 6.496108 (lr=6.5164e-05) (hash(x)=51757372)
+8500 val loss 6.5281
+8500 val perplexity 684.1109
+8500 train 6.627559 (lr=4.6986e-05) (hash(x)=50132971)
+8900 val loss 6.6656
+8900 val perplexity 784.9064
+8900 train 6.556992 (lr=9.3391e-05) (hash(x)=46311615)
+9200 val loss 6.4398
+9200 val perplexity 626.2506
+9200 train 6.312794 (lr=6.5058e-05) (hash(x)=51131708)
+8600 val loss 6.5169
+8600 val perplexity 676.5113
+8600 train 6.478631 (lr=4.6914e-05) (hash(x)=52193699)
+9000 val loss 6.6869
+9000 val perplexity 801.8539
+9000 train 6.547362 (lr=9.3242e-05) (hash(x)=48535188)
+9300 val loss 6.4419
+9300 val perplexity 627.5950
+9300 train 6.396623 (lr=6.4951e-05) (hash(x)=44784276)
+8700 val loss 6.5056
+8700 val perplexity 668.9062
+8700 train 6.499852 (lr=4.6842e-05) (hash(x)=47902319)
+9100 val loss 6.6641
+9100 val perplexity 783.7721
+9100 train 6.704998 (lr=9.3092e-05) (hash(x)=51757372)
+9400 val loss 6.4374
+9400 val perplexity 624.7706
+9400 train 6.555745 (lr=6.4842e-05) (hash(x)=51981169)
+8800 val loss 6.4953
+8800 val perplexity 662.0366
+8800 train 6.775081 (lr=4.6769e-05) (hash(x)=54904230)
+9200 val loss 6.6609
+9200 val perplexity 781.2751
+9200 train 6.553735 (lr=9.2940e-05) (hash(x)=51131708)
+9500 val loss 6.4396
+9500 val perplexity 626.1757
+9500 train 6.402099 (lr=6.4733e-05) (hash(x)=47232936)
+8900 val loss 6.4968
+8900 val perplexity 663.0422
+8900 train 6.389380 (lr=4.6696e-05) (hash(x)=46311615)
+9300 val loss 6.6413
+9300 val perplexity 766.1156
+9300 train 6.591171 (lr=9.2786e-05) (hash(x)=44784276)
+9600 val loss 6.4422
+9600 val perplexity 627.8006
+9600 train 6.452951 (lr=6.4622e-05) (hash(x)=53800450)
+9000 val loss 6.4916
+9000 val perplexity 659.5980
+9000 train 6.363636 (lr=4.6621e-05) (hash(x)=48535188)
+9400 val loss 6.6425
+9400 val perplexity 767.0316
+9400 train 6.770110 (lr=9.2632e-05) (hash(x)=51981169)
+9700 val loss 6.4194
+9700 val perplexity 613.6654
+9700 train 6.515686 (lr=6.4511e-05) (hash(x)=55768123)
+9100 val loss 6.4601
+9100 val perplexity 639.1089
+9100 train 6.511439 (lr=4.6546e-05) (hash(x)=51757372)
+9500 val loss 6.6338
+9500 val perplexity 760.3599
+9500 train 6.565683 (lr=9.2475e-05) (hash(x)=47232936)
+9800 val loss 6.4353
+9800 val perplexity 623.4942
+9800 train 6.359432 (lr=6.4398e-05) (hash(x)=47745177)
+9200 val loss 6.4557
+9200 val perplexity 636.2913
+9600 val loss 6.6392
+9600 val perplexity 764.5008
+9200 train 6.332733 (lr=4.6470e-05) (hash(x)=51131708)
+9600 train 6.650306 (lr=9.2317e-05) (hash(x)=53800450)
+9900 val loss 6.4333
+9900 val perplexity 622.2385
+9900 train 6.698323 (lr=6.4284e-05) (hash(x)=56592246)
+9700 val loss 6.6322
+9700 val perplexity 759.1641
+9700 train 6.719718 (lr=9.2158e-05) (hash(x)=55768123)
+9300 val loss 6.4501
+9300 val perplexity 632.7852
+9300 train 6.408199 (lr=4.6393e-05) (hash(x)=44784276)
+10000 val loss 6.4301
+10000 val perplexity 620.2114
+10000 train 6.434453 (lr=6.4170e-05) (hash(x)=51655963)
+9800 val loss 6.6505
+9800 val perplexity 773.1475
+9800 train 6.584711 (lr=9.1997e-05) (hash(x)=47745177)
+9400 val loss 6.4461
+9400 val perplexity 630.2448
+9400 train 6.568119 (lr=4.6316e-05) (hash(x)=51981169)
+10100 val loss 6.4364
+10100 val perplexity 624.1843
+10100 train 6.337934 (lr=6.4054e-05) (hash(x)=49809511)
+9900 val loss 6.6443
+9900 val perplexity 768.3766
+9900 train 6.882045 (lr=9.1835e-05) (hash(x)=56592246)
+9500 val loss 6.4431
+9500 val perplexity 628.3265
+9500 train 6.391317 (lr=4.6238e-05) (hash(x)=47232936)
+10200 val loss 6.4607
+10200 val perplexity 639.5052
+10200 train 6.072261 (lr=6.3937e-05) (hash(x)=42297812)
+10000 val loss 6.6411
+10000 val perplexity 765.9586
+10000 train 6.647980 (lr=9.1671e-05) (hash(x)=51655963)
+9600 val loss 6.4569
+9600 val perplexity 637.1080
+9600 train 6.479267 (lr=4.6159e-05) (hash(x)=53800450)
+10300 val loss 6.4193
+10300 val perplexity 613.5688
+10300 train 6.110044 (lr=6.3820e-05) (hash(x)=55529820)
+10100 val loss 6.6212
+10100 val perplexity 750.8585
+10100 train 6.521398 (lr=9.1506e-05) (hash(x)=49809511)
+10400 val loss 6.4131
+10400 val perplexity 609.7966
+10400 train 6.424881 (lr=6.3701e-05) (hash(x)=53255684)
+9700 val loss 6.4372
+9700 val perplexity 624.6452
+9700 train 6.544666 (lr=4.6079e-05) (hash(x)=55768123)
+10200 val loss 6.6227
+10200 val perplexity 752.0069
+10200 train 6.219979 (lr=9.1339e-05) (hash(x)=42297812)
+10500 val loss 6.3948
+10500 val perplexity 598.7203
+10500 train 6.502378 (lr=6.3581e-05) (hash(x)=54306191)
+9800 val loss 6.4376
+9800 val perplexity 624.9339
+9800 train 6.367727 (lr=4.5999e-05) (hash(x)=47745177)
+10300 val loss 6.6043
+10300 val perplexity 738.2425
+10300 train 6.311920 (lr=9.1171e-05) (hash(x)=55529820)
+10600 val loss 6.3874
+10600 val perplexity 594.2958
+10600 train 6.597528 (lr=6.3460e-05) (hash(x)=60130567)
+9900 val loss 6.4565
+9900 val perplexity 636.8025
+9900 train 6.694003 (lr=4.5917e-05) (hash(x)=56592246)
+10400 val loss 6.5965
+10400 val perplexity 732.5495
+10400 train 6.600775 (lr=9.1001e-05) (hash(x)=53255684)
+10700 val loss 6.3920
+10700 val perplexity 597.0769
+10700 train 6.317720 (lr=6.3339e-05) (hash(x)=50074737)
+10000 val loss 6.4274
+10000 val perplexity 618.5698
+10000 train 6.419580 (lr=4.5835e-05) (hash(x)=51655963)
+10500 val loss 6.5842
+10500 val perplexity 723.6005
+10500 train 6.704099 (lr=9.0830e-05) (hash(x)=54306191)
+10800 val loss 6.3791
+10800 val perplexity 589.3733
+10800 train 6.496333 (lr=6.3216e-05) (hash(x)=51547220)
+10100 val loss 6.4270
+10100 val perplexity 618.3342
+10100 train 6.324362 (lr=4.5753e-05) (hash(x)=49809511)
+10600 val loss 6.5758
+10600 val perplexity 717.4858
+10600 train 6.799292 (lr=9.0658e-05) (hash(x)=60130567)
+10900 val loss 6.3685
+10900 val perplexity 583.1586
+10900 train 6.451928 (lr=6.3092e-05) (hash(x)=55943981)
+10200 val loss 6.4205
+10200 val perplexity 614.3162
+10200 train 6.012050 (lr=4.5669e-05) (hash(x)=42297812)
+10700 val loss 6.5728
+10700 val perplexity 715.3784
+10700 train 6.510283 (lr=9.0484e-05) (hash(x)=50074737)
+11000 val loss 6.3550
+11000 val perplexity 575.3365
+11000 train 6.410864 (lr=6.2968e-05) (hash(x)=46444570)
+10300 val loss 6.4132
+10300 val perplexity 609.8723
+10300 train 6.100938 (lr=4.5585e-05) (hash(x)=55529820)
+10800 val loss 6.5693
+10800 val perplexity 712.9027
+10800 train 6.683527 (lr=9.0308e-05) (hash(x)=51547220)
+11100 val loss 6.4048
+11100 val perplexity 604.7125
+11100 train 6.424164 (lr=6.2842e-05) (hash(x)=49589063)
+10400 val loss 6.3961
+10400 val perplexity 599.5079
+10400 train 6.398651 (lr=4.5501e-05) (hash(x)=53255684)
+10900 val loss 6.5989
+10900 val perplexity 734.3118
+10900 train 6.665658 (lr=9.0132e-05) (hash(x)=55943981)
+11200 val loss 6.3603
+11200 val perplexity 578.4042
+11200 train 6.354611 (lr=6.2715e-05) (hash(x)=51392283)
+10500 val loss 6.3786
+10500 val perplexity 589.1190
+10500 train 6.479599 (lr=4.5415e-05) (hash(x)=54306191)
+11000 val loss 6.6051
+11000 val perplexity 738.8577
+11000 train 6.646755 (lr=8.9954e-05) (hash(x)=46444570)
+11300 val loss 6.4280
+11300 val perplexity 618.9520
+11300 train 6.434875 (lr=6.2588e-05) (hash(x)=45081133)
+10600 val loss 6.3785
+10600 val perplexity 589.0452
+10600 train 6.581085 (lr=4.5329e-05) (hash(x)=60130567)
+11100 val loss 6.6082
+11100 val perplexity 741.1121
+11100 train 6.600307 (lr=8.9774e-05) (hash(x)=49589063)
+11400 val loss 6.3718
+11400 val perplexity 585.1299
+11400 train 6.458348 (lr=6.2459e-05) (hash(x)=53700397)
+11200 val loss 6.5779
+11200 val perplexity 719.0346
+10700 val loss 6.3648
+10700 val perplexity 581.0134
+11200 train 6.580002 (lr=8.9593e-05) (hash(x)=51392283)
+10700 train 6.285759 (lr=4.5242e-05) (hash(x)=50074737)
+11500 val loss 6.3571
+11500 val perplexity 576.5493
+11500 train 6.090324 (lr=6.2330e-05) (hash(x)=43839088)
+11300 val loss 6.5766
+11300 val perplexity 718.0677
+11300 train 6.538782 (lr=8.9411e-05) (hash(x)=45081133)
+10800 val loss 6.3522
+10800 val perplexity 573.7385
+10800 train 6.471161 (lr=4.5154e-05) (hash(x)=51547220)
+11600 val loss 6.3560
+11600 val perplexity 575.9155
+11600 train 6.365404 (lr=6.2199e-05) (hash(x)=48088111)
+11400 val loss 6.5712
+11400 val perplexity 714.2205
+11400 train 6.654538 (lr=8.9227e-05) (hash(x)=53700397)
+10900 val loss 6.3555
+10900 val perplexity 575.6359
+10900 train 6.447626 (lr=4.5066e-05) (hash(x)=55943981)
+11700 val loss 6.3418
+11700 val perplexity 567.8203
+11700 train 6.648592 (lr=6.2068e-05) (hash(x)=55108226)
+11500 val loss 6.5605
+11500 val perplexity 706.6102
+11500 train 6.303116 (lr=8.9043e-05) (hash(x)=43839088)
+11000 val loss 6.3386
+11000 val perplexity 565.9975
+11000 train 6.404186 (lr=4.4977e-05) (hash(x)=46444570)
+11800 val loss 6.3481
+11800 val perplexity 571.4037
+11800 train 6.587350 (lr=6.1936e-05) (hash(x)=58524839)
+11600 val loss 6.5759
+11600 val perplexity 717.5840
+11600 train 6.604741 (lr=8.8856e-05) (hash(x)=48088111)
+11100 val loss 6.3403
+11100 val perplexity 566.9902
+11100 train 6.350871 (lr=4.4887e-05) (hash(x)=49589063)
+11900 val loss 6.3405
+11900 val perplexity 567.0935
+11900 train 5.980791 (lr=6.1802e-05) (hash(x)=43864078)
+11700 val loss 6.5846
+11700 val perplexity 723.8970
+11700 train 6.892133 (lr=8.8668e-05) (hash(x)=55108226)
+12000 val loss 6.3421
+12000 val perplexity 567.9741
+12000 train 6.039177 (lr=6.1668e-05) (hash(x)=43448544)
+11200 val loss 6.3322
+11200 val perplexity 562.4065
+11200 train 6.340265 (lr=4.4797e-05) (hash(x)=51392283)
+11800 val loss 6.5801
+11800 val perplexity 720.6069
+11800 train 6.812759 (lr=8.8479e-05) (hash(x)=58524839)
+12100 val loss 6.3541
+12100 val perplexity 574.8290
+12100 train 6.301033 (lr=6.1533e-05) (hash(x)=55200399)
+11300 val loss 6.3448
+11300 val perplexity 569.5289
+11300 train 6.311532 (lr=4.4706e-05) (hash(x)=45081133)
+11900 val loss 6.5679
+11900 val perplexity 711.8446
+11900 train 6.241637 (lr=8.8289e-05) (hash(x)=43864078)
+12200 val loss 6.3503
+12200 val perplexity 572.6906
+12200 train 6.491292 (lr=6.1397e-05) (hash(x)=57627314)
+11400 val loss 6.3166
+11400 val perplexity 553.6919
+11400 train 6.405518 (lr=4.4614e-05) (hash(x)=53700397)
+12000 val loss 6.5646
+12000 val perplexity 709.5612
+12000 train 6.279223 (lr=8.8097e-05) (hash(x)=43448544)
+12300 val loss 6.3520
+12300 val perplexity 573.6589
+12300 train 6.656027 (lr=6.1260e-05) (hash(x)=53617087)
+11500 val loss 6.3162
+11500 val perplexity 553.4821
+11500 train 6.044001 (lr=4.4521e-05) (hash(x)=43839088)
+12100 val loss 6.5575
+12100 val perplexity 704.4769
+12100 train 6.543485 (lr=8.7904e-05) (hash(x)=55200399)
+12400 val loss 6.3432
+12400 val perplexity 568.5938
+12400 train 6.196439 (lr=6.1122e-05) (hash(x)=51135678)
+11600 val loss 6.3103
+11600 val perplexity 550.1926
+11600 train 6.319879 (lr=4.4428e-05) (hash(x)=48088111)
+12200 val loss 6.5644
+12200 val perplexity 709.3788
+12200 train 6.671797 (lr=8.7710e-05) (hash(x)=57627314)
+12500 val loss 6.3827
+12500 val perplexity 591.5469
+12500 train 6.429889 (lr=6.0984e-05) (hash(x)=48025130)
+11700 val loss 6.3133
+11700 val perplexity 551.8868
+11700 train 6.645210 (lr=4.4334e-05) (hash(x)=55108226)
+12300 val loss 6.5624
+12300 val perplexity 707.9727
+12300 train 6.810684 (lr=8.7515e-05) (hash(x)=53617087)
+12600 val loss 6.3387
+12600 val perplexity 566.0334
+12600 train 6.383214 (lr=6.0844e-05) (hash(x)=52135695)
+12400 val loss 6.5626
+12400 val perplexity 708.1314
+12400 train 6.393559 (lr=8.7318e-05) (hash(x)=51135678)
+11800 val loss 6.3220
+11800 val perplexity 556.7070
+11800 train 6.576732 (lr=4.4240e-05) (hash(x)=58524839)
+12700 val loss 6.3426
+12700 val perplexity 568.2723
+12700 train 6.201974 (lr=6.0703e-05) (hash(x)=51888613)
+12500 val loss 6.6048
+12500 val perplexity 738.6020
+12500 train 6.596214 (lr=8.7119e-05) (hash(x)=48025130)
+11900 val loss 6.3032
+11900 val perplexity 546.3365
+11900 train 5.970202 (lr=4.4145e-05) (hash(x)=43864078)
+12800 val loss 6.3370
+12800 val perplexity 565.0873
+12800 train 6.260093 (lr=6.0562e-05) (hash(x)=50418818)
+12600 val loss 6.5513
+12600 val perplexity 700.1442
+12600 train 6.610987 (lr=8.6920e-05) (hash(x)=52135695)
+12000 val loss 6.3029
+12000 val perplexity 546.1674
+12000 train 5.999082 (lr=4.4049e-05) (hash(x)=43448544)
+12900 val loss 6.3169
+12900 val perplexity 553.8445
+12900 train 7.201547 (lr=6.0420e-05) (hash(x)=58649585)
+12700 val loss 6.5578
+12700 val perplexity 704.6970
+12700 train 6.436542 (lr=8.6719e-05) (hash(x)=51888613)
+12100 val loss 6.3017
+12100 val perplexity 545.4907
+12100 train 6.246994 (lr=4.3952e-05) (hash(x)=55200399)
+13000 val loss 6.3034
+13000 val perplexity 546.4482
+13000 train 6.579797 (lr=6.0277e-05) (hash(x)=54567307)
+12800 val loss 6.5742
+12800 val perplexity 716.3731
+12800 train 6.467278 (lr=8.6517e-05) (hash(x)=50418818)
+12200 val loss 6.2922
+12200 val perplexity 540.3362
+12200 train 6.435254 (lr=4.3855e-05) (hash(x)=57627314)
+13100 val loss 6.3046
+13100 val perplexity 547.0768
+13100 train 6.428601 (lr=6.0133e-05) (hash(x)=52071473)
+12900 val loss 6.5427
+12900 val perplexity 694.1595
+12900 train 7.406726 (lr=8.6314e-05) (hash(x)=58649585)
+12300 val loss 6.2971
+12300 val perplexity 543.0134
+12300 train 6.599758 (lr=4.3757e-05) (hash(x)=53617087)
+13200 val loss 6.2888
+13200 val perplexity 538.5305
+13200 train 6.058242 (lr=5.9988e-05) (hash(x)=46293092)
+13000 val loss 6.5517
+13000 val perplexity 700.4313
+13000 train 6.792434 (lr=8.6110e-05) (hash(x)=54567307)
+12400 val loss 6.3048
+12400 val perplexity 547.1718
+12400 train 6.146738 (lr=4.3659e-05) (hash(x)=51135678)
+13300 val loss 6.2830
+13300 val perplexity 535.3733
+13300 train 6.390278 (lr=5.9842e-05) (hash(x)=56511467)
+13100 val loss 6.5425
+13100 val perplexity 694.0284
+13100 train 6.697961 (lr=8.5904e-05) (hash(x)=52071473)
+13400 val loss 6.2832
+13400 val perplexity 535.4987
+13400 train 6.380155 (lr=5.9695e-05) (hash(x)=54753763)
+12500 val loss 6.3133
+12500 val perplexity 551.8790
+12500 train 6.356315 (lr=4.3560e-05) (hash(x)=48025130)
+13200 val loss 6.5249
+13200 val perplexity 681.9034
+13200 train 6.271211 (lr=8.5697e-05) (hash(x)=46293092)
+13500 val loss 6.2965
+13500 val perplexity 542.6864
+13500 train 6.437078 (lr=5.9548e-05) (hash(x)=53610247)
+12600 val loss 6.3031
+12600 val perplexity 546.2406
+12600 train 6.349621 (lr=4.3460e-05) (hash(x)=52135695)
+13300 val loss 6.5190
+13300 val perplexity 677.9108
+13300 train 6.633325 (lr=8.5489e-05) (hash(x)=56511467)
+13600 val loss 6.2793
+13600 val perplexity 533.3911
+13600 train 6.178974 (lr=5.9400e-05) (hash(x)=47526249)
+12700 val loss 6.2954
+12700 val perplexity 542.0657
+12700 train 6.138735 (lr=4.3360e-05) (hash(x)=51888613)
+13400 val loss 6.5053
+13400 val perplexity 668.6843
+13400 train 6.609445 (lr=8.5279e-05) (hash(x)=54753763)
+13700 val loss 6.2769
+13700 val perplexity 532.1181
+13700 train 6.125827 (lr=5.9251e-05) (hash(x)=51185517)
+12800 val loss 6.2961
+12800 val perplexity 542.4781
+12800 train 6.221097 (lr=4.3259e-05) (hash(x)=50418818)
+13500 val loss 6.5038
+13500 val perplexity 667.7010
+13500 train 6.635751 (lr=8.5069e-05) (hash(x)=53610247)
+13800 val loss 6.2660
+13800 val perplexity 526.3919
+13800 train 5.934208 (lr=5.9101e-05) (hash(x)=45953529)
+12900 val loss 6.2687
+12900 val perplexity 527.8148
+12900 train 7.175869 (lr=4.3157e-05) (hash(x)=58649585)
+13600 val loss 6.4959
+13600 val perplexity 662.4256
+13600 train 6.370335 (lr=8.4857e-05) (hash(x)=47526249)
+13900 val loss 6.2591
+13900 val perplexity 522.7527
+13900 train 6.160434 (lr=5.8950e-05) (hash(x)=47238157)
+13000 val loss 6.2670
+13000 val perplexity 526.8743
+13000 train 6.563784 (lr=4.3055e-05) (hash(x)=54567307)
+13700 val loss 6.5016
+13700 val perplexity 666.2260
+13700 train 6.336977 (lr=8.4644e-05) (hash(x)=51185517)
+14000 val loss 6.2591
+14000 val perplexity 522.7397
+14000 train 6.215769 (lr=5.8799e-05) (hash(x)=54250750)
+13100 val loss 6.2628
+13100 val perplexity 524.6711
+13100 train 6.379539 (lr=4.2952e-05) (hash(x)=52071473)
+13800 val loss 6.4955
+13800 val perplexity 662.1556
+13800 train 6.183954 (lr=8.4430e-05) (hash(x)=45953529)
+14100 val loss 6.2618
+14100 val perplexity 524.1670
+14100 train 6.260184 (lr=5.8646e-05) (hash(x)=48198552)
+13200 val loss 6.2525
+13200 val perplexity 519.2887
+13200 train 6.040390 (lr=4.2848e-05) (hash(x)=46293092)
+13900 val loss 6.4803
+13900 val perplexity 652.1401
+13900 train 6.410282 (lr=8.4214e-05) (hash(x)=47238157)
+14200 val loss 6.2610
+14200 val perplexity 523.7622
+14200 train 6.367507 (lr=5.8493e-05) (hash(x)=52020690)
+13300 val loss 6.2446
+13300 val perplexity 515.2338
+13300 train 6.354766 (lr=4.2744e-05) (hash(x)=56511467)
+14000 val loss 6.4987
+14000 val perplexity 664.2668
+14000 train 6.451422 (lr=8.3998e-05) (hash(x)=54250750)
+14300 val loss 6.2597
+14300 val perplexity 523.0881
+14300 train 6.066946 (lr=5.8339e-05) (hash(x)=45165483)
+13400 val loss 6.2296
+13400 val perplexity 507.5316
+13400 train 6.346925 (lr=4.2640e-05) (hash(x)=54753763)
+14100 val loss 6.5067
+14100 val perplexity 669.6297
+14100 train 6.501920 (lr=8.3780e-05) (hash(x)=48198552)
+14400 val loss 6.2628
+14400 val perplexity 524.6866
+14400 train 6.359657 (lr=5.8184e-05) (hash(x)=52184072)
+13500 val loss 6.2320
+13500 val perplexity 508.7899
+13500 train 6.372458 (lr=4.2534e-05) (hash(x)=53610247)
+14200 val loss 6.4944
+14200 val perplexity 661.4572
+14200 train 6.618190 (lr=8.3561e-05) (hash(x)=52020690)
+14500 val loss 6.2733
+14500 val perplexity 530.2107
+14500 train 6.625431 (lr=5.8029e-05) (hash(x)=56046436)
+13600 val loss 6.2320
+13600 val perplexity 508.7525
+14300 val loss 6.4961
+14300 val perplexity 662.5339
+13600 train 6.122686 (lr=4.2428e-05) (hash(x)=47526249)
+14300 train 6.288409 (lr=8.3341e-05) (hash(x)=45165483)
+14600 val loss 6.2666
+14600 val perplexity 526.6653
+14600 train 6.079505 (lr=5.7872e-05) (hash(x)=52029694)
+14400 val loss 6.5160
+14400 val perplexity 675.9000
+14400 train 6.597093 (lr=8.3120e-05) (hash(x)=52184072)
+13700 val loss 6.2232
+13700 val perplexity 504.3288
+13700 train 6.069735 (lr=4.2322e-05) (hash(x)=51185517)
+14700 val loss 6.2799
+14700 val perplexity 533.7134
+14700 train 6.125349 (lr=5.7715e-05) (hash(x)=50258224)
+14500 val loss 6.4984
+14500 val perplexity 664.1047
+14500 train 6.804925 (lr=8.2898e-05) (hash(x)=56046436)
+13800 val loss 6.2141
+13800 val perplexity 499.7631
+13800 train 5.898561 (lr=4.2215e-05) (hash(x)=45953529)
+14800 val loss 6.2789
+14800 val perplexity 533.1945
+14800 train 5.578360 (lr=5.7558e-05) (hash(x)=42112262)
+14600 val loss 6.4967
+14600 val perplexity 662.9833
+14600 train 6.377328 (lr=8.2675e-05) (hash(x)=52029694)
+13900 val loss 6.2199
+13900 val perplexity 502.6653
+13900 train 6.139126 (lr=4.2107e-05) (hash(x)=47238157)
+14900 val loss 6.2827
+14900 val perplexity 535.2421
+14900 train 6.032803 (lr=5.7399e-05) (hash(x)=47219933)
+14700 val loss 6.5066
+14700 val perplexity 669.5291
+14700 train 6.350065 (lr=8.2451e-05) (hash(x)=50258224)
+14000 val loss 6.2114
+14000 val perplexity 498.4152
+14000 train 6.168877 (lr=4.1999e-05) (hash(x)=54250750)
+15000 val loss 6.3072
+15000 val perplexity 548.5156
+15000 train 5.927101 (lr=5.7240e-05) (hash(x)=58309309)
+14800 val loss 6.5008
+14800 val perplexity 665.7036
+14800 train 5.805090 (lr=8.2225e-05) (hash(x)=42112262)
+14100 val loss 6.2063
+14100 val perplexity 495.8396
+14100 train 6.204028 (lr=4.1890e-05) (hash(x)=48198552)
+15100 val loss 6.2989
+15100 val perplexity 543.9886
+15100 train 6.204672 (lr=5.7079e-05) (hash(x)=48756049)
+14900 val loss 6.5060
+14900 val perplexity 669.1774
+14900 train 6.271698 (lr=8.1998e-05) (hash(x)=47219933)
+15200 val loss 6.2903
+15200 val perplexity 539.2983
+15200 train 5.983079 (lr=5.6919e-05) (hash(x)=49791737)
+14200 val loss 6.2062
+14200 val perplexity 495.8238
+14200 train 6.306804 (lr=4.1781e-05) (hash(x)=52020690)
+15000 val loss 6.5320
+15000 val perplexity 686.7567
+15000 train 6.149031 (lr=8.1771e-05) (hash(x)=58309309)
+15300 val loss 6.2821
+15300 val perplexity 534.8953
+15300 train 6.211565 (lr=5.6757e-05) (hash(x)=53084126)
+14300 val loss 6.2119
+14300 val perplexity 498.6263
+14300 train 6.032381 (lr=4.1671e-05) (hash(x)=45165483)
+15100 val loss 6.5226
+15100 val perplexity 680.3143
+15100 train 6.455632 (lr=8.1542e-05) (hash(x)=48756049)
+15400 val loss 6.2561
+15400 val perplexity 521.1675
+15400 train 6.303825 (lr=5.6595e-05) (hash(x)=55041679)
+14400 val loss 6.2042
+14400 val perplexity 494.8153
+14400 train 6.308162 (lr=4.1560e-05) (hash(x)=52184072)
+15200 val loss 6.5137
+15200 val perplexity 674.2991
+15200 train 6.220746 (lr=8.1312e-05) (hash(x)=49791737)
+15500 val loss 6.2417
+15500 val perplexity 513.7371
+15500 train 6.041740 (lr=5.6432e-05) (hash(x)=43249867)
+14500 val loss 6.2019
+14500 val perplexity 493.6857
+14500 train 6.565360 (lr=4.1449e-05) (hash(x)=56046436)
+15300 val loss 6.5161
+15300 val perplexity 675.9696
+15300 train 6.489110 (lr=8.1082e-05) (hash(x)=53084126)
+15600 val loss 6.2454
+15600 val perplexity 515.6241
+15600 train 6.214768 (lr=5.6268e-05) (hash(x)=49006517)
+14600 val loss 6.2042
+14600 val perplexity 494.8474
+14600 train 5.997983 (lr=4.1337e-05) (hash(x)=52029694)
+15400 val loss 6.4838
+15400 val perplexity 654.4450
+15400 train 6.545664 (lr=8.0850e-05) (hash(x)=55041679)
+15700 val loss 6.2311
+15700 val perplexity 508.2974
+15700 train 6.525909 (lr=5.6104e-05) (hash(x)=60986839)
+14700 val loss 6.2143
+14700 val perplexity 499.8418
+14700 train 6.074037 (lr=4.1225e-05) (hash(x)=50258224)
+15500 val loss 6.4487
+15500 val perplexity 631.9008
+15500 train 6.254567 (lr=8.0617e-05) (hash(x)=43249867)
+15800 val loss 6.2335
+15800 val perplexity 509.5486
+15800 train 6.157606 (lr=5.5938e-05) (hash(x)=48354906)
+14800 val loss 6.1960
+14800 val perplexity 490.7956
+14800 train 5.508065 (lr=4.1113e-05) (hash(x)=42112262)
+15600 val loss 6.4573
+15600 val perplexity 637.3088
+15600 train 6.429033 (lr=8.0383e-05) (hash(x)=49006517)
+15900 val loss 6.2353
+15900 val perplexity 510.4343
+15900 train 6.071015 (lr=5.5773e-05) (hash(x)=52679780)
+14900 val loss 6.1977
+14900 val perplexity 491.5990
+14900 train 5.965704 (lr=4.0999e-05) (hash(x)=47219933)
+15700 val loss 6.4370
+15700 val perplexity 624.5377
+15700 train 6.727993 (lr=8.0148e-05) (hash(x)=60986839)
+16000 val loss 6.2307
+16000 val perplexity 508.1077
+16000 train 6.364089 (lr=5.5606e-05) (hash(x)=58049587)
+15000 val loss 6.2242
+15000 val perplexity 504.8374
+15000 train 5.856767 (lr=4.0885e-05) (hash(x)=58309309)
+15800 val loss 6.4396
+15800 val perplexity 626.1273
+15800 train 6.376960 (lr=7.9912e-05) (hash(x)=48354906)
+16100 val loss 6.2193
+16100 val perplexity 502.3358
+16100 train 5.868877 (lr=5.5439e-05) (hash(x)=39795825)
+15100 val loss 6.1894
+15100 val perplexity 487.5722
+15100 train 6.100931 (lr=4.0771e-05) (hash(x)=48756049)
+15900 val loss 6.4217
+15900 val perplexity 615.0214
+15900 train 6.283855 (lr=7.9675e-05) (hash(x)=52679780)
+16200 val loss 6.2147
+16200 val perplexity 500.0380
+16200 train 6.207601 (lr=5.5271e-05) (hash(x)=49498689)
+16000 val loss 6.4304
+16000 val perplexity 620.4217
+16000 train 6.587129 (lr=7.9437e-05) (hash(x)=58049587)
+16300 val loss 6.2421
+16300 val perplexity 513.9544
+16300 train 6.600447 (lr=5.5102e-05) (hash(x)=51235194)
+15200 val loss 6.2033
+15200 val perplexity 494.3926
+15200 train 5.858490 (lr=4.0656e-05) (hash(x)=49791737)
+16100 val loss 6.4319
+16100 val perplexity 621.3712
+16100 train 6.077530 (lr=7.9198e-05) (hash(x)=39795825)
+16400 val loss 6.2190
+16400 val perplexity 502.1808
+16400 train 6.082575 (lr=5.4933e-05) (hash(x)=47789417)
+15300 val loss 6.1788
+15300 val perplexity 482.4112
+15300 train 6.114874 (lr=4.0541e-05) (hash(x)=53084126)
+16200 val loss 6.4436
+16200 val perplexity 628.6627
+16200 train 6.434575 (lr=7.8959e-05) (hash(x)=49498689)
+16500 val loss 6.2251
+16500 val perplexity 505.2743
+16500 train 6.784148 (lr=5.4763e-05) (hash(x)=59665889)
+15400 val loss 6.1750
+15400 val perplexity 480.5643
+15400 train 6.214546 (lr=4.0425e-05) (hash(x)=55041679)
+16300 val loss 6.4347
+16300 val perplexity 623.1120
+16300 train 6.750012 (lr=7.8718e-05) (hash(x)=51235194)
+16600 val loss 6.2364
+16600 val perplexity 511.0405
+16600 train 6.131698 (lr=5.4593e-05) (hash(x)=49034893)
+15500 val loss 6.1506
+15500 val perplexity 468.9987
+15500 train 5.963602 (lr=4.0308e-05) (hash(x)=43249867)
+16400 val loss 6.4408
+16400 val perplexity 626.8846
+16400 train 6.300445 (lr=7.8476e-05) (hash(x)=47789417)
+16700 val loss 6.2236
+16700 val perplexity 504.5263
+16700 train 6.050609 (lr=5.4422e-05) (hash(x)=45685955)
+15600 val loss 6.1598
+15600 val perplexity 473.3522
+15600 train 6.138355 (lr=4.0191e-05) (hash(x)=49006517)
+16500 val loss 6.4328
+16500 val perplexity 621.8873
+16500 train 6.958402 (lr=7.8233e-05) (hash(x)=59665889)
+16800 val loss 6.2321
+16800 val perplexity 508.8360
+16800 train 6.032184 (lr=5.4250e-05) (hash(x)=49589835)
+15700 val loss 6.1445
+15700 val perplexity 466.1526
+15700 train 6.424474 (lr=4.0074e-05) (hash(x)=60986839)
+16600 val loss 6.4360
+16600 val perplexity 623.9278
+16600 train 6.318515 (lr=7.7990e-05) (hash(x)=49034893)
+16900 val loss 6.2381
+16900 val perplexity 511.8797
+16900 train 5.675816 (lr=5.4078e-05) (hash(x)=38655565)
+15800 val loss 6.1471
+15800 val perplexity 467.3422
+15800 train 6.088002 (lr=3.9956e-05) (hash(x)=48354906)
+16700 val loss 6.4353
+16700 val perplexity 623.4493
+16700 train 6.250768 (lr=7.7745e-05) (hash(x)=45685955)
+17000 val loss 6.2390
+17000 val perplexity 512.3212
+17000 train 6.162674 (lr=5.3905e-05) (hash(x)=46730732)
+15900 val loss 6.1403
+15900 val perplexity 464.1739
+15900 train 5.986931 (lr=3.9838e-05) (hash(x)=52679780)
+16800 val loss 6.4277
+16800 val perplexity 618.7781
+16800 train 6.249182 (lr=7.7500e-05) (hash(x)=49589835)
+17100 val loss 6.2363
+17100 val perplexity 510.9644
+17100 train 5.835392 (lr=5.3731e-05) (hash(x)=43843722)
+16000 val loss 6.1446
+16000 val perplexity 466.1955
+16000 train 6.260038 (lr=3.9719e-05) (hash(x)=58049587)
+16900 val loss 6.4309
+16900 val perplexity 620.7564
+16900 train 5.830621 (lr=7.7254e-05) (hash(x)=38655565)
+17200 val loss 6.2360
+17200 val perplexity 510.8192
+17200 train 5.961731 (lr=5.3557e-05) (hash(x)=50198978)
+16100 val loss 6.1321
+16100 val perplexity 460.3810
+16100 train 5.801347 (lr=3.9599e-05) (hash(x)=39795825)
+17300 val loss 6.2276
+17300 val perplexity 506.5314
+17300 train 6.210217 (lr=5.3382e-05) (hash(x)=52976276)
+17000 val loss 6.4156
+17000 val perplexity 611.3114
+17000 train 6.364450 (lr=7.7007e-05) (hash(x)=46730732)
+16200 val loss 6.1447
+16200 val perplexity 466.2173
+16200 train 6.124951 (lr=3.9479e-05) (hash(x)=49498689)
+17400 val loss 6.2305
+17400 val perplexity 507.9846
+17400 train 6.068756 (lr=5.3206e-05) (hash(x)=52224275)
+17100 val loss 6.4381
+17100 val perplexity 625.2069
+17100 train 6.037139 (lr=7.6758e-05) (hash(x)=43843722)
+16300 val loss 6.1461
+16300 val perplexity 466.8778
+16300 train 6.477264 (lr=3.9359e-05) (hash(x)=51235194)
+17200 val loss 6.4236
+17200 val perplexity 616.2365
+17200 train 6.175350 (lr=7.6510e-05) (hash(x)=50198978)
+17500 val loss 6.2296
+17500 val perplexity 507.5582
+17500 train 6.032399 (lr=5.3030e-05) (hash(x)=49055112)
+16400 val loss 6.1407
+16400 val perplexity 464.3915
+16400 train 5.994458 (lr=3.9238e-05) (hash(x)=47789417)
+17300 val loss 6.4258
+17300 val perplexity 617.5781
+17300 train 6.377968 (lr=7.6260e-05) (hash(x)=52976276)
+17600 val loss 6.2388
+17600 val perplexity 512.2540
+17600 train 6.254847 (lr=5.2854e-05) (hash(x)=49996498)
+16500 val loss 6.1271
+16500 val perplexity 458.0865
+16500 train 6.665648 (lr=3.9117e-05) (hash(x)=59665889)
+17400 val loss 6.4322
+17400 val perplexity 621.5561
+17400 train 6.273139 (lr=7.6009e-05) (hash(x)=52224275)
+17700 val loss 6.2319
+17700 val perplexity 508.6999
+17700 train 6.250253 (lr=5.2677e-05) (hash(x)=50526074)
+16600 val loss 6.1285
+16600 val perplexity 458.7510
+16600 train 6.013433 (lr=3.8995e-05) (hash(x)=49034893)
+17500 val loss 6.4342
+17500 val perplexity 622.7935
+17500 train 6.241991 (lr=7.5758e-05) (hash(x)=49055112)
+17800 val loss 6.2203
+17800 val perplexity 502.8618
+17800 train 6.085866 (lr=5.2499e-05) (hash(x)=48410966)
+16700 val loss 6.1134
+16700 val perplexity 451.8671
+16700 train 5.965143 (lr=3.8873e-05) (hash(x)=45685955)
+17900 val loss 6.2088
+17900 val perplexity 497.1135
+17900 train 6.120636 (lr=5.2321e-05) (hash(x)=51654550)
+17600 val loss 6.4394
+17600 val perplexity 626.0258
+17600 train 6.431095 (lr=7.5505e-05) (hash(x)=49996498)
+16800 val loss 6.1151
+16800 val perplexity 452.6602
+16800 train 5.916518 (lr=3.8750e-05) (hash(x)=49589835)
+18000 val loss 6.2110
+18000 val perplexity 498.2175
+18000 train 6.256108 (lr=5.2142e-05) (hash(x)=44227619)
+17700 val loss 6.4051
+17700 val perplexity 604.9517
+17700 train 6.401339 (lr=7.5252e-05) (hash(x)=50526074)
+16900 val loss 6.1184
+16900 val perplexity 454.1241
+16900 train 5.535720 (lr=3.8627e-05) (hash(x)=38655565)
+18100 val loss 6.2499
+18100 val perplexity 517.9538
+18100 train 6.060187 (lr=5.1962e-05) (hash(x)=45771090)
+17800 val loss 6.3837
+17800 val perplexity 592.0933
+17800 train 6.248812 (lr=7.4998e-05) (hash(x)=48410966)
+18200 val loss 6.2090
+18200 val perplexity 497.1884
+18200 train 5.813533 (lr=5.1782e-05) (hash(x)=37914739)
+17000 val loss 6.1139
+17000 val perplexity 452.0802
+17000 train 6.051076 (lr=3.8503e-05) (hash(x)=46730732)
+17900 val loss 6.3942
+17900 val perplexity 598.3661
+17900 train 6.317099 (lr=7.4744e-05) (hash(x)=51654550)
+18300 val loss 6.2202
+18300 val perplexity 502.8067
+18300 train 6.649755 (lr=5.1602e-05) (hash(x)=68784189)
+17100 val loss 6.1165
+17100 val perplexity 453.2661
+17100 train 5.728023 (lr=3.8379e-05) (hash(x)=43843722)
+18000 val loss 6.3821
+18000 val perplexity 591.1916
+18000 train 6.441657 (lr=7.4488e-05) (hash(x)=44227619)
+18400 val loss 6.2116
+18400 val perplexity 498.4774
+18400 train 6.305532 (lr=5.1421e-05) (hash(x)=55006900)
+18100 val loss 6.3686
+18100 val perplexity 583.2571
+17200 val loss 6.1154
+17200 val perplexity 452.7926
+18100 train 6.152564 (lr=7.4232e-05) (hash(x)=45771090)
+17200 train 5.870935 (lr=3.8255e-05) (hash(x)=50198978)
+18500 val loss 6.2078
+18500 val perplexity 496.5970
+18500 train 6.639764 (lr=5.1240e-05) (hash(x)=56001248)
+18200 val loss 6.3627
+18200 val perplexity 579.8299
+18200 train 5.982582 (lr=7.3975e-05) (hash(x)=37914739)
+17300 val loss 6.1174
+17300 val perplexity 453.6789
+17300 train 6.124028 (lr=3.8130e-05) (hash(x)=52976276)
+18600 val loss 6.2151
+18600 val perplexity 500.2407
+18600 train 6.427946 (lr=5.1058e-05) (hash(x)=51339352)
+18300 val loss 6.3540
+18300 val perplexity 574.8029
+18300 train 6.817049 (lr=7.3717e-05) (hash(x)=68784189)
+17400 val loss 6.1143
+17400 val perplexity 452.2792
+17400 train 5.953983 (lr=3.8005e-05) (hash(x)=52224275)
+18700 val loss 6.1958
+18700 val perplexity 490.6945
+18700 train 6.316683 (lr=5.0875e-05) (hash(x)=53256371)
+18400 val loss 6.3656
+18400 val perplexity 581.4715
+18400 train 6.470822 (lr=7.3459e-05) (hash(x)=55006900)
+17500 val loss 6.1301
+17500 val perplexity 459.4903
+17500 train 5.912059 (lr=3.7879e-05) (hash(x)=49055112)
+18800 val loss 6.1895
+18800 val perplexity 487.5994
+18800 train 6.153018 (lr=5.0692e-05) (hash(x)=51580096)
+18500 val loss 6.3750
+18500 val perplexity 586.9753
+18500 train 6.800819 (lr=7.3199e-05) (hash(x)=56001248)
+17600 val loss 6.1130
+17600 val perplexity 451.6745
+17600 train 6.148704 (lr=3.7753e-05) (hash(x)=49996498)
+18900 val loss 6.2135
+18900 val perplexity 499.4353
+18900 train 6.289693 (lr=5.0509e-05) (hash(x)=53001073)
+18600 val loss 6.3839
+18600 val perplexity 592.2291
+18600 train 6.638814 (lr=7.2939e-05) (hash(x)=51339352)
+17700 val loss 6.1098
+17700 val perplexity 450.2263
+17700 train 6.122671 (lr=3.7626e-05) (hash(x)=50526074)
+19000 val loss 6.1946
+19000 val perplexity 490.0882
+19000 train 6.040896 (lr=5.0325e-05) (hash(x)=51531298)
+18700 val loss 6.3737
+18700 val perplexity 586.2148
+18700 train 6.463876 (lr=7.2679e-05) (hash(x)=53256371)
+17800 val loss 6.0852
+17800 val perplexity 439.3212
+17800 train 5.954217 (lr=3.7499e-05) (hash(x)=48410966)
+19100 val loss 6.1863
+19100 val perplexity 486.0649
+19100 train 5.977464 (lr=5.0140e-05) (hash(x)=48047406)
+18800 val loss 6.3755
+18800 val perplexity 587.2612
+18800 train 6.335759 (lr=7.2417e-05) (hash(x)=51580096)
+17900 val loss 6.0942
+17900 val perplexity 443.2969
+17900 train 5.995052 (lr=3.7372e-05) (hash(x)=51654550)
+19200 val loss 6.2016
+19200 val perplexity 493.5228
+19200 train 5.989591 (lr=4.9955e-05) (hash(x)=44939100)
+18900 val loss 6.4042
+18900 val perplexity 604.3491
+18900 train 6.465156 (lr=7.2155e-05) (hash(x)=53001073)
+18000 val loss 6.0769
+18000 val perplexity 435.6833
+18000 train 6.125412 (lr=3.7244e-05) (hash(x)=44227619)
+19300 val loss 6.1938
+19300 val perplexity 489.6812
+19300 train 6.048463 (lr=4.9770e-05) (hash(x)=50157419)
+19000 val loss 6.3922
+19000 val perplexity 597.1473
+19000 train 6.266483 (lr=7.1892e-05) (hash(x)=51531298)
+18100 val loss 6.0814
+18100 val perplexity 437.6197
+18100 train 5.878476 (lr=3.7116e-05) (hash(x)=45771090)
+19400 val loss 6.2012
+19400 val perplexity 493.3522
+19400 train 5.994554 (lr=4.9584e-05) (hash(x)=46973593)
+19100 val loss 6.4022
+19100 val perplexity 603.1468
+19100 train 6.178774 (lr=7.1629e-05) (hash(x)=48047406)
+18200 val loss 6.0721
+18200 val perplexity 433.5980
+18200 train 5.698674 (lr=3.6987e-05) (hash(x)=37914739)
+19500 val loss 6.1906
+19500 val perplexity 488.1631
+19500 train 5.973404 (lr=4.9398e-05) (hash(x)=44056368)
+19200 val loss 6.4015
+19200 val perplexity 602.7693
+19200 train 6.190560 (lr=7.1365e-05) (hash(x)=44939100)
+18300 val loss 6.0712
+18300 val perplexity 433.1967
+18300 train 6.515245 (lr=3.6859e-05) (hash(x)=68784189)
+19600 val loss 6.1805
+19600 val perplexity 483.2188
+19600 train 6.429302 (lr=4.9211e-05) (hash(x)=58496552)
+19300 val loss 6.3932
+19300 val perplexity 597.7740
+19300 train 6.245660 (lr=7.1100e-05) (hash(x)=50157419)
+18400 val loss 6.0754
+18400 val perplexity 435.0180
+18400 train 6.186811 (lr=3.6729e-05) (hash(x)=55006900)
+19700 val loss 6.1768
+19700 val perplexity 481.4540
+19700 train 6.024097 (lr=4.9024e-05) (hash(x)=47753232)
+19400 val loss 6.3897
+19400 val perplexity 595.6659
+19400 train 6.178835 (lr=7.0835e-05) (hash(x)=46973593)
+18500 val loss 6.0686
+18500 val perplexity 432.0583
+18500 train 6.523821 (lr=3.6600e-05) (hash(x)=56001248)
+19800 val loss 6.1627
+19800 val perplexity 474.7133
+19800 train 5.977198 (lr=4.8837e-05) (hash(x)=49031231)
+19500 val loss 6.3898
+19500 val perplexity 595.7554
+19500 train 6.186742 (lr=7.0569e-05) (hash(x)=44056368)
+19900 val loss 6.1755
+19900 val perplexity 480.8091
+19900 train 6.226699 (lr=4.8649e-05) (hash(x)=54352451)
+18600 val loss 6.1130
+18600 val perplexity 451.6773
+18600 train 6.311148 (lr=3.6470e-05) (hash(x)=51339352)
+19600 val loss 6.3913
+19600 val perplexity 596.6461
+19600 train 6.729703 (lr=7.0302e-05) (hash(x)=58496552)
+20000 val loss 6.1642
+20000 val perplexity 475.4031
+20000 train 5.978202 (lr=4.8461e-05) (hash(x)=45789621)
+18700 val loss 6.0599
+18700 val perplexity 428.3229
+18700 train 6.181108 (lr=3.6339e-05) (hash(x)=53256371)
+19700 val loss 6.3835
+19700 val perplexity 592.0207
+19700 train 6.216898 (lr=7.0035e-05) (hash(x)=47753232)
+20100 val loss 6.1786
+20100 val perplexity 482.3323
+20100 train 6.119834 (lr=4.8272e-05) (hash(x)=45096720)
+18800 val loss 6.0531
+18800 val perplexity 425.4240
+18800 train 6.023508 (lr=3.6209e-05) (hash(x)=51580096)
+19800 val loss 6.3565
+19800 val perplexity 576.2151
+19800 train 6.147184 (lr=6.9767e-05) (hash(x)=49031231)
+20200 val loss 6.1726
+20200 val perplexity 479.4096
+20200 train 6.233849 (lr=4.8083e-05) (hash(x)=49377287)
+18900 val loss 6.0609
+18900 val perplexity 428.7735
+18900 train 6.137412 (lr=3.6078e-05) (hash(x)=53001073)
+19900 val loss 6.3602
+19900 val perplexity 578.3471
+19900 train 6.447223 (lr=6.9498e-05) (hash(x)=54352451)
+20300 val loss 6.1640
+20300 val perplexity 475.3399
+20300 train 5.945217 (lr=4.7893e-05) (hash(x)=49454716)
+19000 val loss 6.0583
+19000 val perplexity 427.6386
+19000 train 5.912615 (lr=3.5946e-05) (hash(x)=51531298)
+20000 val loss 6.3695
+20000 val perplexity 583.7882
+20000 train 6.155598 (lr=6.9229e-05) (hash(x)=45789621)
+20400 val loss 6.1713
+20400 val perplexity 478.8310
+20400 train 5.680020 (lr=4.7703e-05) (hash(x)=38699401)
+19100 val loss 6.0597
+19100 val perplexity 428.2618
+19100 train 5.843977 (lr=3.5814e-05) (hash(x)=48047406)
+20100 val loss 6.3631
+20100 val perplexity 580.0168
+20100 train 6.303806 (lr=6.8960e-05) (hash(x)=45096720)
+20500 val loss 6.1714
+20500 val perplexity 478.8588
+20500 train 6.270490 (lr=4.7513e-05) (hash(x)=53266981)
+20200 val loss 6.3630
+20200 val perplexity 579.9930
+19200 val loss 6.0528
+19200 val perplexity 425.3007
+19200 train 5.875459 (lr=3.5682e-05) (hash(x)=44939100)
+20200 train 6.376566 (lr=6.8690e-05) (hash(x)=49377287)
+20600 val loss 6.1660
+20600 val perplexity 476.2712
+20600 train 6.099114 (lr=4.7323e-05) (hash(x)=51783804)
+20300 val loss 6.3610
+20300 val perplexity 578.8163
+20300 train 6.128880 (lr=6.8419e-05) (hash(x)=49454716)
+19300 val loss 6.0601
+19300 val perplexity 428.3989
+19300 train 5.926717 (lr=3.5550e-05) (hash(x)=50157419)
+20700 val loss 6.1819
+20700 val perplexity 483.8918
+20700 train 6.071718 (lr=4.7132e-05) (hash(x)=51388467)
+20400 val loss 6.3674
+20400 val perplexity 582.5652
+20400 train 5.859035 (lr=6.8148e-05) (hash(x)=38699401)
+19400 val loss 6.0600
+19400 val perplexity 428.3746
+19400 train 5.875482 (lr=3.5417e-05) (hash(x)=46973593)
+20800 val loss 6.1892
+20800 val perplexity 487.4608
+20800 train 5.933630 (lr=4.6940e-05) (hash(x)=45324903)
+20500 val loss 6.3694
+20500 val perplexity 583.7231
+20500 train 6.425434 (lr=6.7876e-05) (hash(x)=53266981)
+19500 val loss 6.0467
+19500 val perplexity 422.7293
+19500 train 5.862154 (lr=3.5284e-05) (hash(x)=44056368)
+20900 val loss 6.1733
+20900 val perplexity 479.7522
+20900 train 5.804732 (lr=4.6749e-05) (hash(x)=45972467)
+20600 val loss 6.3675
+20600 val perplexity 582.6213
+20600 train 6.303026 (lr=6.7604e-05) (hash(x)=51783804)
+19600 val loss 6.0503
+19600 val perplexity 424.2211
+19600 train 6.313962 (lr=3.5151e-05) (hash(x)=58496552)
+21000 val loss 6.1696
+21000 val perplexity 477.9921
+21000 train 6.117172 (lr=4.6557e-05) (hash(x)=46754507)
+20700 val loss 6.3765
+20700 val perplexity 587.8462
+20700 train 6.280571 (lr=6.7331e-05) (hash(x)=51388467)
+19700 val loss 6.0547
+19700 val perplexity 426.1084
+19700 train 5.911861 (lr=3.5017e-05) (hash(x)=47753232)
+21100 val loss 6.1708
+21100 val perplexity 478.5657
+21100 train 6.458302 (lr=4.6365e-05) (hash(x)=59913667)
+20800 val loss 6.3633
+20800 val perplexity 580.1413
+20800 train 6.099911 (lr=6.7058e-05) (hash(x)=45324903)
+19800 val loss 6.0476
+19800 val perplexity 423.1147
+19800 train 5.875927 (lr=3.4883e-05) (hash(x)=49031231)
+21200 val loss 6.1723
+21200 val perplexity 479.3056
+21200 train 6.133057 (lr=4.6172e-05) (hash(x)=46487759)
+20900 val loss 6.3644
+20900 val perplexity 580.7993
+20900 train 5.990748 (lr=6.6784e-05) (hash(x)=45972467)
+19900 val loss 6.0356
+19900 val perplexity 418.0502
+19900 train 6.095958 (lr=3.4749e-05) (hash(x)=54352451)
+21300 val loss 6.1538
+21300 val perplexity 470.5030
+21300 train 5.945532 (lr=4.5979e-05) (hash(x)=45535495)
+21000 val loss 6.3754
+21000 val perplexity 587.2021
+21000 train 6.294639 (lr=6.6510e-05) (hash(x)=46754507)
+20000 val loss 6.0344
+20000 val perplexity 417.5580
+21400 val loss 6.1425
+21400 val perplexity 465.2253
+20000 train 5.854521 (lr=3.4615e-05) (hash(x)=45789621)
+21400 train 6.150859 (lr=4.5786e-05) (hash(x)=54559895)
+21100 val loss 6.3533
+21100 val perplexity 574.3936
+21100 train 6.633968 (lr=6.6235e-05) (hash(x)=59913667)
+21500 val loss 6.1391
+21500 val perplexity 463.6258
+21500 train 5.982634 (lr=4.5592e-05) (hash(x)=48474176)
+20100 val loss 6.0352
+20100 val perplexity 417.8784
+20100 train 6.019393 (lr=3.4480e-05) (hash(x)=45096720)
+21200 val loss 6.3447
+21200 val perplexity 569.4678
+21200 train 6.289260 (lr=6.5960e-05) (hash(x)=46487759)
+21600 val loss 6.1398
+21600 val perplexity 463.9677
+21600 train 6.109609 (lr=4.5399e-05) (hash(x)=47915584)
+20200 val loss 6.0307
+20200 val perplexity 416.0139
+20200 train 6.110758 (lr=3.4345e-05) (hash(x)=49377287)
+21300 val loss 6.3526
+21300 val perplexity 573.9709
+21300 train 6.203497 (lr=6.5684e-05) (hash(x)=45535495)
+21700 val loss 6.1441
+21700 val perplexity 465.9380
+21700 train 5.896945 (lr=4.5205e-05) (hash(x)=45846696)
+20300 val loss 6.0276
+20300 val perplexity 414.7291
+20300 train 5.837848 (lr=3.4209e-05) (hash(x)=49454716)
+21400 val loss 6.3514
+21400 val perplexity 573.3196
+21400 train 6.349491 (lr=6.5408e-05) (hash(x)=54559895)
+21800 val loss 6.1466
+21800 val perplexity 467.1121
+21800 train 5.892045 (lr=4.5010e-05) (hash(x)=47201143)
+20400 val loss 6.0263
+20400 val perplexity 414.1902
+20400 train 5.551824 (lr=3.4074e-05) (hash(x)=38699401)
+21500 val loss 6.3461
+21500 val perplexity 570.2754
+21500 train 6.196745 (lr=6.5132e-05) (hash(x)=48474176)
+21900 val loss 6.1458
+21900 val perplexity 466.7358
+21900 train 6.086146 (lr=4.4816e-05) (hash(x)=50028395)
+20500 val loss 6.0176
+20500 val perplexity 410.5836
+20500 train 6.118743 (lr=3.3938e-05) (hash(x)=53266981)
+21600 val loss 6.3470
+21600 val perplexity 570.7966
+21600 train 6.301847 (lr=6.4855e-05) (hash(x)=47915584)
+22000 val loss 6.1393
+22000 val perplexity 463.7465
+22000 train 6.101057 (lr=4.4621e-05) (hash(x)=50273781)
+20600 val loss 6.0144
+20600 val perplexity 409.2816
+20600 train 5.957831 (lr=3.3802e-05) (hash(x)=51783804)
+21700 val loss 6.3332
+21700 val perplexity 562.9777
+21700 train 6.098959 (lr=6.4578e-05) (hash(x)=45846696)
+22100 val loss 6.1502
+22100 val perplexity 468.8317
+22100 train 6.074613 (lr=4.4426e-05) (hash(x)=47859152)
+20700 val loss 6.0155
+20700 val perplexity 409.7299
+20700 train 5.928106 (lr=3.3665e-05) (hash(x)=51388467)
+21800 val loss 6.3397
+21800 val perplexity 566.6172
+21800 train 6.066625 (lr=6.4300e-05) (hash(x)=47201143)
+22200 val loss 6.1422
+22200 val perplexity 465.0601
+22200 train 6.002366 (lr=4.4231e-05) (hash(x)=47820296)
+20800 val loss 6.0164
+20800 val perplexity 410.0995
+20800 train 5.793630 (lr=3.3529e-05) (hash(x)=45324903)
+21900 val loss 6.3431
+21900 val perplexity 568.5334
+21900 train 6.304872 (lr=6.4023e-05) (hash(x)=50028395)
+22300 val loss 6.1341
+22300 val perplexity 461.3391
+22300 train 5.852176 (lr=4.4035e-05) (hash(x)=51741508)
+20900 val loss 6.0135
+20900 val perplexity 408.9259
+20900 train 5.664894 (lr=3.3392e-05) (hash(x)=45972467)
+22000 val loss 6.3380
+22000 val perplexity 565.6362
+22000 train 6.293690 (lr=6.3744e-05) (hash(x)=50273781)
+22400 val loss 6.1399
+22400 val perplexity 464.0199
+22400 train 5.707395 (lr=4.3839e-05) (hash(x)=44585679)
+21000 val loss 6.0136
+21000 val perplexity 408.9552
+21000 train 5.986984 (lr=3.3255e-05) (hash(x)=46754507)
+22100 val loss 6.3400
+22100 val perplexity 566.7731
+22100 train 6.252201 (lr=6.3466e-05) (hash(x)=47859152)
+22500 val loss 6.1268
+22500 val perplexity 457.9565
+22500 train 6.158222 (lr=4.3643e-05) (hash(x)=48753202)
+21100 val loss 6.0039
+21100 val perplexity 405.0070
+21100 train 6.288526 (lr=3.3118e-05) (hash(x)=59913667)
+22200 val loss 6.3427
+22200 val perplexity 568.3276
+22200 train 6.203479 (lr=6.3187e-05) (hash(x)=47820296)
+22600 val loss 6.1390
+22600 val perplexity 463.5871
+22600 train 5.701408 (lr=4.3447e-05) (hash(x)=45351902)
+21200 val loss 6.0067
+21200 val perplexity 406.1277
+21200 train 6.005379 (lr=3.2980e-05) (hash(x)=46487759)
+22300 val loss 6.3471
+22300 val perplexity 570.8563
+22300 train 6.079469 (lr=6.2907e-05) (hash(x)=51741508)
+22700 val loss 6.1328
+22700 val perplexity 460.7227
+22700 train 6.620811 (lr=4.3251e-05) (hash(x)=59263323)
+21300 val loss 6.0122
+21300 val perplexity 408.3809
+21300 train 5.797719 (lr=3.2842e-05) (hash(x)=45535495)
+22400 val loss 6.3355
+22400 val perplexity 564.2387
+22400 train 5.933209 (lr=6.2628e-05) (hash(x)=44585679)
+22800 val loss 6.1419
+22800 val perplexity 464.9315
+22800 train 5.694124 (lr=4.3054e-05) (hash(x)=44873637)
+22500 val loss 6.3426
+22500 val perplexity 568.2452
+21400 val loss 6.0007
+21400 val perplexity 403.6951
+21400 train 6.002113 (lr=3.2704e-05) (hash(x)=54559895)
+22500 train 6.350743 (lr=6.2348e-05) (hash(x)=48753202)
+22900 val loss 6.1300
+22900 val perplexity 459.4432
+22900 train 6.244625 (lr=4.2858e-05) (hash(x)=58208930)
+21500 val loss 5.9900
+21500 val perplexity 399.4275
+21500 train 5.842444 (lr=3.2566e-05) (hash(x)=48474176)
+22600 val loss 6.3425
+22600 val perplexity 568.2057
+22600 train 5.930212 (lr=6.2068e-05) (hash(x)=45351902)
+23000 val loss 6.1255
+23000 val perplexity 457.3562
+23000 train 6.102044 (lr=4.2661e-05) (hash(x)=50248612)
+21600 val loss 5.9876
+21600 val perplexity 398.4431
+21600 train 5.968499 (lr=3.2428e-05) (hash(x)=47915584)
+23100 val loss 6.1814
+23100 val perplexity 483.6907
+22700 val loss 6.3480
+22700 val perplexity 571.3630
+23100 train 6.055898 (lr=4.2464e-05) (hash(x)=47902220)
+22700 train 6.802013 (lr=6.1787e-05) (hash(x)=59263323)
+21700 val loss 5.9948
+21700 val perplexity 401.3510
+21700 train 5.751846 (lr=3.2289e-05) (hash(x)=45846696)
+23200 val loss 6.1135
+23200 val perplexity 451.9276
+23200 train 5.991982 (lr=4.2267e-05) (hash(x)=45606218)
+22800 val loss 6.3393
+22800 val perplexity 566.3903
+22800 train 5.898570 (lr=6.1506e-05) (hash(x)=44873637)
+23300 val loss 6.1096
+23300 val perplexity 450.1411
+23300 train 6.233109 (lr=4.2069e-05) (hash(x)=55596334)
+21800 val loss 5.9915
+21800 val perplexity 400.0187
+21800 train 5.753116 (lr=3.2150e-05) (hash(x)=47201143)
+22900 val loss 6.3156
+22900 val perplexity 553.1541
+22900 train 6.436237 (lr=6.1225e-05) (hash(x)=58208930)
+23400 val loss 6.1044
+23400 val perplexity 447.8400
+23400 train 6.133411 (lr=4.1872e-05) (hash(x)=59736985)
+21900 val loss 5.9855
+21900 val perplexity 397.6227
+23000 val loss 6.3196
+23000 val perplexity 555.3450
+21900 train 5.923649 (lr=3.2011e-05) (hash(x)=50028395)
+23000 train 6.307461 (lr=6.0944e-05) (hash(x)=50248612)
+23500 val loss 6.1119
+23500 val perplexity 451.1934
+23500 train 6.046716 (lr=4.1674e-05) (hash(x)=48935694)
+22000 val loss 5.9808
+22000 val perplexity 395.7760
+23100 val loss 6.3458
+23100 val perplexity 570.0873
+22000 train 5.972328 (lr=3.1872e-05) (hash(x)=50273781)
+23100 train 6.207120 (lr=6.0663e-05) (hash(x)=47902220)
+23600 val loss 6.1151
+23600 val perplexity 452.6223
+23600 train 5.924499 (lr=4.1476e-05) (hash(x)=53977113)
+23200 val loss 6.3292
+23200 val perplexity 560.6997
+22100 val loss 5.9880
+22100 val perplexity 398.6314
+23200 train 6.211708 (lr=6.0381e-05) (hash(x)=45606218)
+22100 train 5.931949 (lr=3.1733e-05) (hash(x)=47859152)
+23700 val loss 6.1222
+23700 val perplexity 455.8661
+23700 train 5.764342 (lr=4.1278e-05) (hash(x)=45466992)
+22200 val loss 5.9857
+22200 val perplexity 397.6931
+23300 val loss 6.3152
+23300 val perplexity 552.8865
+22200 train 5.832684 (lr=3.1593e-05) (hash(x)=47820296)
+23300 train 6.430233 (lr=6.0099e-05) (hash(x)=55596334)
+23800 val loss 6.1038
+23800 val perplexity 447.5618
+23800 train 5.909907 (lr=4.1080e-05) (hash(x)=45830878)
+23400 val loss 6.3127
+23400 val perplexity 551.5491
+23400 train 6.347643 (lr=5.9817e-05) (hash(x)=59736985)
+22300 val loss 5.9815
+22300 val perplexity 396.0442
+22300 train 5.725097 (lr=3.1454e-05) (hash(x)=51741508)
+23900 val loss 6.1088
+23900 val perplexity 449.8033
+23900 train 6.000215 (lr=4.0882e-05) (hash(x)=48268066)
+23500 val loss 6.3050
+23500 val perplexity 547.3044
+23500 train 6.209606 (lr=5.9534e-05) (hash(x)=48935694)
+22400 val loss 5.9881
+22400 val perplexity 398.6481
+22400 train 5.545953 (lr=3.1314e-05) (hash(x)=44585679)
+24000 val loss 6.1128
+24000 val perplexity 451.5821
+24000 train 5.937686 (lr=4.0684e-05) (hash(x)=51062942)
+22500 val loss 5.9894
+22500 val perplexity 399.1792
+22500 train 6.055580 (lr=3.1174e-05) (hash(x)=48753202)
+23600 val loss 6.2960
+23600 val perplexity 542.4075
+23600 train 6.142959 (lr=5.9252e-05) (hash(x)=53977113)
+24100 val loss 6.1118
+24100 val perplexity 451.1725
+24100 train 5.957380 (lr=4.0486e-05) (hash(x)=51090246)
+23700 val loss 6.3045
+23700 val perplexity 547.0236
+23700 train 5.952677 (lr=5.8969e-05) (hash(x)=45466992)
+22600 val loss 5.9757
+22600 val perplexity 393.7350
+22600 train 5.555033 (lr=3.1034e-05) (hash(x)=45351902)
+24200 val loss 6.1143
+24200 val perplexity 452.2881
+24200 train 5.828847 (lr=4.0287e-05) (hash(x)=48686225)
+22700 val loss 5.9891
+22700 val perplexity 399.0443
+22700 train 6.472228 (lr=3.0894e-05) (hash(x)=59263323)
+23800 val loss 6.3023
+23800 val perplexity 545.8222
+23800 train 6.130695 (lr=5.8686e-05) (hash(x)=45830878)
+24300 val loss 6.1108
+24300 val perplexity 450.7126
+24300 train 5.765564 (lr=4.0089e-05) (hash(x)=43604814)
+22800 val loss 5.9729
+22800 val perplexity 392.6522
+22800 train 5.568539 (lr=3.0753e-05) (hash(x)=44873637)
+23900 val loss 6.2997
+23900 val perplexity 544.4222
+23900 train 6.194791 (lr=5.8403e-05) (hash(x)=48268066)
+24400 val loss 6.1000
+24400 val perplexity 445.8465
+24400 train 5.904193 (lr=3.9891e-05) (hash(x)=48143655)
+22900 val loss 5.9682
+22900 val perplexity 390.7981
+22900 train 6.094232 (lr=3.0613e-05) (hash(x)=58208930)
+24000 val loss 6.3047
+24000 val perplexity 547.1176
+24000 train 6.130448 (lr=5.8120e-05) (hash(x)=51062942)
+24500 val loss 6.1090
+24500 val perplexity 449.8874
+24500 train 6.140269 (lr=3.9692e-05) (hash(x)=58317397)
+23000 val loss 5.9644
+23000 val perplexity 389.3314
+23000 train 5.966906 (lr=3.0472e-05) (hash(x)=50248612)
+24100 val loss 6.3059
+24100 val perplexity 547.7991
+24100 train 6.162120 (lr=5.7837e-05) (hash(x)=51090246)
+24600 val loss 6.1055
+24600 val perplexity 448.3269
+24600 train 6.242974 (lr=3.9493e-05) (hash(x)=53855425)
+23100 val loss 5.9697
+23100 val perplexity 391.3837
+23100 train 5.868172 (lr=3.0331e-05) (hash(x)=47902220)
+24200 val loss 6.3250
+24200 val perplexity 558.3798
+24200 train 6.033810 (lr=5.7554e-05) (hash(x)=48686225)
+24700 val loss 6.1025
+24700 val perplexity 446.9740
+24700 train 5.891122 (lr=3.9295e-05) (hash(x)=49829785)
+23200 val loss 5.9563
+23200 val perplexity 386.1695
+23200 train 5.812859 (lr=3.0190e-05) (hash(x)=45606218)
+24800 val loss 6.0966
+24800 val perplexity 444.3436
+24800 train 6.130437 (lr=3.9096e-05) (hash(x)=51337298)
+24300 val loss 6.3078
+24300 val perplexity 548.8285
+24300 train 5.955484 (lr=5.7270e-05) (hash(x)=43604814)
+23300 val loss 5.9550
+23300 val perplexity 385.6635
+23300 train 6.090078 (lr=3.0049e-05) (hash(x)=55596334)
+24900 val loss 6.0916
+24900 val perplexity 442.1438
+24900 train 6.120977 (lr=3.8897e-05) (hash(x)=47385279)
+24400 val loss 6.2998
+24400 val perplexity 544.4456
+24400 train 6.106803 (lr=5.6987e-05) (hash(x)=48143655)
+23400 val loss 5.9525
+23400 val perplexity 384.7060
+23400 train 5.954114 (lr=2.9908e-05) (hash(x)=59736985)
+25000 val loss 6.0797
+25000 val perplexity 436.8879
+25000 train 6.504747 (lr=3.8699e-05) (hash(x)=53631975)
+24500 val loss 6.2944
+24500 val perplexity 541.5116
+24500 train 6.330753 (lr=5.6703e-05) (hash(x)=58317397)
+23500 val loss 5.9520
+23500 val perplexity 384.5178
+23500 train 5.869464 (lr=2.9767e-05) (hash(x)=48935694)
+25100 val loss 6.0752
+25100 val perplexity 434.9232
+25100 train 6.107512 (lr=3.8500e-05) (hash(x)=52991852)
+24600 val loss 6.2925
+24600 val perplexity 540.5165
+24600 train 6.439625 (lr=5.6419e-05) (hash(x)=53855425)
+25200 val loss 6.0788
+25200 val perplexity 436.5062
+25200 train 6.457631 (lr=3.8301e-05) (hash(x)=51863290)
+23600 val loss 5.9498
+23600 val perplexity 383.6585
+23600 train 5.757344 (lr=2.9626e-05) (hash(x)=53977113)
+24700 val loss 6.3028
+24700 val perplexity 546.1091
+24700 train 6.100008 (lr=5.6135e-05) (hash(x)=49829785)
+25300 val loss 6.0786
+25300 val perplexity 436.4076
+25300 train 5.847868 (lr=3.8103e-05) (hash(x)=46898810)
+23700 val loss 5.9542
+23700 val perplexity 385.3510
+23700 train 5.572944 (lr=2.9485e-05) (hash(x)=45466992)
+24800 val loss 6.2816
+24800 val perplexity 534.6320
+24800 train 6.316615 (lr=5.5852e-05) (hash(x)=51337298)
+25400 val loss 6.0781
+25400 val perplexity 436.2205
+25400 train 6.139091 (lr=3.7904e-05) (hash(x)=50297951)
+24900 val loss 6.2896
+24900 val perplexity 538.9402
+24900 train 6.324134 (lr=5.5568e-05) (hash(x)=47385279)
+23800 val loss 5.9453
+23800 val perplexity 381.9656
+23800 train 5.733608 (lr=2.9343e-05) (hash(x)=45830878)
+25500 val loss 6.0659
+25500 val perplexity 430.9047
+25500 train 5.887280 (lr=3.7705e-05) (hash(x)=49942718)
+25000 val loss 6.2820
+25000 val perplexity 534.8416
+23900 val loss 5.9411
+23900 val perplexity 380.3693
+23900 train 5.839910 (lr=2.9202e-05) (hash(x)=48268066)
+25000 train 6.650968 (lr=5.5284e-05) (hash(x)=53631975)
+25600 val loss 6.0695
+25600 val perplexity 432.4829
+25600 train 5.939018 (lr=3.7507e-05) (hash(x)=46944979)
+24000 val loss 5.9383
+24000 val perplexity 379.2958
+24000 train 5.761079 (lr=2.9060e-05) (hash(x)=51062942)
+25100 val loss 6.2794
+25100 val perplexity 533.4435
+25100 train 6.317781 (lr=5.5000e-05) (hash(x)=52991852)
+25700 val loss 6.0792
+25700 val perplexity 436.6661
+25700 train 6.337087 (lr=3.7308e-05) (hash(x)=48116500)
+24100 val loss 5.9419
+24100 val perplexity 380.6670
+25200 val loss 6.2777
+25200 val perplexity 532.5540
+24100 train 5.817915 (lr=2.8918e-05) (hash(x)=51090246)
+25200 train 6.661186 (lr=5.4716e-05) (hash(x)=51863290)
+25800 val loss 6.0654
+25800 val perplexity 430.6810
+25800 train 6.344007 (lr=3.7109e-05) (hash(x)=59750893)
+25300 val loss 6.2835
+25300 val perplexity 535.6470
+25300 train 6.038043 (lr=5.4432e-05) (hash(x)=46898810)
+24200 val loss 5.9475
+24200 val perplexity 382.7947
+24200 train 5.666345 (lr=2.8777e-05) (hash(x)=48686225)
+25900 val loss 6.0707
+25900 val perplexity 432.9895
+25900 train 5.797611 (lr=3.6911e-05) (hash(x)=49479938)
+25400 val loss 6.2828
+25400 val perplexity 535.2632
+25400 train 6.348429 (lr=5.4148e-05) (hash(x)=50297951)
+24300 val loss 5.9399
+24300 val perplexity 379.9096
+24300 train 5.606017 (lr=2.8635e-05) (hash(x)=43604814)
+26000 val loss 6.0766
+26000 val perplexity 435.5415
+26000 train 6.209918 (lr=3.6713e-05) (hash(x)=51134577)
+25500 val loss 6.2826
+25500 val perplexity 535.1839
+25500 train 6.111054 (lr=5.3865e-05) (hash(x)=49942718)
+24400 val loss 5.9440
+24400 val perplexity 381.4422
+24400 train 5.755672 (lr=2.8493e-05) (hash(x)=48143655)
+26100 val loss 6.0689
+26100 val perplexity 432.1848
+26100 train 5.996790 (lr=3.6514e-05) (hash(x)=53677080)
+25600 val loss 6.2769
+25600 val perplexity 532.1389
+25600 train 6.146551 (lr=5.3581e-05) (hash(x)=46944979)
+24500 val loss 5.9383
+24500 val perplexity 379.2902
+24500 train 5.992229 (lr=2.8351e-05) (hash(x)=58317397)
+26200 val loss 6.0752
+26200 val perplexity 434.9546
+26200 train 5.937295 (lr=3.6316e-05) (hash(x)=48885091)
+25700 val loss 6.2790
+25700 val perplexity 533.2624
+24600 val loss 5.9361
+24600 val perplexity 378.4429
+25700 train 6.553234 (lr=5.3297e-05) (hash(x)=48116500)
+24600 train 6.060189 (lr=2.8210e-05) (hash(x)=53855425)
+26300 val loss 6.0637
+26300 val perplexity 429.9825
+26300 train 6.134348 (lr=3.6118e-05) (hash(x)=57217953)
+24700 val loss 5.9514
+24700 val perplexity 384.2784
+24700 train 5.756442 (lr=2.8068e-05) (hash(x)=49829785)
+25800 val loss 6.2791
+25800 val perplexity 533.2939
+25800 train 6.653417 (lr=5.3013e-05) (hash(x)=59750893)
+26400 val loss 6.0595
+26400 val perplexity 428.1722
+26400 train 5.841461 (lr=3.5920e-05) (hash(x)=48504109)
+24800 val loss 5.9219
+24800 val perplexity 373.1087
+25900 val loss 6.2764
+25900 val perplexity 531.8708
+24800 train 5.968436 (lr=2.7926e-05) (hash(x)=51337298)
+25900 train 6.000820 (lr=5.2730e-05) (hash(x)=49479938)
+26500 val loss 6.0597
+26500 val perplexity 428.2265
+26500 train 6.635356 (lr=3.5722e-05) (hash(x)=55775918)
+24900 val loss 5.9406
+24900 val perplexity 380.1450
+24900 train 5.984793 (lr=2.7784e-05) (hash(x)=47385279)
+26000 val loss 6.2688
+26000 val perplexity 527.8687
+26000 train 6.396154 (lr=5.2446e-05) (hash(x)=51134577)
+26600 val loss 6.0557
+26600 val perplexity 426.5464
+26600 train 6.022480 (lr=3.5524e-05) (hash(x)=51978817)
+25000 val loss 5.9243
+25000 val perplexity 374.0070
+25000 train 6.377190 (lr=2.7642e-05) (hash(x)=53631975)
+26100 val loss 6.2779
+26100 val perplexity 532.6611
+26100 train 6.178829 (lr=5.2163e-05) (hash(x)=53677080)
+26700 val loss 6.0598
+26700 val perplexity 428.2859
+26700 train 5.925630 (lr=3.5326e-05) (hash(x)=50995814)
+25100 val loss 5.9203
+25100 val perplexity 372.5377
+25100 train 5.939419 (lr=2.7500e-05) (hash(x)=52991852)
+26200 val loss 6.2773
+26200 val perplexity 532.3508
+26800 val loss 6.0507
+26800 val perplexity 424.4202
+26200 train 6.118965 (lr=5.1880e-05) (hash(x)=48885091)
+26800 train 5.966284 (lr=3.5128e-05) (hash(x)=49986643)
+26900 val loss 6.0512
+26900 val perplexity 424.6243
+26900 train 6.157857 (lr=3.4931e-05) (hash(x)=51408788)
+25200 val loss 5.9152
+25200 val perplexity 370.6218
+25200 train 6.300425 (lr=2.7358e-05) (hash(x)=51863290)
+26300 val loss 6.2773
+26300 val perplexity 532.3270
+26300 train 6.354596 (lr=5.1597e-05) (hash(x)=57217953)
+27000 val loss 6.0416
+27000 val perplexity 420.5643
+27000 train 5.814680 (lr=3.4733e-05) (hash(x)=49024980)
+25300 val loss 5.9153
+25300 val perplexity 370.6690
+26400 val loss 6.2635
+26400 val perplexity 525.0340
+25300 train 5.684647 (lr=2.7216e-05) (hash(x)=46898810)
+26400 train 6.056812 (lr=5.1314e-05) (hash(x)=48504109)
+27100 val loss 6.0444
+27100 val perplexity 421.7439
+27100 train 6.281505 (lr=3.4536e-05) (hash(x)=45027677)
+26500 val loss 6.2579
+26500 val perplexity 522.1194
+26500 train 6.795272 (lr=5.1031e-05) (hash(x)=55775918)
+25400 val loss 5.9204
+25400 val perplexity 372.5515
+25400 train 6.009842 (lr=2.7074e-05) (hash(x)=50297951)
+27200 val loss 6.0484
+27200 val perplexity 423.4525
+27200 train 6.103629 (lr=3.4339e-05) (hash(x)=49930937)
+26600 val loss 6.2578
+26600 val perplexity 522.0460
+26600 train 6.216881 (lr=5.0748e-05) (hash(x)=51978817)
+25500 val loss 5.9114
+25500 val perplexity 369.2381
+25500 train 5.744133 (lr=2.6932e-05) (hash(x)=49942718)
+27300 val loss 6.0353
+27300 val perplexity 417.9277
+27300 train 5.947425 (lr=3.4142e-05) (hash(x)=51016030)
+26700 val loss 6.2640
+26700 val perplexity 525.3312
+26700 train 6.130256 (lr=5.0466e-05) (hash(x)=50995814)
+25600 val loss 5.9158
+25600 val perplexity 370.8578
+25600 train 5.784765 (lr=2.6790e-05) (hash(x)=46944979)
+27400 val loss 6.0351
+27400 val perplexity 417.8452
+27400 train 6.264123 (lr=3.3946e-05) (hash(x)=48154914)
+26800 val loss 6.2422
+26800 val perplexity 513.9755
+26800 train 6.156603 (lr=5.0183e-05) (hash(x)=49986643)
+25700 val loss 5.9091
+25700 val perplexity 368.3661
+25700 train 6.181494 (lr=2.6649e-05) (hash(x)=48116500)
+27500 val loss 6.0305
+27500 val perplexity 415.9289
+27500 train 6.128118 (lr=3.3749e-05) (hash(x)=53265478)
+26900 val loss 6.2369
+26900 val perplexity 511.2696
+26900 train 6.337112 (lr=4.9901e-05) (hash(x)=51408788)
+25800 val loss 5.9088
+25800 val perplexity 368.2530
+25800 train 6.189940 (lr=2.6507e-05) (hash(x)=59750893)
+27600 val loss 6.0527
+27600 val perplexity 425.2629
+27600 train 5.856176 (lr=3.3553e-05) (hash(x)=39456920)
+27000 val loss 6.2410
+27000 val perplexity 513.3475
+27000 train 6.025411 (lr=4.9619e-05) (hash(x)=49024980)
+25900 val loss 5.9107
+25900 val perplexity 368.9599
+25900 train 5.638449 (lr=2.6365e-05) (hash(x)=49479938)
+27700 val loss 6.0435
+27700 val perplexity 421.3598
+27700 train 6.216787 (lr=3.3357e-05) (hash(x)=55794844)
+27100 val loss 6.2409
+27100 val perplexity 513.3365
+27100 train 6.460735 (lr=4.9337e-05) (hash(x)=45027677)
+26000 val loss 5.9132
+26000 val perplexity 369.8796
+26000 train 6.063453 (lr=2.6223e-05) (hash(x)=51134577)
+27800 val loss 6.0393
+27800 val perplexity 419.5840
+27800 train 5.958351 (lr=3.3161e-05) (hash(x)=49091362)
+27200 val loss 6.2371
+27200 val perplexity 511.3715
+26100 val loss 5.9118
+26100 val perplexity 369.3658
+27200 train 6.283369 (lr=4.9056e-05) (hash(x)=49930937)
+26100 train 5.850782 (lr=2.6082e-05) (hash(x)=53677080)
+27900 val loss 6.0371
+27900 val perplexity 418.6732
+27900 train 5.852857 (lr=3.2965e-05) (hash(x)=49494590)
+26200 val loss 5.9137
+26200 val perplexity 370.0862
+26200 train 5.798280 (lr=2.5940e-05) (hash(x)=48885091)
+27300 val loss 6.2352
+27300 val perplexity 510.3822
+27300 train 6.146107 (lr=4.8775e-05) (hash(x)=51016030)
+28000 val loss 6.0347
+28000 val perplexity 417.6547
+28000 train 5.417509 (lr=3.2769e-05) (hash(x)=42333356)
+26300 val loss 5.9087
+26300 val perplexity 368.2252
+27400 val loss 6.2294
+27400 val perplexity 507.4462
+26300 train 5.979610 (lr=2.5798e-05) (hash(x)=57217953)
+27400 train 6.446771 (lr=4.8494e-05) (hash(x)=48154914)
+28100 val loss 6.0293
+28100 val perplexity 415.4191
+28100 train 6.210233 (lr=3.2574e-05) (hash(x)=49378958)
+26400 val loss 5.9082
+26400 val perplexity 368.0288
+26400 train 5.696656 (lr=2.5657e-05) (hash(x)=48504109)
+27500 val loss 6.2373
+27500 val perplexity 511.4503
+27500 train 6.334397 (lr=4.8213e-05) (hash(x)=53265478)
+28200 val loss 6.0243
+28200 val perplexity 413.3595
+28200 train 6.198831 (lr=3.2379e-05) (hash(x)=56595308)
+26500 val loss 5.8980
+26500 val perplexity 364.3212
+26500 train 6.510257 (lr=2.5515e-05) (hash(x)=55775918)
+27600 val loss 6.2322
+27600 val perplexity 508.8639
+27600 train 6.047086 (lr=4.7932e-05) (hash(x)=39456920)
+28300 val loss 6.0461
+28300 val perplexity 422.4748
+28300 train 6.702070 (lr=3.2184e-05) (hash(x)=52311024)
+26600 val loss 5.8978
+26600 val perplexity 364.2238
+26600 train 5.865256 (lr=2.5374e-05) (hash(x)=51978817)
+27700 val loss 6.2495
+27700 val perplexity 517.7338
+27700 train 6.436790 (lr=4.7652e-05) (hash(x)=55794844)
+28400 val loss 6.0277
+28400 val perplexity 414.7752
+28400 train 6.030992 (lr=3.1990e-05) (hash(x)=47997766)
+26700 val loss 5.9000
+26700 val perplexity 365.0312
+26700 train 5.699328 (lr=2.5233e-05) (hash(x)=50995814)
+27800 val loss 6.2441
+27800 val perplexity 514.9801
+27800 train 6.175504 (lr=4.7372e-05) (hash(x)=49091362)
+28500 val loss 6.0278
+28500 val perplexity 414.7870
+28500 train 5.981230 (lr=3.1795e-05) (hash(x)=46906589)
+26800 val loss 5.8952
+26800 val perplexity 363.3040
+26800 train 5.803399 (lr=2.5092e-05) (hash(x)=49986643)
+27900 val loss 6.2364
+27900 val perplexity 511.0261
+28600 val loss 6.0144
+28600 val perplexity 409.3001
+27900 train 6.035953 (lr=4.7093e-05) (hash(x)=49494590)
+28600 train 5.802358 (lr=3.1601e-05) (hash(x)=44172312)
+26900 val loss 5.8867
+26900 val perplexity 360.2006
+26900 train 6.010286 (lr=2.4951e-05) (hash(x)=51408788)
+28700 val loss 6.0143
+28700 val perplexity 409.2209
+28700 train 6.113423 (lr=3.1408e-05) (hash(x)=50435687)
+28000 val loss 6.2374
+28000 val perplexity 511.5454
+28000 train 5.639122 (lr=4.6813e-05) (hash(x)=42333356)
+27000 val loss 5.8821
+27000 val perplexity 358.5791
+27000 train 5.639222 (lr=2.4810e-05) (hash(x)=49024980)
+28800 val loss 6.0265
+28800 val perplexity 414.2587
+28800 train 5.955416 (lr=3.1214e-05) (hash(x)=53678477)
+28100 val loss 6.2366
+28100 val perplexity 511.1124
+28100 train 6.408118 (lr=4.6534e-05) (hash(x)=49378958)
+27100 val loss 5.8795
+27100 val perplexity 357.6401
+27100 train 6.145059 (lr=2.4669e-05) (hash(x)=45027677)
+28900 val loss 6.0131
+28900 val perplexity 408.7520
+28900 train 5.830149 (lr=3.1021e-05) (hash(x)=47353092)
+28200 val loss 6.2355
+28200 val perplexity 510.5514
+28200 train 6.413532 (lr=4.6256e-05) (hash(x)=56595308)
+27200 val loss 5.8851
+27200 val perplexity 359.6531
+27200 train 5.970792 (lr=2.4528e-05) (hash(x)=49930937)
+29000 val loss 6.0132
+29000 val perplexity 408.7924
+29000 train 6.236730 (lr=3.0828e-05) (hash(x)=56723899)
+28300 val loss 6.2460
+28300 val perplexity 515.9637
+28300 train 6.775725 (lr=4.5977e-05) (hash(x)=52311024)
+29100 val loss 6.0087
+29100 val perplexity 406.9632
+27300 val loss 5.8865
+27300 val perplexity 360.1448
+29100 train 6.015871 (lr=3.0635e-05) (hash(x)=56056600)
+27300 train 5.807554 (lr=2.4387e-05) (hash(x)=51016030)
+28400 val loss 6.2328
+28400 val perplexity 509.1544
+28400 train 6.245226 (lr=4.5700e-05) (hash(x)=47997766)
+29200 val loss 6.0170
+29200 val perplexity 410.3560
+29200 train 5.952146 (lr=3.0443e-05) (hash(x)=50538322)
+27400 val loss 5.8778
+27400 val perplexity 357.0374
+27400 train 6.118532 (lr=2.4247e-05) (hash(x)=48154914)
+28500 val loss 6.2164
+28500 val perplexity 500.9097
+28500 train 6.170032 (lr=4.5422e-05) (hash(x)=46906589)
+29300 val loss 6.0125
+29300 val perplexity 408.5009
+29300 train 6.051044 (lr=3.0251e-05) (hash(x)=57261305)
+27500 val loss 5.8801
+27500 val perplexity 357.8611
+27500 train 5.987864 (lr=2.4106e-05) (hash(x)=53265478)
+28600 val loss 6.2102
+28600 val perplexity 497.8000
+28600 train 6.002185 (lr=4.5145e-05) (hash(x)=44172312)
+29400 val loss 6.0143
+29400 val perplexity 409.2464
+29400 train 5.828960 (lr=3.0060e-05) (hash(x)=49014077)
+27600 val loss 5.9269
+27600 val perplexity 374.9927
+27600 train 5.704084 (lr=2.3966e-05) (hash(x)=39456920)
+28700 val loss 6.2190
+28700 val perplexity 502.2180
+28700 train 6.295090 (lr=4.4868e-05) (hash(x)=50435687)
+29500 val loss 6.0043
+29500 val perplexity 405.1839
+29500 train 5.747086 (lr=2.9868e-05) (hash(x)=48698820)
+27700 val loss 5.8864
+27700 val perplexity 360.1145
+27700 train 6.045403 (lr=2.3826e-05) (hash(x)=55794844)
+28800 val loss 6.2307
+28800 val perplexity 508.0989
+28800 train 6.177702 (lr=4.4592e-05) (hash(x)=53678477)
+29600 val loss 6.0036
+29600 val perplexity 404.9002
+29600 train 5.857091 (lr=2.9677e-05) (hash(x)=48327445)
+28900 val loss 6.2273
+28900 val perplexity 506.4051
+28900 train 6.015232 (lr=4.4316e-05) (hash(x)=47353092)
+27800 val loss 5.8906
+27800 val perplexity 361.6364
+27800 train 5.769773 (lr=2.3686e-05) (hash(x)=49091362)
+29700 val loss 6.0087
+29700 val perplexity 406.9378
+29700 train 6.164454 (lr=2.9487e-05) (hash(x)=52084503)
+29000 val loss 6.2123
+29000 val perplexity 498.8557
+29000 train 6.440894 (lr=4.4040e-05) (hash(x)=56723899)
+27900 val loss 5.8769
+27900 val perplexity 356.7052
+27900 train 5.697681 (lr=2.3546e-05) (hash(x)=49494590)
+29800 val loss 6.0307
+29800 val perplexity 416.0104
+29800 train 5.802908 (lr=2.9297e-05) (hash(x)=44402411)
+29100 val loss 6.2136
+29100 val perplexity 499.5199
+29100 train 6.206547 (lr=4.3765e-05) (hash(x)=56056600)
+28000 val loss 5.8756
+28000 val perplexity 356.2436
+28000 train 5.260273 (lr=2.3407e-05) (hash(x)=42333356)
+29900 val loss 6.0100
+29900 val perplexity 407.5032
+29900 train 5.680004 (lr=2.9107e-05) (hash(x)=43932550)
+29200 val loss 6.2180
+29200 val perplexity 501.7069
+29200 train 6.145995 (lr=4.3490e-05) (hash(x)=50538322)
+28100 val loss 5.8717
+28100 val perplexity 354.8508
+28100 train 6.080104 (lr=2.3267e-05) (hash(x)=49378958)
+30000 val loss 5.9983
+30000 val perplexity 402.7516
+30000 train 5.870724 (lr=2.8917e-05) (hash(x)=48250169)
+29300 val loss 6.2152
+29300 val perplexity 500.2998
+29300 train 6.333208 (lr=4.3216e-05) (hash(x)=57261305)
+28200 val loss 5.8669
+28200 val perplexity 353.1412
+28200 train 6.020795 (lr=2.3128e-05) (hash(x)=56595308)
+30100 val loss 5.9937
+30100 val perplexity 400.8860
+30100 train 6.050492 (lr=2.8728e-05) (hash(x)=46727253)
+29400 val loss 6.2179
+29400 val perplexity 501.6275
+29400 train 6.051409 (lr=4.2942e-05) (hash(x)=49014077)
+28300 val loss 5.8746
+28300 val perplexity 355.8778
+28300 train 6.543456 (lr=2.2989e-05) (hash(x)=52311024)
+30200 val loss 5.9869
+30200 val perplexity 398.1633
+30200 train 6.059178 (lr=2.8539e-05) (hash(x)=52740773)
+29500 val loss 6.2182
+29500 val perplexity 501.8105
+29500 train 5.973386 (lr=4.2669e-05) (hash(x)=48698820)
+28400 val loss 5.8685
+28400 val perplexity 353.7059
+28400 train 5.846119 (lr=2.2850e-05) (hash(x)=47997766)
+30300 val loss 5.9918
+30300 val perplexity 400.1261
+30300 train 6.092599 (lr=2.8351e-05) (hash(x)=52481310)
+29600 val loss 6.2193
+29600 val perplexity 502.3401
+29600 train 6.098068 (lr=4.2396e-05) (hash(x)=48327445)
+28500 val loss 5.8600
+28500 val perplexity 350.7257
+28500 train 5.825363 (lr=2.2711e-05) (hash(x)=46906589)
+30400 val loss 6.0039
+30400 val perplexity 405.0247
+30400 train 5.917053 (lr=2.8163e-05) (hash(x)=44942827)
+29700 val loss 6.2219
+29700 val perplexity 503.6554
+29700 train 6.411894 (lr=4.2124e-05) (hash(x)=52084503)
+28600 val loss 5.8538
+28600 val perplexity 348.5419
+28600 train 5.667266 (lr=2.2572e-05) (hash(x)=44172312)
+30500 val loss 5.9957
+30500 val perplexity 401.7155
+30500 train 5.870349 (lr=2.7976e-05) (hash(x)=53118706)
+29800 val loss 6.2254
+29800 val perplexity 505.4112
+29800 train 5.972321 (lr=4.1852e-05) (hash(x)=44402411)
+28700 val loss 5.8625
+28700 val perplexity 351.6016
+28700 train 5.945757 (lr=2.2434e-05) (hash(x)=50435687)
+30600 val loss 5.9892
+30600 val perplexity 399.0977
+30600 train 5.704517 (lr=2.7789e-05) (hash(x)=46278699)
+29900 val loss 6.2241
+29900 val perplexity 504.7782
+29900 train 5.888961 (lr=4.1581e-05) (hash(x)=43932550)
+28800 val loss 5.8593
+28800 val perplexity 350.4829
+28800 train 5.783760 (lr=2.2296e-05) (hash(x)=53678477)
+30700 val loss 5.9854
+30700 val perplexity 397.5801
+30700 train 5.937806 (lr=2.7602e-05) (hash(x)=51949086)
+30000 val loss 6.2064
+30000 val perplexity 495.9354
+30000 train 6.103238 (lr=4.1310e-05) (hash(x)=48250169)
+28900 val loss 5.8612
+28900 val perplexity 351.1393
+28900 train 5.676228 (lr=2.2158e-05) (hash(x)=47353092)
+30800 val loss 5.9892
+30800 val perplexity 399.1051
+30800 train 5.973952 (lr=2.7416e-05) (hash(x)=50575709)
+30100 val loss 6.2129
+30100 val perplexity 499.1432
+30100 train 6.235704 (lr=4.1040e-05) (hash(x)=46727253)
+29000 val loss 5.8532
+29000 val perplexity 348.3626
+29000 train 6.064645 (lr=2.2020e-05) (hash(x)=56723899)
+30900 val loss 5.9934
+30900 val perplexity 400.7560
+30900 train 5.830714 (lr=2.7230e-05) (hash(x)=53992116)
+30200 val loss 6.2046
+30200 val perplexity 495.0435
+30200 train 6.227543 (lr=4.0771e-05) (hash(x)=52740773)
+29100 val loss 5.8571
+29100 val perplexity 349.7240
+29100 train 5.874152 (lr=2.1882e-05) (hash(x)=56056600)
+31000 val loss 5.9884
+31000 val perplexity 398.7793
+31000 train 5.784818 (lr=2.7045e-05) (hash(x)=44967737)
+30300 val loss 6.1977
+30300 val perplexity 491.6121
+30300 train 6.350532 (lr=4.0502e-05) (hash(x)=52481310)
+29200 val loss 5.8520
+29200 val perplexity 347.9445
+29200 train 5.807312 (lr=2.1745e-05) (hash(x)=50538322)
+31100 val loss 5.9902
+31100 val perplexity 399.4781
+31100 train 5.687879 (lr=2.6860e-05) (hash(x)=41850245)
+30400 val loss 6.2095
+30400 val perplexity 497.4491
+30400 train 6.109884 (lr=4.0233e-05) (hash(x)=44942827)
+29300 val loss 5.8500
+29300 val perplexity 347.2274
+29300 train 5.855346 (lr=2.1608e-05) (hash(x)=57261305)
+31200 val loss 5.9812
+31200 val perplexity 395.9185
+31200 train 5.836033 (lr=2.6675e-05) (hash(x)=49921656)
+30500 val loss 6.1996
+30500 val perplexity 492.5657
+30500 train 6.063776 (lr=3.9965e-05) (hash(x)=53118706)
+29400 val loss 5.8523
+29400 val perplexity 348.0262
+29400 train 5.677557 (lr=2.1471e-05) (hash(x)=49014077)
+31300 val loss 5.9833
+31300 val perplexity 396.7625
+31300 train 5.862272 (lr=2.6491e-05) (hash(x)=47068006)
+30600 val loss 6.1976
+30600 val perplexity 491.5683
+30600 train 5.907681 (lr=3.9698e-05) (hash(x)=46278699)
+29500 val loss 5.8536
+29500 val perplexity 348.5013
+29500 train 5.582386 (lr=2.1335e-05) (hash(x)=48698820)
+31400 val loss 5.9852
+31400 val perplexity 397.5016
+31400 train 6.159078 (lr=2.6308e-05) (hash(x)=52608679)
+30700 val loss 6.1976
+30700 val perplexity 491.5591
+30700 train 6.130133 (lr=3.9431e-05) (hash(x)=51949086)
+29600 val loss 5.8484
+29600 val perplexity 346.6938
+29600 train 5.701379 (lr=2.1198e-05) (hash(x)=48327445)
+31500 val loss 5.9958
+31500 val perplexity 401.7412
+31500 train 5.899184 (lr=2.6125e-05) (hash(x)=51542050)
+30800 val loss 6.1912
+30800 val perplexity 488.4259
+30800 train 6.164798 (lr=3.9165e-05) (hash(x)=50575709)
+31600 val loss 5.9843
+31600 val perplexity 397.1341
+31600 train 5.896556 (lr=2.5942e-05) (hash(x)=49170260)
+29700 val loss 5.8503
+29700 val perplexity 347.3433
+29700 train 6.005876 (lr=2.1062e-05) (hash(x)=52084503)
+30900 val loss 6.1910
+30900 val perplexity 488.3528
+30900 train 6.037580 (lr=3.8900e-05) (hash(x)=53992116)
+31700 val loss 5.9763
+31700 val perplexity 393.9861
+31700 train 6.023350 (lr=2.5760e-05) (hash(x)=57677051)
+29800 val loss 5.8533
+29800 val perplexity 348.3767
+29800 train 5.676028 (lr=2.0926e-05) (hash(x)=44402411)
+31000 val loss 6.1961
+31000 val perplexity 490.8546
+31000 train 6.038806 (lr=3.8635e-05) (hash(x)=44967737)
+31800 val loss 5.9717
+31800 val perplexity 392.1629
+31800 train 6.152301 (lr=2.5579e-05) (hash(x)=52567943)
+29900 val loss 5.8530
+29900 val perplexity 348.2892
+29900 train 5.540175 (lr=2.0791e-05) (hash(x)=43932550)
+31100 val loss 6.1820
+31100 val perplexity 483.9361
+31100 train 5.881097 (lr=3.8371e-05) (hash(x)=41850245)
+31900 val loss 5.9706
+31900 val perplexity 391.7267
+31900 train 6.161015 (lr=2.5398e-05) (hash(x)=51246750)
+30000 val loss 5.8419
+30000 val perplexity 344.4402
+30000 train 5.702406 (lr=2.0655e-05) (hash(x)=48250169)
+31200 val loss 6.1892
+31200 val perplexity 487.4655
+31200 train 6.056690 (lr=3.8108e-05) (hash(x)=49921656)
+32000 val loss 5.9709
+32000 val perplexity 391.8543
+32000 train 6.447259 (lr=2.5218e-05) (hash(x)=63256210)
+30100 val loss 5.8364
+30100 val perplexity 342.5384
+30100 train 5.934769 (lr=2.0520e-05) (hash(x)=46727253)
+31300 val loss 6.1902
+31300 val perplexity 487.9301
+31300 train 6.051952 (lr=3.7845e-05) (hash(x)=47068006)
+32100 val loss 5.9726
+32100 val perplexity 392.5269
+32100 train 6.162525 (lr=2.5038e-05) (hash(x)=52238178)
+30200 val loss 5.8326
+30200 val perplexity 341.2585
+30200 train 5.895270 (lr=2.0385e-05) (hash(x)=52740773)
+31400 val loss 6.1884
+31400 val perplexity 487.0724
+31400 train 6.287651 (lr=3.7583e-05) (hash(x)=52608679)
+32200 val loss 5.9707
+32200 val perplexity 391.7810
+32200 train 6.205959 (lr=2.4858e-05) (hash(x)=56290278)
+30300 val loss 5.8379
+30300 val perplexity 343.0556
+30300 train 5.890831 (lr=2.0251e-05) (hash(x)=52481310)
+31500 val loss 6.2008
+31500 val perplexity 493.1667
+31500 train 6.076781 (lr=3.7321e-05) (hash(x)=51542050)
+32300 val loss 5.9640
+32300 val perplexity 389.1691
+32300 train 6.014713 (lr=2.4679e-05) (hash(x)=48455511)
+30400 val loss 5.8431
+30400 val perplexity 344.8558
+30400 train 5.780347 (lr=2.0117e-05) (hash(x)=44942827)
+31600 val loss 6.1891
+31600 val perplexity 487.4055
+31600 train 6.056008 (lr=3.7061e-05) (hash(x)=49170260)
+32400 val loss 5.9667
+32400 val perplexity 390.2041
+32400 train 5.776800 (lr=2.4501e-05) (hash(x)=42754343)
+30500 val loss 5.8344
+30500 val perplexity 341.8611
+30500 train 5.713102 (lr=1.9983e-05) (hash(x)=53118706)
+31700 val loss 6.1859
+31700 val perplexity 485.8332
+31700 train 6.241374 (lr=3.6801e-05) (hash(x)=57677051)
+32500 val loss 5.9725
+32500 val perplexity 392.4980
+32500 train 5.962894 (lr=2.4323e-05) (hash(x)=51062421)
+30600 val loss 5.8376
+30600 val perplexity 342.9691
+30600 train 5.548477 (lr=1.9849e-05) (hash(x)=46278699)
+31800 val loss 6.1742
+31800 val perplexity 480.2129
+31800 train 6.333062 (lr=3.6541e-05) (hash(x)=52567943)
+32600 val loss 5.9705
+32600 val perplexity 391.7160
+32600 train 5.763828 (lr=2.4146e-05) (hash(x)=43957725)
+30700 val loss 5.8322
+30700 val perplexity 341.1252
+30700 train 5.787788 (lr=1.9716e-05) (hash(x)=51949086)
+31900 val loss 6.1778
+31900 val perplexity 481.9407
+31900 train 6.360960 (lr=3.6283e-05) (hash(x)=51246750)
+32700 val loss 5.9618
+32700 val perplexity 388.3044
+32700 train 5.987890 (lr=2.3970e-05) (hash(x)=55971557)
+30800 val loss 5.8322
+30800 val perplexity 341.1127
+30800 train 5.843060 (lr=1.9583e-05) (hash(x)=50575709)
+32000 val loss 6.1807
+32000 val perplexity 483.3308
+32000 train 6.709186 (lr=3.6025e-05) (hash(x)=63256210)
+32800 val loss 5.9656
+32800 val perplexity 389.7902
+32800 train 5.970448 (lr=2.3794e-05) (hash(x)=49429775)
+30900 val loss 5.8312
+30900 val perplexity 340.7760
+30900 train 5.671123 (lr=1.9450e-05) (hash(x)=53992116)
+32100 val loss 6.1815
+32100 val perplexity 483.7410
+32100 train 6.340555 (lr=3.5768e-05) (hash(x)=52238178)
+32900 val loss 5.9773
+32900 val perplexity 394.3556
+32900 train 5.824573 (lr=2.3618e-05) (hash(x)=49564803)
+32200 val loss 6.1679
+32200 val perplexity 477.1887
+32200 train 6.374283 (lr=3.5512e-05) (hash(x)=56290278)
+31000 val loss 5.8332
+31000 val perplexity 341.4471
+31000 train 5.659619 (lr=1.9318e-05) (hash(x)=44967737)
+33000 val loss 5.9679
+33000 val perplexity 390.7012
+33000 train 5.847964 (lr=2.3443e-05) (hash(x)=46343736)
+32300 val loss 6.1604
+32300 val perplexity 473.5974
+32300 train 6.191844 (lr=3.5256e-05) (hash(x)=48455511)
+31100 val loss 5.8327
+31100 val perplexity 341.2876
+31100 train 5.544391 (lr=1.9186e-05) (hash(x)=41850245)
+33100 val loss 5.9698
+33100 val perplexity 391.4238
+33100 train 5.662004 (lr=2.3269e-05) (hash(x)=45424347)
+32400 val loss 6.1660
+32400 val perplexity 476.2751
+32400 train 5.952959 (lr=3.5002e-05) (hash(x)=42754343)
+31200 val loss 5.8291
+31200 val perplexity 340.0558
+31200 train 5.683711 (lr=1.9054e-05) (hash(x)=49921656)
+33200 val loss 5.9647
+33200 val perplexity 389.4277
+33200 train 6.005509 (lr=2.3095e-05) (hash(x)=50207861)
+32500 val loss 6.1714
+32500 val perplexity 478.8346
+32500 train 6.157789 (lr=3.4748e-05) (hash(x)=51062421)
+31300 val loss 5.8296
+31300 val perplexity 340.2083
+31300 train 5.687627 (lr=1.8922e-05) (hash(x)=47068006)
+33300 val loss 5.9633
+33300 val perplexity 388.8740
+33300 train 5.886430 (lr=2.2922e-05) (hash(x)=45847922)
+32600 val loss 6.1692
+32600 val perplexity 477.8271
+32600 train 5.951297 (lr=3.4495e-05) (hash(x)=43957725)
+31400 val loss 5.8276
+31400 val perplexity 339.5481
+31400 train 5.998035 (lr=1.8791e-05) (hash(x)=52608679)
+33400 val loss 5.9622
+33400 val perplexity 388.4505
+33400 train 5.791804 (lr=2.2750e-05) (hash(x)=44723617)
+32700 val loss 6.1683
+32700 val perplexity 477.3633
+32700 train 6.196739 (lr=3.4242e-05) (hash(x)=55971557)
+33500 val loss 5.9567
+33500 val perplexity 386.3194
+33500 train 6.093992 (lr=2.2578e-05) (hash(x)=54058855)
+31500 val loss 5.8449
+31500 val perplexity 345.4549
+31500 train 5.749641 (lr=1.8661e-05) (hash(x)=51542050)
+32800 val loss 6.1635
+32800 val perplexity 475.0682
+32800 train 6.165715 (lr=3.3991e-05) (hash(x)=49429775)
+33600 val loss 5.9519
+33600 val perplexity 384.4791
+33600 train 5.929941 (lr=2.2407e-05) (hash(x)=53052866)
+31600 val loss 5.8288
+31600 val perplexity 339.9622
+31600 train 5.715558 (lr=1.8530e-05) (hash(x)=49170260)
+32900 val loss 6.1666
+32900 val perplexity 476.5743
+32900 train 6.011598 (lr=3.3740e-05) (hash(x)=49564803)
+33700 val loss 5.9595
+33700 val perplexity 387.4097
+33700 train 5.876254 (lr=2.2237e-05) (hash(x)=50458295)
+31700 val loss 5.8296
+31700 val perplexity 340.2090
+31700 train 5.814413 (lr=1.8400e-05) (hash(x)=57677051)
+33000 val loss 6.1660
+33000 val perplexity 476.2631
+33000 train 6.041053 (lr=3.3490e-05) (hash(x)=46343736)
+33800 val loss 5.9636
+33800 val perplexity 389.0148
+33800 train 6.121011 (lr=2.2067e-05) (hash(x)=58204533)
+31800 val loss 5.8223
+31800 val perplexity 337.7436
+31800 train 5.994304 (lr=1.8271e-05) (hash(x)=52567943)
+33900 val loss 5.9465
+33900 val perplexity 382.4198
+33900 train 6.083571 (lr=2.1898e-05) (hash(x)=52246445)
+33100 val loss 6.1740
+33100 val perplexity 480.1154
+33100 train 5.879986 (lr=3.3242e-05) (hash(x)=45424347)
+31900 val loss 5.8270
+31900 val perplexity 339.3340
+31900 train 6.048852 (lr=1.8141e-05) (hash(x)=51246750)
+34000 val loss 5.9464
+34000 val perplexity 382.3709
+34000 train 5.890872 (lr=2.1729e-05) (hash(x)=49316883)
+33200 val loss 6.1753
+33200 val perplexity 480.7034
+33200 train 6.173397 (lr=3.2993e-05) (hash(x)=50207861)
+32000 val loss 5.8267
+32000 val perplexity 339.2486
+32000 train 6.286895 (lr=1.8013e-05) (hash(x)=63256210)
+34100 val loss 5.9498
+34100 val perplexity 383.6810
+34100 train 5.636144 (lr=2.1561e-05) (hash(x)=42991129)
+33300 val loss 6.1664
+33300 val perplexity 476.4754
+33300 train 6.078833 (lr=3.2746e-05) (hash(x)=45847922)
+32100 val loss 5.8271
+32100 val perplexity 339.3733
+32100 train 6.022761 (lr=1.7884e-05) (hash(x)=52238178)
+34200 val loss 5.9550
+34200 val perplexity 385.6611
+34200 train 5.614861 (lr=2.1394e-05) (hash(x)=45199853)
+33400 val loss 6.1709
+33400 val perplexity 478.6107
+33400 train 5.986867 (lr=3.2500e-05) (hash(x)=44723617)
+32200 val loss 5.8140
+32200 val perplexity 334.9713
+32200 train 6.064962 (lr=1.7756e-05) (hash(x)=56290278)
+34300 val loss 5.9525
+34300 val perplexity 384.7282
+34300 train 5.832723 (lr=2.1227e-05) (hash(x)=48404386)
+33500 val loss 6.1614
+33500 val perplexity 474.0743
+33500 train 6.271558 (lr=3.2255e-05) (hash(x)=54058855)
+32300 val loss 5.8125
+32300 val perplexity 334.4654
+32300 train 5.886790 (lr=1.7628e-05) (hash(x)=48455511)
+34400 val loss 5.9482
+34400 val perplexity 383.0553
+34400 train 5.740652 (lr=2.1062e-05) (hash(x)=45614997)
+33600 val loss 6.1572
+33600 val perplexity 472.1153
+33600 train 6.153094 (lr=3.2010e-05) (hash(x)=53052866)
+32400 val loss 5.8132
+32400 val perplexity 334.7010
+32400 train 5.625381 (lr=1.7501e-05) (hash(x)=42754343)
+33700 val loss 6.1547
+33700 val perplexity 470.9484
+33700 train 6.073953 (lr=3.1767e-05) (hash(x)=50458295)
+34500 val loss 5.9496
+34500 val perplexity 383.6016
+34500 train 5.980463 (lr=2.0896e-05) (hash(x)=53157684)
+32500 val loss 5.8101
+32500 val perplexity 333.6473
+32500 train 5.810159 (lr=1.7374e-05) (hash(x)=51062421)
+34600 val loss 5.9588
+34600 val perplexity 387.1570
+34600 train 6.175712 (lr=2.0732e-05) (hash(x)=57152287)
+33800 val loss 6.1629
+33800 val perplexity 474.8112
+33800 train 6.277860 (lr=3.1524e-05) (hash(x)=58204533)
+32600 val loss 5.8166
+32600 val perplexity 335.8415
+32600 train 5.610869 (lr=1.7247e-05) (hash(x)=43957725)
+34700 val loss 5.9519
+34700 val perplexity 384.4753
+34700 train 5.406115 (lr=2.0568e-05) (hash(x)=46088213)
+33900 val loss 6.1466
+33900 val perplexity 467.1330
+33900 train 6.256472 (lr=3.1282e-05) (hash(x)=52246445)
+32700 val loss 5.8128
+32700 val perplexity 334.5435
+32700 train 5.838300 (lr=1.7121e-05) (hash(x)=55971557)
+34800 val loss 5.9471
+34800 val perplexity 382.6555
+34800 train 5.693446 (lr=2.0405e-05) (hash(x)=47098930)
+34000 val loss 6.1448
+34000 val perplexity 466.2920
+34000 train 6.076751 (lr=3.1041e-05) (hash(x)=49316883)
+34900 val loss 5.9476
+34900 val perplexity 382.8204
+34900 train 5.945630 (lr=2.0243e-05) (hash(x)=51033119)
+32800 val loss 5.8090
+32800 val perplexity 333.2708
+32800 train 5.821727 (lr=1.6995e-05) (hash(x)=49429775)
+34100 val loss 6.1457
+34100 val perplexity 466.7285
+34100 train 5.839543 (lr=3.0802e-05) (hash(x)=42991129)
+35000 val loss 5.9482
+35000 val perplexity 383.0675
+35000 train 5.622958 (lr=2.0081e-05) (hash(x)=47111881)
+32900 val loss 5.8096
+32900 val perplexity 333.4849
+32900 train 5.661140 (lr=1.6870e-05) (hash(x)=49564803)
+34200 val loss 6.1458
+34200 val perplexity 466.7634
+34200 train 5.827536 (lr=3.0563e-05) (hash(x)=45199853)
+35100 val loss 5.9462
+35100 val perplexity 382.2891
+35100 train 5.967577 (lr=1.9921e-05) (hash(x)=52899306)
+34300 val loss 6.1517
+34300 val perplexity 469.5308
+34300 train 6.048966 (lr=3.0325e-05) (hash(x)=48404386)
+33000 val loss 5.8099
+33000 val perplexity 333.5746
+33000 train 5.681557 (lr=1.6745e-05) (hash(x)=46343736)
+35200 val loss 5.9419
+35200 val perplexity 380.6734
+35200 train 5.869995 (lr=1.9760e-05) (hash(x)=49124834)
+34400 val loss 6.1486
+34400 val perplexity 468.0514
+34400 train 5.962949 (lr=3.0088e-05) (hash(x)=45614997)
+33100 val loss 5.8131
+33100 val perplexity 334.6570
+33100 train 5.508917 (lr=1.6621e-05) (hash(x)=45424347)
+35300 val loss 5.9407
+35300 val perplexity 380.2195
+35300 train 6.108963 (lr=1.9601e-05) (hash(x)=54127282)
+34500 val loss 6.1566
+34500 val perplexity 471.8180
+34500 train 6.178240 (lr=2.9852e-05) (hash(x)=53157684)
+33200 val loss 5.8256
+33200 val perplexity 338.8622
+33200 train 5.892923 (lr=1.6497e-05) (hash(x)=50207861)
+35400 val loss 5.9397
+35400 val perplexity 379.8055
+35400 train 5.801503 (lr=1.9442e-05) (hash(x)=51154858)
+34600 val loss 6.1572
+34600 val perplexity 472.1272
+34600 train 6.362251 (lr=2.9617e-05) (hash(x)=57152287)
+33300 val loss 5.8135
+33300 val perplexity 334.7842
+33300 train 5.733989 (lr=1.6373e-05) (hash(x)=45847922)
+35500 val loss 5.9351
+35500 val perplexity 378.0600
+35500 train 5.985872 (lr=1.9285e-05) (hash(x)=51585534)
+34700 val loss 6.1456
+34700 val perplexity 466.6777
+34700 train 5.622260 (lr=2.9383e-05) (hash(x)=46088213)
+33400 val loss 5.8126
+33400 val perplexity 334.4757
+33400 train 5.656195 (lr=1.6250e-05) (hash(x)=44723617)
+35600 val loss 5.9394
+35600 val perplexity 379.6975
+35600 train 5.750732 (lr=1.9128e-05) (hash(x)=46228695)
+34800 val loss 6.1489
+34800 val perplexity 468.2012
+34800 train 5.881421 (lr=2.9150e-05) (hash(x)=47098930)
+33500 val loss 5.8089
+33500 val perplexity 333.2426
+33500 train 5.919910 (lr=1.6127e-05) (hash(x)=54058855)
+35700 val loss 5.9294
+35700 val perplexity 375.9344
+35700 train 5.940912 (lr=1.8971e-05) (hash(x)=49080481)
+34900 val loss 6.1519
+34900 val perplexity 469.5917
+34900 train 6.178014 (lr=2.8918e-05) (hash(x)=51033119)
+33600 val loss 5.8011
+33600 val perplexity 330.6728
+33600 train 5.768274 (lr=1.6005e-05) (hash(x)=53052866)
+35800 val loss 5.9297
+35800 val perplexity 376.0268
+35800 train 6.078786 (lr=1.8816e-05) (hash(x)=56743084)
+35000 val loss 6.1568
+35000 val perplexity 471.9262
+35000 train 5.839863 (lr=2.8688e-05) (hash(x)=47111881)
+33700 val loss 5.8030
+33700 val perplexity 331.2847
+33700 train 5.727721 (lr=1.5883e-05) (hash(x)=50458295)
+35900 val loss 5.9304
+35900 val perplexity 376.3233
+35900 train 6.033379 (lr=1.8661e-05) (hash(x)=53068229)
+35100 val loss 6.1500
+35100 val perplexity 468.7335
+35100 train 6.190754 (lr=2.8458e-05) (hash(x)=52899306)
+33800 val loss 5.8021
+33800 val perplexity 331.0087
+33800 train 5.948277 (lr=1.5762e-05) (hash(x)=58204533)
+36000 val loss 5.9267
+36000 val perplexity 374.9042
+36000 train 6.096465 (lr=1.8507e-05) (hash(x)=54363049)
+35200 val loss 6.1501
+35200 val perplexity 468.7583
+35200 train 6.098011 (lr=2.8229e-05) (hash(x)=49124834)
+36100 val loss 5.9290
+36100 val perplexity 375.7826
+36100 train 5.854910 (lr=1.8354e-05) (hash(x)=50555767)
+33900 val loss 5.7948
+33900 val perplexity 328.6017
+33900 train 5.935129 (lr=1.5641e-05) (hash(x)=52246445)
+35300 val loss 6.1463
+35300 val perplexity 466.9794
+35300 train 6.313648 (lr=2.8002e-05) (hash(x)=54127282)
+36200 val loss 5.9271
+36200 val perplexity 375.0605
+36200 train 6.072759 (lr=1.8201e-05) (hash(x)=52103710)
+34000 val loss 5.7918
+34000 val perplexity 327.5931
+34000 train 5.721344 (lr=1.5521e-05) (hash(x)=49316883)
+35400 val loss 6.1389
+35400 val perplexity 463.5280
+35400 train 5.998002 (lr=2.7775e-05) (hash(x)=51154858)
+36300 val loss 5.9308
+36300 val perplexity 376.4374
+36300 train 5.712657 (lr=1.8050e-05) (hash(x)=43347374)
+34100 val loss 5.7962
+34100 val perplexity 329.0480
+34100 train 5.451437 (lr=1.5401e-05) (hash(x)=42991129)
+35500 val loss 6.1332
+35500 val perplexity 460.9086
+35500 train 6.169389 (lr=2.7549e-05) (hash(x)=51585534)
+36400 val loss 5.9217
+36400 val perplexity 373.0446
+36400 train 5.978114 (lr=1.7899e-05) (hash(x)=52661268)
+34200 val loss 5.7939
+34200 val perplexity 328.2848
+34200 train 5.440014 (lr=1.5281e-05) (hash(x)=45199853)
+35600 val loss 6.1330
+35600 val perplexity 460.8136
+35600 train 5.942529 (lr=2.7325e-05) (hash(x)=46228695)
+36500 val loss 5.9290
+36500 val perplexity 375.7711
+36500 train 5.829330 (lr=1.7749e-05) (hash(x)=47765383)
+34300 val loss 5.7996
+34300 val perplexity 330.1776
+34300 train 5.690372 (lr=1.5162e-05) (hash(x)=48404386)
+35700 val loss 6.1292
+35700 val perplexity 459.0685
+35700 train 6.098458 (lr=2.7102e-05) (hash(x)=49080481)
+36600 val loss 5.9310
+36600 val perplexity 376.5367
+36600 train 5.769433 (lr=1.7600e-05) (hash(x)=49302411)
+34400 val loss 5.7971
+34400 val perplexity 329.3420
+34400 train 5.594779 (lr=1.5044e-05) (hash(x)=45614997)
+35800 val loss 6.1308
+35800 val perplexity 459.8204
+35800 train 6.276246 (lr=2.6880e-05) (hash(x)=56743084)
+36700 val loss 5.9329
+36700 val perplexity 377.2295
+36700 train 5.456398 (lr=1.7452e-05) (hash(x)=44113066)
+34500 val loss 5.8002
+34500 val perplexity 330.3516
+34500 train 5.815750 (lr=1.4926e-05) (hash(x)=53157684)
+35900 val loss 6.1383
+35900 val perplexity 463.2589
+35900 train 6.221340 (lr=2.6659e-05) (hash(x)=53068229)
+36800 val loss 5.9301
+36800 val perplexity 376.1776
+36800 train 5.837457 (lr=1.7305e-05) (hash(x)=55896927)
+34600 val loss 5.8000
+34600 val perplexity 330.2973
+34600 train 6.033842 (lr=1.4809e-05) (hash(x)=57152287)
+36000 val loss 6.1363
+36000 val perplexity 462.3354
+36000 train 6.289048 (lr=2.6439e-05) (hash(x)=54363049)
+36900 val loss 5.9274
+36900 val perplexity 375.1843
+36900 train 5.742035 (lr=1.7158e-05) (hash(x)=49812554)
+34700 val loss 5.7947
+34700 val perplexity 328.5693
+34700 train 5.238037 (lr=1.4692e-05) (hash(x)=46088213)
+36100 val loss 6.1482
+36100 val perplexity 467.8798
+36100 train 6.052368 (lr=2.6220e-05) (hash(x)=50555767)
+37000 val loss 5.9316
+37000 val perplexity 376.7454
+37000 train 5.750156 (lr=1.7012e-05) (hash(x)=49391771)
+34800 val loss 5.7920
+34800 val perplexity 327.6733
+34800 train 5.528825 (lr=1.4575e-05) (hash(x)=47098930)
+36200 val loss 6.1339
+36200 val perplexity 461.2412
+36200 train 6.267734 (lr=2.6002e-05) (hash(x)=52103710)
+37100 val loss 5.9311
+37100 val perplexity 376.5597
+37100 train 5.879148 (lr=1.6867e-05) (hash(x)=50515328)
+34900 val loss 5.7981
+34900 val perplexity 329.6606
+34900 train 5.806500 (lr=1.4459e-05) (hash(x)=51033119)
+36300 val loss 6.1314
+36300 val perplexity 460.0860
+36300 train 5.908183 (lr=2.5786e-05) (hash(x)=43347374)
+37200 val loss 5.9214
+37200 val perplexity 372.9274
+37200 train 6.117315 (lr=1.6723e-05) (hash(x)=57788407)
+35000 val loss 5.7945
+35000 val perplexity 328.4980
+35000 train 5.473814 (lr=1.4344e-05) (hash(x)=47111881)
+36400 val loss 6.1271
+36400 val perplexity 458.0939
+36400 train 6.201708 (lr=2.5570e-05) (hash(x)=52661268)
+37300 val loss 5.9186
+37300 val perplexity 371.8936
+37300 train 5.694643 (lr=1.6580e-05) (hash(x)=46333526)
+35100 val loss 5.7934
+35100 val perplexity 328.1255
+35100 train 5.811529 (lr=1.4229e-05) (hash(x)=52899306)
+36500 val loss 6.1314
+36500 val perplexity 460.0856
+36500 train 6.002098 (lr=2.5356e-05) (hash(x)=47765383)
+37400 val loss 5.9194
+37400 val perplexity 372.1934
+37400 train 5.719808 (lr=1.6438e-05) (hash(x)=45862038)
+35200 val loss 5.7948
+35200 val perplexity 328.5847
+35200 train 5.748470 (lr=1.4115e-05) (hash(x)=49124834)
+36600 val loss 6.1339
+36600 val perplexity 461.2357
+36600 train 5.971117 (lr=2.5143e-05) (hash(x)=49302411)
+37500 val loss 5.9198
+37500 val perplexity 372.3443
+37500 train 5.635836 (lr=1.6297e-05) (hash(x)=46543051)
+35300 val loss 5.7896
+35300 val perplexity 326.8839
+35300 train 5.945773 (lr=1.4001e-05) (hash(x)=54127282)
+36700 val loss 6.1357
+36700 val perplexity 462.0757
+36700 train 5.676261 (lr=2.4931e-05) (hash(x)=44113066)
+37600 val loss 5.9128
+37600 val perplexity 369.7480
+37600 train 5.879233 (lr=1.6156e-05) (hash(x)=51314450)
+35400 val loss 5.8002
+35400 val perplexity 330.3776
+35400 train 5.669840 (lr=1.3887e-05) (hash(x)=51154858)
+36800 val loss 6.1364
+36800 val perplexity 462.4066
+36800 train 6.062473 (lr=2.4721e-05) (hash(x)=55896927)
+37700 val loss 5.9144
+37700 val perplexity 370.3425
+37700 train 5.778914 (lr=1.6016e-05) (hash(x)=46706763)
+35500 val loss 5.7805
+35500 val perplexity 323.9211
+35500 train 5.829420 (lr=1.3775e-05) (hash(x)=51585534)
+36900 val loss 6.1396
+36900 val perplexity 463.8754
+36900 train 5.966776 (lr=2.4511e-05) (hash(x)=49812554)
+37800 val loss 5.9107
+37800 val perplexity 368.9532
+37800 train 5.765634 (lr=1.5878e-05) (hash(x)=45793085)
+35600 val loss 5.7790
+35600 val perplexity 323.4417
+35600 train 5.583309 (lr=1.3663e-05) (hash(x)=46228695)
+37000 val loss 6.1361
+37000 val perplexity 462.2492
+37000 train 5.911458 (lr=2.4303e-05) (hash(x)=49391771)
+37900 val loss 5.9105
+37900 val perplexity 368.8726
+37900 train 5.771966 (lr=1.5740e-05) (hash(x)=49969493)
+35700 val loss 5.7855
+35700 val perplexity 325.5457
+37100 val loss 6.1279
+37100 val perplexity 458.4543
+35700 train 5.764114 (lr=1.3551e-05) (hash(x)=49080481)
+37100 train 6.064731 (lr=2.4096e-05) (hash(x)=50515328)
+38000 val loss 5.9120
+38000 val perplexity 369.4412
+38000 train 5.700391 (lr=1.5603e-05) (hash(x)=47184489)
+37200 val loss 6.1302
+37200 val perplexity 459.5096
+37200 train 6.320570 (lr=2.3890e-05) (hash(x)=57788407)
+35800 val loss 5.7784
+35800 val perplexity 323.2407
+35800 train 5.946089 (lr=1.3440e-05) (hash(x)=56743084)
+38100 val loss 5.9118
+38100 val perplexity 369.3829
+38100 train 6.399194 (lr=1.5467e-05) (hash(x)=54905742)
+37300 val loss 6.1166
+37300 val perplexity 453.3305
+37300 train 5.894850 (lr=2.3686e-05) (hash(x)=46333526)
+35900 val loss 5.7812
+35900 val perplexity 324.1445
+35900 train 5.871631 (lr=1.3329e-05) (hash(x)=53068229)
+38200 val loss 5.9090
+38200 val perplexity 368.3536
+38200 train 6.178787 (lr=1.5332e-05) (hash(x)=48831719)
+37400 val loss 6.1123
+37400 val perplexity 451.3916
+37400 train 5.903808 (lr=2.3483e-05) (hash(x)=45862038)
+36000 val loss 5.7791
+36000 val perplexity 323.4611
+36000 train 5.956838 (lr=1.3219e-05) (hash(x)=54363049)
+38300 val loss 5.9125
+38300 val perplexity 369.6474
+38300 train 5.759835 (lr=1.5198e-05) (hash(x)=48944700)
+37500 val loss 6.1163
+37500 val perplexity 453.2015
+37500 train 5.844558 (lr=2.3281e-05) (hash(x)=46543051)
+36100 val loss 5.7844
+36100 val perplexity 325.1889
+36100 train 5.697713 (lr=1.3110e-05) (hash(x)=50555767)
+38400 val loss 5.9094
+38400 val perplexity 368.4789
+38400 train 5.544498 (lr=1.5064e-05) (hash(x)=46504848)
+37600 val loss 6.1285
+37600 val perplexity 458.7383
+37600 train 6.088625 (lr=2.3080e-05) (hash(x)=51314450)
+36200 val loss 5.7789
+36200 val perplexity 323.4190
+36200 train 5.935506 (lr=1.3001e-05) (hash(x)=52103710)
+38500 val loss 5.9102
+38500 val perplexity 368.7902
+38500 train 5.647353 (lr=1.4932e-05) (hash(x)=50148948)
+37700 val loss 6.1204
+37700 val perplexity 455.0315
+37700 train 5.980180 (lr=2.2881e-05) (hash(x)=46706763)
+36300 val loss 5.7805
+36300 val perplexity 323.9157
+36300 train 5.575034 (lr=1.2893e-05) (hash(x)=43347374)
+38600 val loss 5.9123
+38600 val perplexity 369.5478
+38600 train 6.689418 (lr=1.4801e-05) (hash(x)=60660913)
+37800 val loss 6.1189
+37800 val perplexity 454.3582
+37800 train 5.963765 (lr=2.2682e-05) (hash(x)=45793085)
+36400 val loss 5.7763
+36400 val perplexity 322.5495
+38700 val loss 5.9106
+38700 val perplexity 368.9344
+36400 train 5.812431 (lr=1.2785e-05) (hash(x)=52661268)
+38700 train 5.705847 (lr=1.4670e-05) (hash(x)=50256252)
+37900 val loss 6.1147
+37900 val perplexity 452.4417
+37900 train 5.977998 (lr=2.2485e-05) (hash(x)=49969493)
+38800 val loss 5.9057
+38800 val perplexity 367.1224
+38800 train 5.963464 (lr=1.4541e-05) (hash(x)=52592415)
+36500 val loss 5.7815
+36500 val perplexity 324.2580
+36500 train 5.695755 (lr=1.2678e-05) (hash(x)=47765383)
+38000 val loss 6.1155
+38000 val perplexity 452.8185
+38000 train 5.888125 (lr=2.2290e-05) (hash(x)=47184489)
+38900 val loss 5.9059
+38900 val perplexity 367.2014
+38900 train 5.842779 (lr=1.4412e-05) (hash(x)=51945430)
+36600 val loss 5.7770
+36600 val perplexity 322.7749
+36600 train 5.621260 (lr=1.2572e-05) (hash(x)=49302411)
+38100 val loss 6.1191
+38100 val perplexity 454.4779
+38100 train 6.524765 (lr=2.2096e-05) (hash(x)=54905742)
+39000 val loss 5.9003
+39000 val perplexity 365.1581
+39000 train 6.111910 (lr=1.4285e-05) (hash(x)=51838869)
+36700 val loss 5.7835
+36700 val perplexity 324.8792
+36700 train 5.340133 (lr=1.2466e-05) (hash(x)=44113066)
+38200 val loss 6.1137
+38200 val perplexity 451.9871
+38200 train 6.343554 (lr=2.1903e-05) (hash(x)=48831719)
+39100 val loss 5.8984
+39100 val perplexity 364.4516
+39100 train 6.018938 (lr=1.4158e-05) (hash(x)=54556054)
+36800 val loss 5.7789
+36800 val perplexity 323.3910
+36800 train 5.646072 (lr=1.2360e-05) (hash(x)=55896927)
+38300 val loss 6.1177
+38300 val perplexity 453.8232
+38300 train 5.962441 (lr=2.1711e-05) (hash(x)=48944700)
+39200 val loss 5.8986
+39200 val perplexity 364.5089
+39200 train 6.256696 (lr=1.4032e-05) (hash(x)=48915509)
+36900 val loss 5.7768
+36900 val perplexity 322.7098
+36900 train 5.590997 (lr=1.2256e-05) (hash(x)=49812554)
+38400 val loss 6.1172
+38400 val perplexity 453.5770
+38400 train 5.777618 (lr=2.1521e-05) (hash(x)=46504848)
+39300 val loss 5.9054
+39300 val perplexity 367.0246
+39300 train 5.709824 (lr=1.3908e-05) (hash(x)=46121589)
+37000 val loss 5.7821
+37000 val perplexity 324.4427
+37000 train 5.613402 (lr=1.2152e-05) (hash(x)=49391771)
+38500 val loss 6.1143
+38500 val perplexity 452.2820
+38500 train 5.850760 (lr=2.1332e-05) (hash(x)=50148948)
+39400 val loss 5.9016
+39400 val perplexity 365.6397
+39400 train 5.460722 (lr=1.3784e-05) (hash(x)=45556201)
+37100 val loss 5.7746
+37100 val perplexity 322.0222
+37100 train 5.731025 (lr=1.2048e-05) (hash(x)=50515328)
+38600 val loss 6.1168
+38600 val perplexity 453.4129
+38600 train 6.839215 (lr=2.1144e-05) (hash(x)=60660913)
+39500 val loss 5.8980
+39500 val perplexity 364.2948
+39500 train 5.912900 (lr=1.3661e-05) (hash(x)=47014749)
+37200 val loss 5.7738
+37200 val perplexity 321.7449
+37200 train 5.990785 (lr=1.1945e-05) (hash(x)=57788407)
+38700 val loss 6.1175
+38700 val perplexity 453.7334
+38700 train 5.911247 (lr=2.0957e-05) (hash(x)=50256252)
+39600 val loss 5.9021
+39600 val perplexity 365.8049
+39600 train 5.702456 (lr=1.3540e-05) (hash(x)=47395114)
+37300 val loss 5.7715
+37300 val perplexity 321.0235
+37300 train 5.544409 (lr=1.1843e-05) (hash(x)=46333526)
+38800 val loss 6.1133
+38800 val perplexity 451.8223
+38800 train 6.149075 (lr=2.0773e-05) (hash(x)=52592415)
+39700 val loss 5.9055
+39700 val perplexity 367.0372
+39700 train 6.017768 (lr=1.3419e-05) (hash(x)=50858008)
+37400 val loss 5.7680
+37400 val perplexity 319.8995
+37400 train 5.567012 (lr=1.1741e-05) (hash(x)=45862038)
+38900 val loss 6.1136
+38900 val perplexity 451.9839
+38900 train 6.061025 (lr=2.0589e-05) (hash(x)=51945430)
+39800 val loss 5.8998
+39800 val perplexity 364.9705
+39800 train 5.619897 (lr=1.3299e-05) (hash(x)=44096434)
+37500 val loss 5.7672
+37500 val perplexity 319.6557
+37500 train 5.480631 (lr=1.1640e-05) (hash(x)=46543051)
+39000 val loss 6.1031
+39000 val perplexity 447.2501
+39000 train 6.308416 (lr=2.0407e-05) (hash(x)=51838869)
+39900 val loss 5.9007
+39900 val perplexity 365.2943
+39900 train 5.654382 (lr=1.3180e-05) (hash(x)=46902151)
+37600 val loss 5.7685
+37600 val perplexity 320.0498
+37600 train 5.724071 (lr=1.1540e-05) (hash(x)=51314450)
+39100 val loss 6.1018
+39100 val perplexity 446.6725
+39100 train 6.199195 (lr=2.0226e-05) (hash(x)=54556054)
+40000 val loss 5.8979
+40000 val perplexity 364.2790
+40000 train 5.670296 (lr=1.3063e-05) (hash(x)=46735297)
+37700 val loss 5.7655
+37700 val perplexity 319.1130
+37700 train 5.618969 (lr=1.1440e-05) (hash(x)=46706763)
+39200 val loss 6.1041
+39200 val perplexity 447.7082
+39200 train 6.436305 (lr=2.0046e-05) (hash(x)=48915509)
+40100 val loss 5.8982
+40100 val perplexity 364.3735
+40100 train 5.852563 (lr=1.2946e-05) (hash(x)=48035170)
+37800 val loss 5.7649
+37800 val perplexity 318.8958
+37800 train 5.625665 (lr=1.1341e-05) (hash(x)=45793085)
+39300 val loss 6.1047
+39300 val perplexity 447.9523
+39300 train 5.907686 (lr=1.9868e-05) (hash(x)=46121589)
+40200 val loss 5.8964
+40200 val perplexity 363.7285
+40200 train 5.715671 (lr=1.2830e-05) (hash(x)=46651322)
+37900 val loss 5.7650
+37900 val perplexity 318.9442
+37900 train 5.611680 (lr=1.1243e-05) (hash(x)=49969493)
+39400 val loss 6.0985
+39400 val perplexity 445.2034
+39400 train 5.652236 (lr=1.9692e-05) (hash(x)=45556201)
+40300 val loss 5.8976
+40300 val perplexity 364.1667
+40300 train 5.741758 (lr=1.2716e-05) (hash(x)=46378099)
+38000 val loss 5.7623
+38000 val perplexity 318.0941
+38000 train 5.559613 (lr=1.1145e-05) (hash(x)=47184489)
+39500 val loss 6.1009
+39500 val perplexity 446.2423
+39500 train 6.076194 (lr=1.9516e-05) (hash(x)=47014749)
+40400 val loss 5.8924
+40400 val perplexity 362.2703
+40400 train 5.743531 (lr=1.2602e-05) (hash(x)=46495438)
+38100 val loss 5.7646
+38100 val perplexity 318.8094
+38100 train 6.227178 (lr=1.1048e-05) (hash(x)=54905742)
+39600 val loss 6.1081
+39600 val perplexity 449.4911
+39600 train 5.875370 (lr=1.9342e-05) (hash(x)=47395114)
+40500 val loss 5.8888
+40500 val perplexity 360.9549
+40500 train 6.038678 (lr=1.2489e-05) (hash(x)=52059147)
+38200 val loss 5.7636
+38200 val perplexity 318.5088
+38200 train 6.033066 (lr=1.0951e-05) (hash(x)=48831719)
+39700 val loss 6.1090
+39700 val perplexity 449.8992
+39700 train 6.190334 (lr=1.9170e-05) (hash(x)=50858008)
+40600 val loss 5.8898
+40600 val perplexity 361.3468
+40600 train 6.023067 (lr=1.2378e-05) (hash(x)=54885045)
+38300 val loss 5.7640
+38300 val perplexity 318.6113
+38300 train 5.599991 (lr=1.0855e-05) (hash(x)=48944700)
+39800 val loss 6.0991
+39800 val perplexity 445.4457
+39800 train 5.810238 (lr=1.8999e-05) (hash(x)=44096434)
+40700 val loss 5.8854
+40700 val perplexity 359.7634
+40700 train 6.309203 (lr=1.2267e-05) (hash(x)=53213971)
+38400 val loss 5.7637
+38400 val perplexity 318.5376
+38400 train 5.379463 (lr=1.0760e-05) (hash(x)=46504848)
+39900 val loss 6.0979
+39900 val perplexity 444.9122
+39900 train 5.845292 (lr=1.8829e-05) (hash(x)=46902151)
+40800 val loss 5.8834
+40800 val perplexity 359.0162
+40800 train 5.555313 (lr=1.2158e-05) (hash(x)=45133794)
+38500 val loss 5.7629
+38500 val perplexity 318.2579
+38500 train 5.494953 (lr=1.0666e-05) (hash(x)=50148948)
+40000 val loss 6.1015
+40000 val perplexity 446.5135
+40000 train 5.857305 (lr=1.8661e-05) (hash(x)=46735297)
+40900 val loss 5.8835
+40900 val perplexity 359.0653
+40900 train 6.047424 (lr=1.2049e-05) (hash(x)=56546547)
+38600 val loss 5.7637
+38600 val perplexity 318.5261
+38600 train 6.555881 (lr=1.0572e-05) (hash(x)=60660913)
+40100 val loss 6.1004
+40100 val perplexity 446.0259
+40100 train 6.029654 (lr=1.8494e-05) (hash(x)=48035170)
+41000 val loss 5.8819
+41000 val perplexity 358.4984
+41000 train 6.263039 (lr=1.1942e-05) (hash(x)=49552270)
+38700 val loss 5.7654
+38700 val perplexity 319.0681
+38700 train 5.562728 (lr=1.0479e-05) (hash(x)=50256252)
+40200 val loss 6.0984
+40200 val perplexity 445.1240
+40200 train 5.900217 (lr=1.8329e-05) (hash(x)=46651322)
+41100 val loss 5.8834
+41100 val perplexity 359.0388
+41100 train 6.105730 (lr=1.1836e-05) (hash(x)=51222370)
+38800 val loss 5.7634
+38800 val perplexity 318.4179
+38800 train 5.842839 (lr=1.0386e-05) (hash(x)=52592415)
+40300 val loss 6.1051
+40300 val perplexity 448.1296
+40300 train 5.948763 (lr=1.8165e-05) (hash(x)=46378099)
+41200 val loss 5.8920
+41200 val perplexity 362.1415
+41200 train 5.928735 (lr=1.1730e-05) (hash(x)=50883608)
+38900 val loss 5.7602
+38900 val perplexity 317.4065
+38900 train 5.703421 (lr=1.0294e-05) (hash(x)=51945430)
+40400 val loss 6.0977
+40400 val perplexity 444.8231
+40400 train 5.962878 (lr=1.8003e-05) (hash(x)=46495438)
+41300 val loss 5.8873
+41300 val perplexity 360.4363
+41300 train 6.133073 (lr=1.1626e-05) (hash(x)=52996828)
+39000 val loss 5.7592
+39000 val perplexity 317.1058
+39000 train 5.992633 (lr=1.0203e-05) (hash(x)=51838869)
+40500 val loss 6.0893
+40500 val perplexity 441.0985
+40500 train 6.257522 (lr=1.7842e-05) (hash(x)=52059147)
+41400 val loss 5.8895
+41400 val perplexity 361.2230
+41400 train 5.853330 (lr=1.1523e-05) (hash(x)=48822716)
+39100 val loss 5.7557
+39100 val perplexity 315.9767
+39100 train 5.885694 (lr=1.0113e-05) (hash(x)=54556054)
+40600 val loss 6.0907
+40600 val perplexity 441.7514
+40600 train 6.233351 (lr=1.7683e-05) (hash(x)=54885045)
+41500 val loss 5.8887
+41500 val perplexity 360.9361
+41500 train 5.860559 (lr=1.1421e-05) (hash(x)=50936577)
+39200 val loss 5.7568
+39200 val perplexity 316.3292
+39200 train 6.130731 (lr=1.0023e-05) (hash(x)=48915509)
+40700 val loss 6.0848
+40700 val perplexity 439.1400
+40700 train 6.507118 (lr=1.7525e-05) (hash(x)=53213971)
+41600 val loss 5.8957
+41600 val perplexity 363.4856
+41600 train 5.931068 (lr=1.1320e-05) (hash(x)=44375046)
+39300 val loss 5.7592
+39300 val perplexity 317.0928
+39300 train 5.582416 (lr=9.9341e-06) (hash(x)=46121589)
+40800 val loss 6.0815
+40800 val perplexity 437.7063
+40800 train 5.749034 (lr=1.7368e-05) (hash(x)=45133794)
+41700 val loss 5.8856
+41700 val perplexity 359.8251
+41700 train 5.654349 (lr=1.1220e-05) (hash(x)=44060021)
+40900 val loss 6.0838
+40900 val perplexity 438.6890
+40900 train 6.203508 (lr=1.7214e-05) (hash(x)=56546547)
+39400 val loss 5.7545
+39400 val perplexity 315.6202
+39400 train 5.293779 (lr=9.8458e-06) (hash(x)=45556201)
+41800 val loss 5.8851
+41800 val perplexity 359.6262
+41800 train 6.167309 (lr=1.1121e-05) (hash(x)=57765221)
+41000 val loss 6.0783
+41000 val perplexity 436.3056
+41000 train 6.447348 (lr=1.7060e-05) (hash(x)=49552270)
+39500 val loss 5.7557
+39500 val perplexity 315.9789
+39500 train 5.796134 (lr=9.7581e-06) (hash(x)=47014749)
+41900 val loss 5.8858
+41900 val perplexity 359.8881
+41900 train 5.638933 (lr=1.1024e-05) (hash(x)=46051470)
+41100 val loss 6.0833
+41100 val perplexity 438.4789
+41100 train 6.289819 (lr=1.6908e-05) (hash(x)=51222370)
+39600 val loss 5.7580
+39600 val perplexity 316.7281
+39600 train 5.561605 (lr=9.6712e-06) (hash(x)=47395114)
+42000 val loss 5.8856
+42000 val perplexity 359.8166
+42000 train 6.123823 (lr=1.0927e-05) (hash(x)=52077616)
+39700 val loss 5.7605
+39700 val perplexity 317.5004
+39700 train 5.875614 (lr=9.5849e-06) (hash(x)=50858008)
+41200 val loss 6.0801
+41200 val perplexity 437.0869
+41200 train 6.125017 (lr=1.6758e-05) (hash(x)=50883608)
+42100 val loss 5.8841
+42100 val perplexity 359.2952
+42100 train 5.741609 (lr=1.0831e-05) (hash(x)=47845199)
+39800 val loss 5.7552
+39800 val perplexity 315.8245
+39800 train 5.476305 (lr=9.4994e-06) (hash(x)=44096434)
+41300 val loss 6.0845
+41300 val perplexity 438.9784
+41300 train 6.330495 (lr=1.6609e-05) (hash(x)=52996828)
+42200 val loss 5.8834
+42200 val perplexity 359.0181
+42200 train 6.128910 (lr=1.0737e-05) (hash(x)=51549823)
+39900 val loss 5.7576
+39900 val perplexity 316.5782
+39900 train 5.512799 (lr=9.4146e-06) (hash(x)=46902151)
+41400 val loss 6.0846
+41400 val perplexity 439.0529
+41400 train 6.062356 (lr=1.6462e-05) (hash(x)=48822716)
+42300 val loss 5.8727
+42300 val perplexity 355.2144
+42300 train 6.019602 (lr=1.0644e-05) (hash(x)=56922131)
+41500 val loss 6.0899
+41500 val perplexity 441.3937
+40000 val loss 5.7535
+40000 val perplexity 315.2927
+41500 train 6.054962 (lr=1.6316e-05) (hash(x)=50936577)
+40000 train 5.562321 (lr=9.3305e-06) (hash(x)=46735297)
+42400 val loss 5.8732
+42400 val perplexity 355.3774
+42400 train 5.817928 (lr=1.0552e-05) (hash(x)=49004372)
+41600 val loss 6.0941
+41600 val perplexity 443.2212
+41600 train 6.104293 (lr=1.6172e-05) (hash(x)=44375046)
+40100 val loss 5.7523
+40100 val perplexity 314.9220
+40100 train 5.715792 (lr=9.2472e-06) (hash(x)=48035170)
+42500 val loss 5.8702
+42500 val perplexity 354.3189
+42500 train 5.842021 (lr=1.0460e-05) (hash(x)=50651839)
+41700 val loss 6.0870
+41700 val perplexity 440.1031
+41700 train 5.825902 (lr=1.6029e-05) (hash(x)=44060021)
+40200 val loss 5.7523
+40200 val perplexity 314.9005
+40200 train 5.575696 (lr=9.1646e-06) (hash(x)=46651322)
+42600 val loss 5.8698
+42600 val perplexity 354.1635
+42600 train 6.166778 (lr=1.0370e-05) (hash(x)=50767721)
+41800 val loss 6.0914
+41800 val perplexity 442.0266
+41800 train 6.348953 (lr=1.5888e-05) (hash(x)=57765221)
+40300 val loss 5.7509
+40300 val perplexity 314.4647
+40300 train 5.593280 (lr=9.0827e-06) (hash(x)=46378099)
+42700 val loss 5.8683
+42700 val perplexity 353.6617
+42700 train 5.581938 (lr=1.0282e-05) (hash(x)=49099183)
+41900 val loss 6.0879
+41900 val perplexity 440.4934
+41900 train 5.849139 (lr=1.5748e-05) (hash(x)=46051470)
+40400 val loss 5.7455
+40400 val perplexity 312.7840
+40400 train 5.585403 (lr=9.0015e-06) (hash(x)=46495438)
+42800 val loss 5.8679
+42800 val perplexity 353.4950
+42800 train 6.413287 (lr=1.0194e-05) (hash(x)=42272413)
+42000 val loss 6.0918
+42000 val perplexity 442.2172
+42000 train 6.311883 (lr=1.5610e-05) (hash(x)=52077616)
+40500 val loss 5.7442
+40500 val perplexity 312.3704
+40500 train 5.897409 (lr=8.9211e-06) (hash(x)=52059147)
+42900 val loss 5.8702
+42900 val perplexity 354.3152
+42900 train 5.672939 (lr=1.0107e-05) (hash(x)=48582863)
+42100 val loss 6.0902
+42100 val perplexity 441.4968
+42100 train 5.944408 (lr=1.5474e-05) (hash(x)=47845199)
+40600 val loss 5.7484
+40600 val perplexity 313.7004
+40600 train 5.902262 (lr=8.8414e-06) (hash(x)=54885045)
+43000 val loss 5.8681
+43000 val perplexity 353.5818
+43000 train 5.837842 (lr=1.0022e-05) (hash(x)=48703446)
+42200 val loss 6.0861
+42200 val perplexity 439.7129
+42200 train 6.341642 (lr=1.5339e-05) (hash(x)=51549823)
+40700 val loss 5.7442
+40700 val perplexity 312.3774
+40700 train 6.190853 (lr=8.7624e-06) (hash(x)=53213971)
+43100 val loss 5.8689
+43100 val perplexity 353.8753
+43100 train 5.806827 (lr=9.9373e-06) (hash(x)=48730321)
+42300 val loss 6.0799
+42300 val perplexity 436.9808
+42300 train 6.246227 (lr=1.5205e-05) (hash(x)=56922131)
+40800 val loss 5.7420
+40800 val perplexity 311.6921
+40800 train 5.419962 (lr=8.6842e-06) (hash(x)=45133794)
+43200 val loss 5.8685
+43200 val perplexity 353.7095
+43200 train 6.086263 (lr=9.8541e-06) (hash(x)=56536090)
+42400 val loss 6.0783
+42400 val perplexity 436.2948
+42400 train 6.009583 (lr=1.5074e-05) (hash(x)=49004372)
+43300 val loss 5.8681
+43300 val perplexity 353.5628
+43300 train 5.982976 (lr=9.7720e-06) (hash(x)=54154116)
+40900 val loss 5.7432
+40900 val perplexity 312.0596
+40900 train 5.886569 (lr=8.6068e-06) (hash(x)=56546547)
+42500 val loss 6.0735
+42500 val perplexity 434.1761
+42500 train 6.045653 (lr=1.4943e-05) (hash(x)=50651839)
+43400 val loss 5.8719
+43400 val perplexity 354.9122
+43400 train 5.750636 (lr=9.6911e-06) (hash(x)=50058055)
+41000 val loss 5.7402
+41000 val perplexity 311.1349
+41000 train 6.131865 (lr=8.5301e-06) (hash(x)=49552270)
+42600 val loss 6.0687
+42600 val perplexity 432.1059
+42600 train 6.374171 (lr=1.4815e-05) (hash(x)=50767721)
+43500 val loss 5.8686
+43500 val perplexity 353.7496
+43500 train 5.801444 (lr=9.6113e-06) (hash(x)=48743802)
+41100 val loss 5.7393
+41100 val perplexity 310.8314
+41100 train 5.977370 (lr=8.4541e-06) (hash(x)=51222370)
+42700 val loss 6.0701
+42700 val perplexity 432.7057
+42700 train 5.794621 (lr=1.4688e-05) (hash(x)=49099183)
+43600 val loss 5.8708
+43600 val perplexity 354.5286
+43600 train 5.519044 (lr=9.5326e-06) (hash(x)=42792886)
+41200 val loss 5.7427
+41200 val perplexity 311.9193
+41200 train 5.779368 (lr=8.3789e-06) (hash(x)=50883608)
+42800 val loss 6.0708
+42800 val perplexity 433.0135
+42800 train 6.560636 (lr=1.4563e-05) (hash(x)=42272413)
+43700 val loss 5.8688
+43700 val perplexity 353.8321
+43700 train 6.247068 (lr=9.4552e-06) (hash(x)=56446070)
+41300 val loss 5.7443
+41300 val perplexity 312.3972
+41300 train 5.995403 (lr=8.3045e-06) (hash(x)=52996828)
+42900 val loss 6.0737
+42900 val perplexity 434.2632
+42900 train 5.865055 (lr=1.4439e-05) (hash(x)=48582863)
+43800 val loss 5.8704
+43800 val perplexity 354.3770
+43800 train 5.713625 (lr=9.3788e-06) (hash(x)=45584354)
+41400 val loss 5.7400
+41400 val perplexity 311.0588
+41400 train 5.687901 (lr=8.2308e-06) (hash(x)=48822716)
+43000 val loss 6.0699
+43000 val perplexity 432.6428
+43000 train 6.018393 (lr=1.4317e-05) (hash(x)=48703446)
+43900 val loss 5.8692
+43900 val perplexity 353.9805
+43900 train 5.711111 (lr=9.3036e-06) (hash(x)=49339253)
+41500 val loss 5.7436
+41500 val perplexity 312.1848
+41500 train 5.718872 (lr=8.1579e-06) (hash(x)=50936577)
+43100 val loss 6.0662
+43100 val perplexity 431.0327
+43100 train 5.987951 (lr=1.4196e-05) (hash(x)=48730321)
+44000 val loss 5.8701
+44000 val perplexity 354.2998
+44000 train 5.685949 (lr=9.2296e-06) (hash(x)=46183203)
+41600 val loss 5.7488
+41600 val perplexity 313.8020
+41600 train 5.803505 (lr=8.0858e-06) (hash(x)=44375046)
+43200 val loss 6.0691
+43200 val perplexity 432.3110
+43200 train 6.268177 (lr=1.4077e-05) (hash(x)=56536090)
+44100 val loss 5.8633
+44100 val perplexity 351.8832
+44100 train 5.847696 (lr=9.1568e-06) (hash(x)=47849630)
+41700 val loss 5.7438
+41700 val perplexity 312.2614
+41700 train 5.521793 (lr=8.0144e-06) (hash(x)=44060021)
+43300 val loss 6.0682
+43300 val perplexity 431.8818
+43300 train 6.172038 (lr=1.3960e-05) (hash(x)=54154116)
+44200 val loss 5.8638
+44200 val perplexity 352.0425
+44200 train 6.163978 (lr=9.0851e-06) (hash(x)=49834275)
+41800 val loss 5.7417
+41800 val perplexity 311.6032
+41800 train 6.056554 (lr=7.9438e-06) (hash(x)=57765221)
+43400 val loss 6.0726
+43400 val perplexity 433.8108
+43400 train 5.960909 (lr=1.3844e-05) (hash(x)=50058055)
+44300 val loss 5.8622
+44300 val perplexity 351.4916
+44300 train 6.319765 (lr=9.0146e-06) (hash(x)=62535257)
+41900 val loss 5.7419
+41900 val perplexity 311.6466
+41900 train 5.494161 (lr=7.8740e-06) (hash(x)=46051470)
+43500 val loss 6.0713
+43500 val perplexity 433.2254
+43500 train 5.989706 (lr=1.3730e-05) (hash(x)=48743802)
+44400 val loss 5.8612
+44400 val perplexity 351.1554
+44400 train 5.913574 (lr=8.9453e-06) (hash(x)=49253957)
+42000 val loss 5.7454
+42000 val perplexity 312.7515
+42000 train 5.963854 (lr=7.8050e-06) (hash(x)=52077616)
+43600 val loss 6.0686
+43600 val perplexity 432.0546
+43600 train 5.709352 (lr=1.3618e-05) (hash(x)=42792886)
+44500 val loss 5.8609
+44500 val perplexity 351.0523
+44500 train 5.816366 (lr=8.8771e-06) (hash(x)=55368339)
+42100 val loss 5.7410
+42100 val perplexity 311.3617
+42100 train 5.579000 (lr=7.7368e-06) (hash(x)=47845199)
+43700 val loss 6.0663
+43700 val perplexity 431.0718
+43700 train 6.482207 (lr=1.3507e-05) (hash(x)=56446070)
+44600 val loss 5.8624
+44600 val perplexity 351.5555
+44600 train 5.944849 (lr=8.8101e-06) (hash(x)=47098476)
+42200 val loss 5.7395
+42200 val perplexity 310.9067
+42200 train 5.996667 (lr=7.6693e-06) (hash(x)=51549823)
+43800 val loss 6.0681
+43800 val perplexity 431.8659
+43800 train 5.916797 (lr=1.3398e-05) (hash(x)=45584354)
+44700 val loss 5.8614
+44700 val perplexity 351.2013
+44700 train 5.784379 (lr=8.7443e-06) (hash(x)=48280562)
+43900 val loss 6.0683
+43900 val perplexity 431.9393
+42300 val loss 5.7336
+42300 val perplexity 309.0953
+43900 train 5.914937 (lr=1.3291e-05) (hash(x)=49339253)
+42300 train 5.855470 (lr=7.6027e-06) (hash(x)=56922131)
+44800 val loss 5.8671
+44800 val perplexity 353.2086
+44800 train 6.136794 (lr=8.6797e-06) (hash(x)=55591638)
+44000 val loss 6.0671
+44000 val perplexity 431.4092
+44000 train 5.857544 (lr=1.3185e-05) (hash(x)=46183203)
+42400 val loss 5.7331
+42400 val perplexity 308.9134
+42400 train 5.686144 (lr=7.5368e-06) (hash(x)=49004372)
+44900 val loss 5.8650
+44900 val perplexity 352.4780
+44900 train 6.042705 (lr=8.6163e-06) (hash(x)=53757748)
+44100 val loss 6.0625
+44100 val perplexity 429.4332
+44100 train 6.027684 (lr=1.3081e-05) (hash(x)=47849630)
+42500 val loss 5.7304
+42500 val perplexity 308.0865
+42500 train 5.707544 (lr=7.4717e-06) (hash(x)=50651839)
+45000 val loss 5.8605
+45000 val perplexity 350.8926
+45000 train 6.009489 (lr=8.5540e-06) (hash(x)=51685087)
+44200 val loss 6.0632
+44200 val perplexity 429.7634
+44200 train 6.403512 (lr=1.2979e-05) (hash(x)=49834275)
+42600 val loss 5.7309
+42600 val perplexity 308.2605
+42600 train 6.003824 (lr=7.4074e-06) (hash(x)=50767721)
+45100 val loss 5.8595
+45100 val perplexity 350.5572
+45100 train 5.871101 (lr=8.4930e-06) (hash(x)=50093774)
+44300 val loss 6.0562
+44300 val perplexity 426.7346
+44300 train 6.502149 (lr=1.2878e-05) (hash(x)=62535257)
+42700 val loss 5.7305
+42700 val perplexity 308.1259
+42700 train 5.442407 (lr=7.3440e-06) (hash(x)=49099183)
+45200 val loss 5.8601
+45200 val perplexity 350.7615
+45200 train 5.580861 (lr=8.4331e-06) (hash(x)=43460450)
+44400 val loss 6.0591
+44400 val perplexity 427.9775
+44400 train 6.109484 (lr=1.2779e-05) (hash(x)=49253957)
+42800 val loss 5.7267
+42800 val perplexity 306.9621
+42800 train 6.192804 (lr=7.2813e-06) (hash(x)=42272413)
+45300 val loss 5.8632
+45300 val perplexity 351.8497
+45300 train 6.055310 (lr=8.3745e-06) (hash(x)=49935488)
+44500 val loss 6.0592
+44500 val perplexity 428.0197
+44500 train 6.028224 (lr=1.2682e-05) (hash(x)=55368339)
+45400 val loss 5.8620
+45400 val perplexity 351.4144
+45400 train 5.832631 (lr=8.3170e-06) (hash(x)=49447929)
+42900 val loss 5.7333
+42900 val perplexity 308.9987
+42900 train 5.529157 (lr=7.2194e-06) (hash(x)=48582863)
+44600 val loss 6.0570
+44600 val perplexity 427.1129
+44600 train 6.093005 (lr=1.2586e-05) (hash(x)=47098476)
+45500 val loss 5.8609
+45500 val perplexity 351.0280
+45500 train 5.806209 (lr=8.2607e-06) (hash(x)=50713904)
+43000 val loss 5.7272
+43000 val perplexity 307.1086
+43000 train 5.717058 (lr=7.1583e-06) (hash(x)=48703446)
+45600 val loss 5.8931
+45600 val perplexity 362.5154
+45600 train 5.801927 (lr=8.2057e-06) (hash(x)=47674606)
+44700 val loss 6.0582
+44700 val perplexity 427.6148
+44700 train 5.950908 (lr=1.2492e-05) (hash(x)=48280562)
+43100 val loss 5.7278
+43100 val perplexity 307.2908
+43100 train 5.654952 (lr=7.0981e-06) (hash(x)=48730321)
+45700 val loss 5.8621
+45700 val perplexity 351.4761
+45700 train 5.754256 (lr=8.1518e-06) (hash(x)=51539617)
+44800 val loss 6.0574
+44800 val perplexity 427.2695
+44800 train 6.307104 (lr=1.2400e-05) (hash(x)=55591638)
+43200 val loss 5.7288
+43200 val perplexity 307.6085
+43200 train 5.956896 (lr=7.0386e-06) (hash(x)=56536090)
+44900 val loss 6.0572
+44900 val perplexity 427.1797
+44900 train 6.209588 (lr=1.2309e-05) (hash(x)=53757748)
+45800 val loss 5.8620
+45800 val perplexity 351.4243
+45800 train 5.833173 (lr=8.0992e-06) (hash(x)=44448785)
+43300 val loss 5.7274
+43300 val perplexity 307.1761
+43300 train 5.866704 (lr=6.9800e-06) (hash(x)=54154116)
+45000 val loss 6.0571
+45000 val perplexity 427.1190
+45000 train 6.189182 (lr=1.2220e-05) (hash(x)=51685087)
+45900 val loss 5.8606
+45900 val perplexity 350.9274
+45900 train 5.672136 (lr=8.0478e-06) (hash(x)=51499105)
+43400 val loss 5.7301
+43400 val perplexity 307.9856
+43400 train 5.614723 (lr=6.9222e-06) (hash(x)=50058055)
+45100 val loss 6.0571
+45100 val perplexity 427.1398
+45100 train 6.055492 (lr=1.2133e-05) (hash(x)=50093774)
+46000 val loss 5.8559
+46000 val perplexity 349.2964
+46000 train 5.756037 (lr=7.9976e-06) (hash(x)=48359464)
+43500 val loss 5.7306
+43500 val perplexity 308.1495
+43500 train 5.641300 (lr=6.8652e-06) (hash(x)=48743802)
+45200 val loss 6.0550
+45200 val perplexity 426.2207
+45200 train 5.782551 (lr=1.2047e-05) (hash(x)=43460450)
+46100 val loss 5.8582
+46100 val perplexity 350.1075
+46100 train 5.897924 (lr=7.9485e-06) (hash(x)=51885986)
+43600 val loss 5.7286
+43600 val perplexity 307.5317
+43600 train 5.374582 (lr=6.8090e-06) (hash(x)=42792886)
+45300 val loss 6.0555
+45300 val perplexity 426.4360
+45300 train 6.190984 (lr=1.1964e-05) (hash(x)=49935488)
+46200 val loss 5.8500
+46200 val perplexity 347.2492
+46200 train 6.391445 (lr=7.9008e-06) (hash(x)=65186615)
+43700 val loss 5.7287
+43700 val perplexity 307.5639
+43700 train 6.129159 (lr=6.7537e-06) (hash(x)=56446070)
+45400 val loss 6.0603
+45400 val perplexity 428.5037
+45400 train 6.018839 (lr=1.1881e-05) (hash(x)=49447929)
+46300 val loss 5.8518
+46300 val perplexity 347.8571
+46300 train 5.823817 (lr=7.8542e-06) (hash(x)=49626999)
+43800 val loss 5.7300
+43800 val perplexity 307.9766
+43800 train 5.549335 (lr=6.6992e-06) (hash(x)=45584354)
+45500 val loss 6.0581
+45500 val perplexity 427.5738
+45500 train 5.983250 (lr=1.1801e-05) (hash(x)=50713904)
+46400 val loss 5.8525
+46400 val perplexity 348.0972
+46400 train 5.741714 (lr=7.8088e-06) (hash(x)=43325701)
+43900 val loss 5.7271
+43900 val perplexity 307.0764
+43900 train 5.571040 (lr=6.6455e-06) (hash(x)=49339253)
+45600 val loss 6.0809
+45600 val perplexity 437.4357
+45600 train 5.990339 (lr=1.1722e-05) (hash(x)=47674606)
+46500 val loss 5.8523
+46500 val perplexity 348.0345
+46500 train 6.141178 (lr=7.7647e-06) (hash(x)=54028595)
+44000 val loss 5.7273
+44000 val perplexity 307.1531
+44000 train 5.546133 (lr=6.5926e-06) (hash(x)=46183203)
+45700 val loss 6.0576
+45700 val perplexity 427.3579
+45700 train 5.974851 (lr=1.1645e-05) (hash(x)=51539617)
+46600 val loss 5.8544
+46600 val perplexity 348.7543
+46600 train 5.707172 (lr=7.7218e-06) (hash(x)=44519175)
+44100 val loss 5.7220
+44100 val perplexity 305.5253
+44100 train 5.738289 (lr=6.5406e-06) (hash(x)=47849630)
+45800 val loss 6.0620
+45800 val perplexity 429.2244
+46700 val loss 5.8538
+46700 val perplexity 348.5613
+45800 train 6.021019 (lr=1.1570e-05) (hash(x)=44448785)
+46700 train 5.908800 (lr=7.6801e-06) (hash(x)=48357998)
+44200 val loss 5.7261
+44200 val perplexity 306.7854
+44200 train 6.044593 (lr=6.4894e-06) (hash(x)=49834275)
+45900 val loss 6.0583
+45900 val perplexity 427.6523
+46800 val loss 5.8497
+46800 val perplexity 347.1213
+45900 train 5.886087 (lr=1.1497e-05) (hash(x)=51499105)
+46800 train 6.041516 (lr=7.6397e-06) (hash(x)=55911353)
+44300 val loss 5.7182
+44300 val perplexity 304.3456
+44300 train 6.210258 (lr=6.4390e-06) (hash(x)=62535257)
+46900 val loss 5.8517
+46900 val perplexity 347.8130
+46000 val loss 6.0527
+46000 val perplexity 425.2731
+46900 train 5.773681 (lr=7.6004e-06) (hash(x)=47897187)
+46000 train 5.946332 (lr=1.1425e-05) (hash(x)=48359464)
+47000 val loss 5.8494
+47000 val perplexity 347.0211
+47000 train 5.433744 (lr=7.5624e-06) (hash(x)=43196571)
+44400 val loss 5.7177
+44400 val perplexity 304.2032
+44400 train 5.762341 (lr=6.3895e-06) (hash(x)=49253957)
+46100 val loss 6.0516
+46100 val perplexity 424.7917
+46100 train 6.077628 (lr=1.1355e-05) (hash(x)=51885986)
+47100 val loss 5.8501
+47100 val perplexity 347.2580
+47100 train 5.718368 (lr=7.5257e-06) (hash(x)=51224987)
+44500 val loss 5.7176
+44500 val perplexity 304.1715
+46200 val loss 6.0464
+46200 val perplexity 422.5769
+44500 train 5.675300 (lr=6.3408e-06) (hash(x)=55368339)
+46200 train 6.599186 (lr=1.1287e-05) (hash(x)=65186615)
+47200 val loss 5.8540
+47200 val perplexity 348.6273
+47200 train 5.550478 (lr=7.4901e-06) (hash(x)=47943697)
+46300 val loss 6.0460
+46300 val perplexity 422.4326
+46300 train 6.010680 (lr=1.1220e-05) (hash(x)=49626999)
+44600 val loss 5.7185
+44600 val perplexity 304.4501
+44600 train 5.850135 (lr=6.2929e-06) (hash(x)=47098476)
+47300 val loss 5.8548
+47300 val perplexity 348.9183
+47300 train 5.792697 (lr=7.4558e-06) (hash(x)=47351003)
+46400 val loss 6.0468
+46400 val perplexity 422.7397
+46400 train 5.928525 (lr=1.1155e-05) (hash(x)=43325701)
+44700 val loss 5.7176
+44700 val perplexity 304.1606
+44700 train 5.639624 (lr=6.2459e-06) (hash(x)=48280562)
+47400 val loss 5.8528
+47400 val perplexity 348.2169
+47400 train 5.919446 (lr=7.4228e-06) (hash(x)=55562243)
+46500 val loss 6.0487
+46500 val perplexity 423.5460
+46500 train 6.316158 (lr=1.1092e-05) (hash(x)=54028595)
+44800 val loss 5.7194
+44800 val perplexity 304.7143
+44800 train 6.008594 (lr=6.1998e-06) (hash(x)=55591638)
+47500 val loss 5.8510
+47500 val perplexity 347.5935
+47500 train 5.729331 (lr=7.3909e-06) (hash(x)=53544850)
+46600 val loss 6.0485
+46600 val perplexity 423.4939
+46600 train 5.879036 (lr=1.1031e-05) (hash(x)=44519175)
+44900 val loss 5.7217
+44900 val perplexity 305.4323
+44900 train 5.873918 (lr=6.1545e-06) (hash(x)=53757748)
+47600 val loss 5.8521
+47600 val perplexity 347.9716
+47600 train 6.031857 (lr=7.3603e-06) (hash(x)=43634907)
+46700 val loss 6.0500
+46700 val perplexity 424.1289
+46700 train 6.119004 (lr=1.0972e-05) (hash(x)=48357998)
+45000 val loss 5.7176
+45000 val perplexity 304.1754
+45000 train 5.889555 (lr=6.1100e-06) (hash(x)=51685087)
+47700 val loss 5.8513
+47700 val perplexity 347.6973
+47700 train 5.841846 (lr=7.3310e-06) (hash(x)=47909383)
+46800 val loss 6.0481
+46800 val perplexity 423.3104
+46800 train 6.212865 (lr=1.0914e-05) (hash(x)=55911353)
+45100 val loss 5.7174
+45100 val perplexity 304.1154
+45100 train 5.737835 (lr=6.0664e-06) (hash(x)=50093774)
+47800 val loss 5.8470
+47800 val perplexity 346.2068
+47800 train 5.702913 (lr=7.3029e-06) (hash(x)=45871079)
+46900 val loss 6.0458
+46900 val perplexity 422.3537
+46900 train 5.969927 (lr=1.0858e-05) (hash(x)=47897187)
+45200 val loss 5.7165
+45200 val perplexity 303.8303
+45200 train 5.434013 (lr=6.0237e-06) (hash(x)=43460450)
+47900 val loss 5.8458
+47900 val perplexity 345.7645
+47900 train 5.847965 (lr=7.2760e-06) (hash(x)=47333324)
+47000 val loss 6.0454
+47000 val perplexity 422.1632
+47000 train 5.631547 (lr=1.0803e-05) (hash(x)=43196571)
+45300 val loss 5.7197
+45300 val perplexity 304.8224
+45300 train 5.949478 (lr=5.9818e-06) (hash(x)=49935488)
+48000 val loss 5.8457
+48000 val perplexity 345.7522
+48000 train 5.926929 (lr=7.2504e-06) (hash(x)=52758020)
+47100 val loss 6.0468
+47100 val perplexity 422.7625
+47100 train 5.896503 (lr=1.0751e-05) (hash(x)=51224987)
+45400 val loss 5.7188
+45400 val perplexity 304.5240
+45400 train 5.714464 (lr=5.9407e-06) (hash(x)=49447929)
+48100 val loss 5.8417
+48100 val perplexity 344.3708
+48100 train 5.853239 (lr=7.2260e-06) (hash(x)=49806349)
+47200 val loss 6.0475
+47200 val perplexity 423.0751
+47200 train 5.746195 (lr=1.0700e-05) (hash(x)=47943697)
+45500 val loss 5.7183
+45500 val perplexity 304.3906
+45500 train 5.659379 (lr=5.9005e-06) (hash(x)=50713904)
+48200 val loss 5.8440
+48200 val perplexity 345.1598
+48200 train 5.899774 (lr=7.2029e-06) (hash(x)=53220839)
+47300 val loss 6.0511
+47300 val perplexity 424.5892
+47300 train 5.972193 (lr=1.0651e-05) (hash(x)=47351003)
+48300 val loss 5.8452
+48300 val perplexity 345.5555
+48300 train 5.856475 (lr=7.1810e-06) (hash(x)=56052541)
+45600 val loss 5.7401
+45600 val perplexity 311.0983
+45600 train 5.674036 (lr=5.8612e-06) (hash(x)=47674606)
+47400 val loss 6.0474
+47400 val perplexity 423.0206
+47400 train 6.107438 (lr=1.0604e-05) (hash(x)=55562243)
+48400 val loss 5.8411
+48400 val perplexity 344.1588
+48400 train 5.725373 (lr=7.1603e-06) (hash(x)=44482356)
+45700 val loss 5.7195
+45700 val perplexity 304.7594
+45700 train 5.608114 (lr=5.8227e-06) (hash(x)=51539617)
+47500 val loss 6.0458
+47500 val perplexity 422.3304
+47500 train 5.906383 (lr=1.0558e-05) (hash(x)=53544850)
+48500 val loss 5.8416
+48500 val perplexity 344.3341
+48500 train 5.554918 (lr=7.1409e-06) (hash(x)=45714818)
+45800 val loss 5.7209
+45800 val perplexity 305.1709
+45800 train 5.699752 (lr=5.7851e-06) (hash(x)=44448785)
+47600 val loss 6.0501
+47600 val perplexity 424.1624
+47600 train 6.212331 (lr=1.0515e-05) (hash(x)=43634907)
+48600 val loss 5.8423
+48600 val perplexity 344.5629
+48600 train 5.602136 (lr=7.1228e-06) (hash(x)=49476556)
+45900 val loss 5.7186
+45900 val perplexity 304.4800
+45900 train 5.515782 (lr=5.7484e-06) (hash(x)=51499105)
+47700 val loss 6.0454
+47700 val perplexity 422.1596
+47700 train 6.017856 (lr=1.0473e-05) (hash(x)=47909383)
+48700 val loss 5.8426
+48700 val perplexity 344.6624
+48700 train 5.536637 (lr=7.1059e-06) (hash(x)=42508579)
+46000 val loss 5.7148
+46000 val perplexity 303.3325
+46000 train 5.614646 (lr=5.7125e-06) (hash(x)=48359464)
+47800 val loss 6.0433
+47800 val perplexity 421.2838
+47800 train 5.923380 (lr=1.0433e-05) (hash(x)=45871079)
+48800 val loss 5.8451
+48800 val perplexity 345.5361
+48800 train 6.052786 (lr=7.0902e-06) (hash(x)=52737449)
+46100 val loss 5.7171
+46100 val perplexity 304.0354
+46100 train 5.770822 (lr=5.6775e-06) (hash(x)=51885986)
+47900 val loss 6.0408
+47900 val perplexity 420.2432
+47900 train 6.021464 (lr=1.0394e-05) (hash(x)=47333324)
+48900 val loss 5.8437
+48900 val perplexity 345.0556
+48900 train 5.683334 (lr=7.0758e-06) (hash(x)=47057569)
+46200 val loss 5.7129
+46200 val perplexity 302.7400
+46200 train 6.260228 (lr=5.6434e-06) (hash(x)=65186615)
+48000 val loss 6.0431
+48000 val perplexity 421.1812
+48000 train 6.103298 (lr=1.0358e-05) (hash(x)=52758020)
+49000 val loss 5.8492
+49000 val perplexity 346.9598
+49000 train 5.825048 (lr=7.0627e-06) (hash(x)=49908975)
+46300 val loss 5.7136
+46300 val perplexity 302.9592
+46300 train 5.667426 (lr=5.6101e-06) (hash(x)=49626999)
+48100 val loss 6.0383
+48100 val perplexity 419.1610
+48100 train 6.059237 (lr=1.0323e-05) (hash(x)=49806349)
+49100 val loss 5.8464
+49100 val perplexity 345.9872
+49100 train 5.681278 (lr=7.0508e-06) (hash(x)=48427414)
+46400 val loss 5.7126
+46400 val perplexity 302.6713
+48200 val loss 6.0352
+48200 val perplexity 417.8884
+46400 train 5.609452 (lr=5.5777e-06) (hash(x)=43325701)
+48200 train 6.101655 (lr=1.0290e-05) (hash(x)=53220839)
+49200 val loss 5.8453
+49200 val perplexity 345.6060
+49200 train 5.658861 (lr=7.0401e-06) (hash(x)=50246074)
+48300 val loss 6.0348
+48300 val perplexity 417.7147
+48300 train 6.066577 (lr=1.0259e-05) (hash(x)=56052541)
+46500 val loss 5.7142
+46500 val perplexity 303.1279
+46500 train 6.013310 (lr=5.5462e-06) (hash(x)=54028595)
+49300 val loss 5.8465
+49300 val perplexity 346.0131
+49300 train 5.907238 (lr=7.0307e-06) (hash(x)=47715359)
+48400 val loss 6.0330
+48400 val perplexity 416.9503
+46600 val loss 5.7100
+46600 val perplexity 301.8763
+46600 train 5.570869 (lr=5.5156e-06) (hash(x)=44519175)
+48400 train 5.892282 (lr=1.0229e-05) (hash(x)=44482356)
+49400 val loss 5.8469
+49400 val perplexity 346.1647
+49400 train 5.822443 (lr=7.0226e-06) (hash(x)=50175867)
+46700 val loss 5.7136
+46700 val perplexity 302.9503
+46700 train 5.749952 (lr=5.4858e-06) (hash(x)=48357998)
+48500 val loss 6.0343
+48500 val perplexity 417.5217
+48500 train 5.764053 (lr=1.0201e-05) (hash(x)=45714818)
+46800 val loss 5.7132
+46800 val perplexity 302.8454
+46800 train 5.879297 (lr=5.4569e-06) (hash(x)=55911353)
+49500 val loss 5.8512
+49500 val perplexity 347.6361
+49500 train 5.676702 (lr=7.0157e-06) (hash(x)=49336040)
+48600 val loss 6.0346
+48600 val perplexity 417.6282
+48600 train 5.787707 (lr=1.0175e-05) (hash(x)=49476556)
+46900 val loss 5.7112
+46900 val perplexity 302.2209
+46900 train 5.637585 (lr=5.4289e-06) (hash(x)=47897187)
+49600 val loss 5.8519
+49600 val perplexity 347.8839
+49600 train 5.900337 (lr=7.0100e-06) (hash(x)=52039357)
+48700 val loss 6.0347
+48700 val perplexity 417.6910
+48700 train 5.723627 (lr=1.0151e-05) (hash(x)=42508579)
+47000 val loss 5.7112
+47000 val perplexity 302.2247
+47000 train 5.284904 (lr=5.4017e-06) (hash(x)=43196571)
+49700 val loss 5.8527
+49700 val perplexity 348.1769
+49700 train 5.638364 (lr=7.0056e-06) (hash(x)=47568707)
+48800 val loss 6.0380
+48800 val perplexity 419.0475
+48800 train 6.229567 (lr=1.0129e-05) (hash(x)=52737449)
+47100 val loss 5.7113
+47100 val perplexity 302.2637
+47100 train 5.542101 (lr=5.3755e-06) (hash(x)=51224987)
+49800 val loss 5.8571
+49800 val perplexity 349.7072
+49800 train 5.771938 (lr=7.0025e-06) (hash(x)=48451274)
+48900 val loss 6.0367
+48900 val perplexity 418.5146
+47200 val loss 5.7128
+47200 val perplexity 302.7038
+48900 train 5.891450 (lr=1.0108e-05) (hash(x)=47057569)
+47200 train 5.417474 (lr=5.3501e-06) (hash(x)=47943697)
+49900 val loss 5.8548
+49900 val perplexity 348.9065
+49900 train 5.546834 (lr=7.0006e-06) (hash(x)=44523603)
+47300 val loss 5.7134
+47300 val perplexity 302.8896
+47300 train 5.650457 (lr=5.3256e-06) (hash(x)=47351003)
+49000 val loss 6.0453
+49000 val perplexity 422.1228
+49000 train 6.011272 (lr=1.0090e-05) (hash(x)=49908975)
+49999 val loss 5.8445
+49999 val perplexity 345.3444
+47400 val loss 5.7120
+47400 val perplexity 302.4785
+47400 train 5.765558 (lr=5.3020e-06) (hash(x)=55562243)
+49100 val loss 6.0398
+49100 val perplexity 419.8189
+49100 train 5.864618 (lr=1.0073e-05) (hash(x)=48427414)
+47500 val loss 5.7122
+47500 val perplexity 302.5348
+47500 train 5.578689 (lr=5.2792e-06) (hash(x)=53544850)
+49200 val loss 6.0385
+49200 val perplexity 419.2824
+49200 train 5.857993 (lr=1.0057e-05) (hash(x)=50246074)
+47600 val loss 5.7123
+47600 val perplexity 302.5630
+47600 train 5.958771 (lr=5.2574e-06) (hash(x)=43634907)
+49300 val loss 6.0412
+49300 val perplexity 420.4065
+49300 train 6.071166 (lr=1.0044e-05) (hash(x)=47715359)
+47700 val loss 5.7123
+47700 val perplexity 302.5704
+47700 train 5.694991 (lr=5.2364e-06) (hash(x)=47909383)
+49400 val loss 6.0446
+49400 val perplexity 421.8107
+49400 train 6.025251 (lr=1.0032e-05) (hash(x)=50175867)
+47800 val loss 5.7095
+47800 val perplexity 301.7066
+47800 train 5.585462 (lr=5.2163e-06) (hash(x)=45871079)
+49500 val loss 6.0472
+49500 val perplexity 422.9289
+49500 train 5.871203 (lr=1.0022e-05) (hash(x)=49336040)
+47900 val loss 5.7073
+47900 val perplexity 301.0569
+47900 train 5.720872 (lr=5.1972e-06) (hash(x)=47333324)
+49600 val loss 6.0544
+49600 val perplexity 425.9853
+49600 train 6.064258 (lr=1.0014e-05) (hash(x)=52039357)
+48000 val loss 5.7077
+48000 val perplexity 301.1822
+48000 train 5.798008 (lr=5.1788e-06) (hash(x)=52758020)
+49700 val loss 6.0505
+49700 val perplexity 424.3354
+49700 train 5.830864 (lr=1.0008e-05) (hash(x)=47568707)
+48100 val loss 5.7056
+48100 val perplexity 300.5576
+48100 train 5.700547 (lr=5.1614e-06) (hash(x)=49806349)
+49800 val loss 6.0525
+49800 val perplexity 425.1691
+49800 train 5.946240 (lr=1.0004e-05) (hash(x)=48451274)
+48200 val loss 5.7051
+48200 val perplexity 300.4040
+48200 train 5.781303 (lr=5.1449e-06) (hash(x)=53220839)
+49900 val loss 6.0540
+49900 val perplexity 425.8073
+49900 train 5.733783 (lr=1.0001e-05) (hash(x)=44523603)
+48300 val loss 5.7046
+48300 val perplexity 300.2323
+48300 train 5.714827 (lr=5.1293e-06) (hash(x)=56052541)
+49999 val loss 6.0411
+49999 val perplexity 420.3536
+48400 val loss 5.7032
+48400 val perplexity 299.8373
+48400 train 5.594895 (lr=5.1145e-06) (hash(x)=44482356)
+48500 val loss 5.7032
+48500 val perplexity 299.8374
+48500 train 5.403605 (lr=5.1007e-06) (hash(x)=45714818)
+48600 val loss 5.7045
+48600 val perplexity 300.2178
+48600 train 5.458147 (lr=5.0877e-06) (hash(x)=49476556)
+48700 val loss 5.7060
+48700 val perplexity 300.6585
+48700 train 5.399137 (lr=5.0756e-06) (hash(x)=42508579)
diff --git a/attention_kindselective_n_heads2_seed1340/model_02500.pt b/attention_kindselective_n_heads2_seed1340/model_02500.pt
index 473b3cc45f008ad013670fa4bacbf70244a8f26b..8617b408eaa5c03c920f0207feaa12718d83e800 100644
--- a/attention_kindselective_n_heads2_seed1340/model_02500.pt
+++ b/attention_kindselective_n_heads2_seed1340/model_02500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db5e71e2651a7e4c04adf1b1bdd2fd67898b9838ff7ce2857cd735f4766b5532
+oid sha256:a43c3fea68b44f6e94a64fabd4bec96f4af33541fb79fc8ab291548d61ad3ba8
 size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_05000.pt b/attention_kindselective_n_heads2_seed1340/model_05000.pt
index 6a5728e2e1b189f44a9462ec4aed9dd4b14ad527..534f5ce06d11d6963eb798dc85425cb1b5f1d97d 100644
--- a/attention_kindselective_n_heads2_seed1340/model_05000.pt
+++ b/attention_kindselective_n_heads2_seed1340/model_05000.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87a3a4e1a720bf19b1c7f850b83a50505c837892b3964a6f17b36ac335b8a606
+oid sha256:cedd4c5d0d207b268fe96d32662f008ce85caf9422332a5f205480009a0bae8a
 size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_07500.pt b/attention_kindselective_n_heads2_seed1340/model_07500.pt
index b716248628433516e6300010672ae386bd9d8dfa..66ebfd3368d945ea47dfb35e4e2cb4f8fa1fa235 100644
--- a/attention_kindselective_n_heads2_seed1340/model_07500.pt
+++ b/attention_kindselective_n_heads2_seed1340/model_07500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:933638ce43638ef9bd5900408f7bbcb30a32be65bd1c3e240216a1265da81d33
+oid sha256:cf3ae491c351dd98c03ca4270a47dadf6b7299816fd4d4d28e612be080c8eaf6
 size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_10000.pt b/attention_kindselective_n_heads2_seed1340/model_10000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f5d0be82aafd14dada0797f701e6666b91eaa03c
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_10000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3917438735ab535c976c8cf747154e5ff137b3226377c17494de2db52a5dbc9
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_12500.pt b/attention_kindselective_n_heads2_seed1340/model_12500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e1a7eb2266e99c7b8c4ed79fbfc21587b956a9f4
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_12500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82f39c782dc0ecfa52b440f257e68f9fbd5d1c95cb1327b31ba1a4a90fddbd14
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_15000.pt b/attention_kindselective_n_heads2_seed1340/model_15000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a0b673ba6972275623da4056004c290a90c37588
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_15000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9a999d0f39f00dac906e79883b2c60cacea327110d17ce2aeb2ca063f6dd405
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_17500.pt b/attention_kindselective_n_heads2_seed1340/model_17500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f21606eafcd7ee65de929519901d9ecbf05c4585
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_17500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77730d4395b567ca5bcbb5ee491c47a6347ea335cdafbcaa57606e056f99d5be
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_20000.pt b/attention_kindselective_n_heads2_seed1340/model_20000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..b4dd22160671a19d595f7a51401bb98352f36acc
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_20000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1b0f211ab6773a5b61676a5550d5e97c97a33cb69ba778acf90934dbc1edd6b
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_22500.pt b/attention_kindselective_n_heads2_seed1340/model_22500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e1931a1089e622d35f7f338038a60eb4105ce4a6
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_22500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:704498bc3ab665573a16170d899809e4f5b2e0433ee9568c75b1a299e86e63d0
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_25000.pt b/attention_kindselective_n_heads2_seed1340/model_25000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..8e44cddaf203cafb2b7b34973b270ac7f068201e
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_25000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abad2d98112c3e440e66c5dbf182d0a97537a774c60116df1cc4e601a19b6d31
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_27500.pt b/attention_kindselective_n_heads2_seed1340/model_27500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..25562fc194bd00fed67c27c4b8f75181027ce733
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_27500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:231c011ec9ed5d97d25c7481c92c73d78b16474dea2f6e86258c8ef2055ce77f
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_30000.pt b/attention_kindselective_n_heads2_seed1340/model_30000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..8e6da3aa7a7d28b84c404a91b637a58465f39419
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_30000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c5fa0b4c0e1c96c7546e72b0608230b2b1be0609b274c5f83c59202f5463a12
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_32500.pt b/attention_kindselective_n_heads2_seed1340/model_32500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..307835a5183aa325552d286901b2d9c9a266d92c
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_32500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7db2e6312252b187d62e6ebdbc4915b0d541396bd5728579d3b5167316a4611
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_35000.pt b/attention_kindselective_n_heads2_seed1340/model_35000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..6e324026bfe83798b49d0c0d6f070465bc5ff463
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_35000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49399b8ac5d522d2716a77e73e6a22da180c585b1c865fdc6647f0186d199c0a
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_37500.pt b/attention_kindselective_n_heads2_seed1340/model_37500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..2abba94f607d621041bde491cffb6eb62378ad2e
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_37500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cee4edf4b83913b1bfc2b1150a9871f06b8205e075d2edd1d1f43cc015cd3ef
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_40000.pt b/attention_kindselective_n_heads2_seed1340/model_40000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..6de2d7103a58c63c65ab1530d4e29943c0b2ef95
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_40000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e12690a1fa5e37625d4caa7ed93f085e14857ea2fd92fde5d9c686eaadf988cd
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_42500.pt b/attention_kindselective_n_heads2_seed1340/model_42500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7b55aae57ae4e861546d7699642429a6b9cc64ad
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_42500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8fbfa201aad46b6a3b9db8b8a6e3ffe19b2b0bf7135c941a10352a0f659dbd2f
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_45000.pt b/attention_kindselective_n_heads2_seed1340/model_45000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..3b9bfdb758f96e75cf5506b553cee2844ccaf284
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_45000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dd2d44c2ef7a35bb4a5a6a4caae579ea0b5bbefab52fcc539b46efb5e230c44
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_47500.pt b/attention_kindselective_n_heads2_seed1340/model_47500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..06a66d3a3ac8ee1c4dd4f64ec85d1e1efb886d46
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_47500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:671e09d821aad0b05acfd58c4cb5943479619d40c185ec2b6e6081321e58589a
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/model_49999.pt b/attention_kindselective_n_heads2_seed1340/model_49999.pt
new file mode 100644
index 0000000000000000000000000000000000000000..810f732bb1702e0df2534b034e0a9dd8b3ba8cfa
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/model_49999.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ade59c1a7b1f27604ba9d522da69929f07674b7e48af67cb6f07699745ecf35
+size 38587970
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_02500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_02500.pt
index 035c55e1b50e39811bdd6a3852edc14545d28dec..3aa37668665ab051744b172656f49288d0e804e8 100644
--- a/attention_kindselective_n_heads2_seed1340/optimizer_02500.pt
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_02500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ad182381a14e875a3c04a9d0fce780b0641e040b738f8454b13df6255a740e34
+oid sha256:f8b2588354f55480e0234fb6a453af90f8dce086ada918b3cfb81acfd1ade067
 size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_05000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_05000.pt
index 667bea6c6237453d6d2f1d66f75747ef029a1f26..93ba36664ddacf6bbc231d55047909ff0a511313 100644
--- a/attention_kindselective_n_heads2_seed1340/optimizer_05000.pt
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_05000.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:89e51f2fe0afb3163bdb42b8d6c2d3968fb0c1d3a51b33ea7ae5061524b2b719
+oid sha256:a0fd46e04dbb546ef5f7f70b8792710a6b2c3d746613f6b100e414a89e996728
 size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_07500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_07500.pt
index 84af818a6b45e3df41354c297304dfff3c01917a..3d7b82171cc3c9567754aa300a42d0d1dded51cf 100644
--- a/attention_kindselective_n_heads2_seed1340/optimizer_07500.pt
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_07500.pt
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2c475776587ad7676fd824bc7c9c9570e94a3ffb32776966605d3ccf70784164
+oid sha256:06a62406b060c40f3fc2ef5b420b68e53741a74ba8388e186bc8ea3757ec7472
 size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_10000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_10000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..592cb6db43b371b2a1d56454da5a3775fd3ed0ba
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_10000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fca96eb06f8fcae983adc5b5fd192a2b465e885c3ce28a93a16dff13711aed2a
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_12500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_12500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e88626bc78c9ac9dd32fa225f24f68bc35733a8a
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_12500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:976e03738be53dea6d55c73ba3ec3c146d8423d50974c2d70bf10f0f5942be6a
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_15000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_15000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f3018d53784aba1d372484db42f40270fd0cda22
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_15000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccb9e25a713df0525fc8663aa13e51a23de1821fe91a03628167abe36ea38647
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_17500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_17500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..7634a54943f4f011c3c82902431864b85a260dc4
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_17500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f09d15b8a367be854baf5c8608688081f0710680a6364d6cfac81c9d751acf67
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_20000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_20000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..a0ee652acbf3a347d34dfbe4427bb3e7cf9a8ecb
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_20000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e238640eaa708aa5d306f56d91790321b5bfa8e6c0dc770b07dea34c1eb7de1f
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_22500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_22500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..84e6ab01d60ca18cc119a285a013a76d3bc63555
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_22500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b84fcade25645820cc712b0b50986685542b78af0f11c9d5779617e41883b28
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_25000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_25000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9f6a6c47fa6c94911221e15600ef85761c2dd254
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_25000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0345fb665f4a9124835ba1fa36906d37e6402ad0cc34129b35cf08f1f03a412
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_27500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_27500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..02f46a54167993958dcec795e3d5294c1b23bc62
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_27500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a7b48a02e56eaa60ecd353f6f6ca16dabbe3aca0844593838434ead44e59fa4
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_30000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_30000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..251420b3569496e38ffebfdea90d17f9009b363f
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_30000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b2deb47856a5a5c04ac084381911f4eb2ed9d4f2e122cf75eaa373d6a0d06d9
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_32500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_32500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9db9de0a936a8cc94cbcdc4b2f807195876caa43
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_32500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2758cd788b5b1aa7913b9d292af233fdef11fced987cca41dc5f9a8ebfdd73f6
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_35000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_35000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..e5358ea7ff52cddf598f473011d4eb2c9c1e3dc7
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_35000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:287a9f12979f3a2fe6db8b8ed14e540067d6f702530f08fae301c9ed10d640b0
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_37500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_37500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..f9aed3ba8951bd7e6de0a2fceffad98f76d2661d
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_37500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39099accc12524192b5e056c310a41a15cb2732a5494880dc53425c10f951945
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_40000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_40000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..21d171adaebf334ce24615a075c0b8f82b149188
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_40000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfa8c38e106b13173b0d03014522cc51af92bb1d830433a5e97e26aa89b43989
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_42500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_42500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..d8bbb71550a87f48ef7b4157cdcfc558e759541f
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_42500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1624e56abfb76aa3296c9513beeac34e72980c3261c49b8b9fff585341b61b64
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_45000.pt b/attention_kindselective_n_heads2_seed1340/optimizer_45000.pt
new file mode 100644
index 0000000000000000000000000000000000000000..9274490b3ec45c144e024765474854960787cab0
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_45000.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db558aca5edc7dc4484cfa7ed272ef293cac9a8a59c2e7ecb072dcf22b3c4fc1
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_47500.pt b/attention_kindselective_n_heads2_seed1340/optimizer_47500.pt
new file mode 100644
index 0000000000000000000000000000000000000000..bb737ad70de4bd04e828bb7e077d639d7996f34d
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_47500.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a33cd629f7b0ddcb5fe3a20bc9eeb53fd933901026e245136c3cc30eeeb0cf06
+size 70895430
diff --git a/attention_kindselective_n_heads2_seed1340/optimizer_49999.pt b/attention_kindselective_n_heads2_seed1340/optimizer_49999.pt
new file mode 100644
index 0000000000000000000000000000000000000000..99f19698ac2e6817ae276b499b50670157d25f66
--- /dev/null
+++ b/attention_kindselective_n_heads2_seed1340/optimizer_49999.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0b2e4ee681ee034716693975b2f3696f223c079ff54afce878108a7d45599c2
+size 70895430