diff --git a/attention_kindselective_n_heads4_seed1340/args.json b/attention_kindselective_n_heads4_seed1340/args.json index a87c86842010708e180c55ef14cd9ae27f332dba..32f27c1a2708e73351550f04ff619196a5478c5d 100644 --- a/attention_kindselective_n_heads4_seed1340/args.json +++ b/attention_kindselective_n_heads4_seed1340/args.json @@ -1 +1 @@ -{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads4_seed1340", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 4, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 50000, "warmup_steps": 200, "group": "wider_is_better_7", "use_wandb": true, "kill_self_after_run": false, "random_seed": 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": 7e-05, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "7e-5_10240_4_1340", "n_embd": 256} \ No newline at end of file +{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads4_seed1340", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 4, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 50000, "warmup_steps": 200, "group": "wider_is_better_7", "use_wandb": true, "kill_self_after_run": false, "random_seed": 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": 2e-05, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "2e-5_10240_4_1340", "n_embd": 256} \ No newline at end of file diff --git a/attention_kindselective_n_heads4_seed1340/dataloader_15000.pt b/attention_kindselective_n_heads4_seed1340/dataloader_15000.pt new file mode 100644 index 0000000000000000000000000000000000000000..0ea0f5a1bfab75667c4ebb0ca01b358cdc836a54 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_17500.pt b/attention_kindselective_n_heads4_seed1340/dataloader_17500.pt new file mode 100644 index 0000000000000000000000000000000000000000..eb9392348b1209c827e3e376b05eeda80e779aa8 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_20000.pt b/attention_kindselective_n_heads4_seed1340/dataloader_20000.pt new file mode 100644 index 0000000000000000000000000000000000000000..ba32fca059aec40e4f758de96bdac6df23b9d9f5 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_22500.pt b/attention_kindselective_n_heads4_seed1340/dataloader_22500.pt new file mode 100644 index 0000000000000000000000000000000000000000..498dc444f528d893090328a5bd1e2f37da46dc12 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_25000.pt b/attention_kindselective_n_heads4_seed1340/dataloader_25000.pt new file mode 100644 index 0000000000000000000000000000000000000000..b657bc134192d0ea956f984c289d0c682979a1f4 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_27500.pt b/attention_kindselective_n_heads4_seed1340/dataloader_27500.pt new file mode 100644 index 0000000000000000000000000000000000000000..d92d43d89390714f43db4f0782e49af0145b4a90 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_30000.pt b/attention_kindselective_n_heads4_seed1340/dataloader_30000.pt new file mode 100644 index 0000000000000000000000000000000000000000..06856f53253b6c8cbefd7d595d9b9b7266b22621 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_32500.pt b/attention_kindselective_n_heads4_seed1340/dataloader_32500.pt new file mode 100644 index 0000000000000000000000000000000000000000..dc0b4b40e57f41aa1046b3bd2697256635160c09 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_35000.pt b/attention_kindselective_n_heads4_seed1340/dataloader_35000.pt new file mode 100644 index 0000000000000000000000000000000000000000..62647b540fa7925361626b9f8dfa3959eebb7608 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_37500.pt b/attention_kindselective_n_heads4_seed1340/dataloader_37500.pt new file mode 100644 index 0000000000000000000000000000000000000000..475d8b538138a8e39b76a4cf04c8eaeac074d295 --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/dataloader_40000.pt b/attention_kindselective_n_heads4_seed1340/dataloader_40000.pt new file mode 100644 index 0000000000000000000000000000000000000000..50ec00ef6be330bbdb4cdf88e9a1097345da0d4d --- /dev/null +++ b/attention_kindselective_n_heads4_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_heads4_seed1340/log2.txt b/attention_kindselective_n_heads4_seed1340/log2.txt index f8f7d483410df33c09d779c3dcd5fe9d4cfcefcb..951c444b10974906effca27f7d91d3af187bd125 100644 --- a/attention_kindselective_n_heads4_seed1340/log2.txt +++ b/attention_kindselective_n_heads4_seed1340/log2.txt @@ -1,1113 +1,3231 @@ max_steps: 50000 -1100 val loss 7.3709 -1100 val perplexity 1589.1416 -1100 train 6.997883 (lr=9.9927e-05) (hash(x)=37984588) -40100 val loss 5.8536 -40100 val perplexity 348.4711 -40100 train 5.793257 (lr=9.2472e-06) (hash(x)=48035170) -1200 val loss 7.3289 -1200 val perplexity 1523.7263 -1200 train 7.504585 (lr=9.9910e-05) (hash(x)=56333817) +200 val loss 9.4488 +200 val perplexity 12693.1924 +200 train 9.438283 (lr=3.0000e-05) (hash(x)=50375849) +33000 val loss 5.8441 +33000 val perplexity 345.1911 +33000 train 5.711804 (lr=5.0236e-06) (hash(x)=46343736) 0 val loss 11.2703 0 val perplexity 78458.0234 -1300 val loss 7.3412 -1300 val perplexity 1542.5471 -1300 train 7.483638 (lr=9.9892e-05) (hash(x)=53454056) -40200 val loss 5.8481 -40200 val perplexity 346.5657 -40200 train 5.675584 (lr=9.1646e-06) (hash(x)=46651322) -1400 val loss 7.2560 -1400 val perplexity 1416.6073 -1400 train 7.362457 (lr=9.9871e-05) (hash(x)=55284163) -40300 val loss 5.8508 -40300 val perplexity 347.5022 -40300 train 5.707502 (lr=9.0827e-06) (hash(x)=46378099) -1500 val loss 7.3036 -1500 val perplexity 1485.6129 -1500 train 7.158344 (lr=9.9849e-05) (hash(x)=48162598) -0 train 11.272942 (lr=3.5000e-07) (hash(x)=57791809) -1600 val loss 7.2907 -1600 val perplexity 1466.6249 -1600 train 7.350899 (lr=9.9825e-05) (hash(x)=54214535) -40400 val loss 5.8443 -40400 val perplexity 345.2604 -40400 train 5.696771 (lr=9.0015e-06) (hash(x)=46495438) -100 val loss 9.7259 -100 val perplexity 16745.7324 -100 train 9.735422 (lr=3.5350e-05) (hash(x)=48211824) -1700 val loss 7.2683 -1700 val perplexity 1434.0831 -1700 train 7.474521 (lr=9.9799e-05) (hash(x)=53525003) -200 val loss 8.1014 -200 val perplexity 3298.9370 -200 train 8.058278 (lr=7.0000e-05) (hash(x)=50375849) -40500 val loss 5.8388 -40500 val perplexity 343.3772 -40500 train 5.996032 (lr=8.9211e-06) (hash(x)=52059147) -1800 val loss 7.2337 -1800 val perplexity 1385.3571 -1800 train 7.303545 (lr=9.9771e-05) (hash(x)=51848994) -300 val loss 7.6442 -300 val perplexity 2088.4370 -300 train 7.951274 (lr=6.9999e-05) (hash(x)=57250808) -400 val loss 7.5686 -400 val perplexity 1936.4075 -400 train 8.221130 (lr=6.9997e-05) (hash(x)=62519858) -40600 val loss 5.8407 -40600 val perplexity 344.0050 -40600 train 5.996709 (lr=8.8414e-06) (hash(x)=54885045) -1900 val loss 7.2157 -1900 val perplexity 1360.6837 -1900 train 7.053372 (lr=9.9741e-05) (hash(x)=48405987) -500 val loss 7.5147 -500 val perplexity 1834.7661 -500 train 7.395103 (lr=6.9994e-05) (hash(x)=47226806) -40700 val loss 5.8398 -40700 val perplexity 343.7109 -40700 train 6.230294 (lr=8.7624e-06) (hash(x)=53213971) -600 val loss 7.4733 -600 val perplexity 1760.4728 -600 train 7.511261 (lr=6.9990e-05) (hash(x)=51149322) -2000 val loss 7.1997 -2000 val perplexity 1338.9938 -2000 train 7.547995 (lr=9.9710e-05) (hash(x)=58592291) -700 val loss 7.4403 -700 val perplexity 1703.2408 -700 train 7.431039 (lr=6.9984e-05) (hash(x)=51564551) -40800 val loss 5.8359 -40800 val perplexity 342.3692 -40800 train 5.498599 (lr=8.6842e-06) (hash(x)=45133794) -2100 val loss 7.1929 -2100 val perplexity 1329.9009 -2100 train 7.272621 (lr=9.9677e-05) (hash(x)=51167081) -800 val loss 7.4083 -800 val perplexity 1649.6486 -800 train 7.160203 (lr=6.9977e-05) (hash(x)=45093459) -40900 val loss 5.8379 -40900 val perplexity 343.0574 -40900 train 5.988767 (lr=8.6068e-06) (hash(x)=56546547) -900 val loss 7.3864 -900 val perplexity 1613.9653 -900 train 7.685310 (lr=6.9969e-05) (hash(x)=54988361) -2200 val loss 7.1644 -2200 val perplexity 1292.5675 -2200 train 7.169672 (lr=9.9642e-05) (hash(x)=47994988) -1000 val loss 7.3699 -1000 val perplexity 1587.5004 -1000 train 7.238439 (lr=6.9960e-05) (hash(x)=47588648) -41000 val loss 5.8354 -41000 val perplexity 342.1882 -41000 train 6.221378 (lr=8.5301e-06) (hash(x)=49552270) -2300 val loss 7.1801 -2300 val perplexity 1313.0338 -2300 train 7.161005 (lr=9.9606e-05) (hash(x)=47377604) -1100 val loss 7.3479 -1100 val perplexity 1552.9474 -1100 train 6.955181 (lr=6.9949e-05) (hash(x)=37984588) -1200 val loss 7.3201 -1200 val perplexity 1510.3381 -1200 train 7.489902 (lr=6.9937e-05) (hash(x)=56333817) -41100 val loss 5.8354 -41100 val perplexity 342.2155 -41100 train 6.067726 (lr=8.4541e-06) (hash(x)=51222370) -2400 val loss 7.1777 -2400 val perplexity 1309.8896 -2400 train 7.223081 (lr=9.9567e-05) (hash(x)=53554323) -1300 val loss 7.3031 -1300 val perplexity 1484.8395 -1300 train 7.442085 (lr=6.9924e-05) (hash(x)=53454056) -41200 val loss 5.8369 -41200 val perplexity 342.7026 -41200 train 5.903095 (lr=8.3789e-06) (hash(x)=50883608) -2500 val loss 7.1743 -2500 val perplexity 1305.3995 -2500 train 7.199438 (lr=9.9527e-05) (hash(x)=50780417) -1400 val loss 7.2946 -1400 val perplexity 1472.2584 -1400 train 7.407144 (lr=6.9910e-05) (hash(x)=55284163) -1500 val loss 7.3019 -1500 val perplexity 1483.1653 -1500 train 7.143060 (lr=6.9894e-05) (hash(x)=48162598) -41300 val loss 5.8382 -41300 val perplexity 343.1662 -41300 train 6.058548 (lr=8.3045e-06) (hash(x)=52996828) -2600 val loss 7.1702 -2600 val perplexity 1300.0759 -2600 train 7.071150 (lr=9.9485e-05) (hash(x)=46453562) -1600 val loss 7.2579 -1600 val perplexity 1419.3342 -1600 train 7.321304 (lr=6.9877e-05) (hash(x)=54214535) -41400 val loss 5.8361 -41400 val perplexity 342.4566 -41400 train 5.811955 (lr=8.2308e-06) (hash(x)=48822716) -2700 val loss 7.2418 -2700 val perplexity 1396.6613 -2700 train 7.236431 (lr=9.9442e-05) (hash(x)=54404221) -1700 val loss 7.2608 -1700 val perplexity 1423.4137 -1700 train 7.450740 (lr=6.9859e-05) (hash(x)=53525003) -1800 val loss 7.2223 -1800 val perplexity 1369.6931 -1800 train 7.294693 (lr=6.9840e-05) (hash(x)=51848994) -41500 val loss 5.8375 -41500 val perplexity 342.9267 -41500 train 5.816633 (lr=8.1579e-06) (hash(x)=50936577) -2800 val loss 7.1606 -2800 val perplexity 1287.7007 -2800 train 7.852295 (lr=9.9396e-05) (hash(x)=59318895) -1900 val loss 7.2239 -1900 val perplexity 1371.8162 -1900 train 7.046358 (lr=6.9819e-05) (hash(x)=48405987) -41600 val loss 5.8377 -41600 val perplexity 343.0029 -41600 train 5.881047 (lr=8.0858e-06) (hash(x)=44375046) -2900 val loss 7.1619 -2900 val perplexity 1289.3486 -2900 train 7.149612 (lr=9.9349e-05) (hash(x)=47845760) -2000 val loss 7.2136 -2000 val perplexity 1357.7904 -2000 train 7.559543 (lr=6.9797e-05) (hash(x)=58592291) -2100 val loss 7.2009 -2100 val perplexity 1340.6804 -2100 train 7.274922 (lr=6.9774e-05) (hash(x)=51167081) -41700 val loss 5.8362 -41700 val perplexity 342.4806 -41700 train 5.599354 (lr=8.0144e-06) (hash(x)=44060021) -3000 val loss 7.1622 -3000 val perplexity 1289.7507 -3000 train 6.873025 (lr=9.9300e-05) (hash(x)=44336167) -2200 val loss 7.1883 -2200 val perplexity 1323.8024 -2200 train 7.174533 (lr=6.9750e-05) (hash(x)=47994988) -41800 val loss 5.8358 -41800 val perplexity 342.3502 -41800 train 6.128627 (lr=7.9438e-06) (hash(x)=57765221) -2300 val loss 7.2004 -2300 val perplexity 1340.0284 -3100 val loss 7.1277 -3100 val perplexity 1246.0415 -2300 train 7.185493 (lr=6.9724e-05) (hash(x)=47377604) -3100 train 7.390228 (lr=9.9249e-05) (hash(x)=44479330) -2400 val loss 7.1732 -2400 val perplexity 1303.9562 -2400 train 7.220889 (lr=6.9697e-05) (hash(x)=53554323) -3200 val loss 7.1646 -3200 val perplexity 1292.8153 -3200 train 7.241447 (lr=9.9197e-05) (hash(x)=54593096) -41900 val loss 5.8348 -41900 val perplexity 341.9940 -41900 train 5.597972 (lr=7.8740e-06) (hash(x)=46051470) -2500 val loss 7.1729 -2500 val perplexity 1303.6392 -2500 train 7.214406 (lr=6.9669e-05) (hash(x)=50780417) -3300 val loss 7.1778 -3300 val perplexity 1309.9883 -3300 train 7.088508 (lr=9.9142e-05) (hash(x)=45347643) -2600 val loss 7.1263 -2600 val perplexity 1244.2609 -2600 train 7.028562 (lr=6.9640e-05) (hash(x)=46453562) -42000 val loss 5.8415 -42000 val perplexity 344.2894 -42000 train 6.066402 (lr=7.8050e-06) (hash(x)=52077616) -2700 val loss 7.1115 -2700 val perplexity 1225.9312 -2700 train 7.080199 (lr=6.9609e-05) (hash(x)=54404221) -3400 val loss 7.1109 -3400 val perplexity 1225.2228 -3400 train 7.224405 (lr=9.9086e-05) (hash(x)=47797247) -42100 val loss 5.8336 -42100 val perplexity 341.5976 -42100 train 5.685880 (lr=7.7368e-06) (hash(x)=47845199) -2800 val loss 7.1089 -2800 val perplexity 1222.8175 -2800 train 7.750087 (lr=6.9577e-05) (hash(x)=59318895) -2900 val loss 7.1042 -2900 val perplexity 1217.0992 -2900 train 7.102781 (lr=6.9544e-05) (hash(x)=47845760) -3500 val loss 7.1129 -3500 val perplexity 1227.7335 -3500 train 7.017200 (lr=9.9028e-05) (hash(x)=46115683) -42200 val loss 5.8312 -42200 val perplexity 340.7710 -42200 train 6.088301 (lr=7.6693e-06) (hash(x)=51549823) -3000 val loss 7.0754 -3000 val perplexity 1182.5112 -3000 train 6.773137 (lr=6.9510e-05) (hash(x)=44336167) -3600 val loss 7.1055 -3600 val perplexity 1218.6689 -3600 train 6.937795 (lr=9.8969e-05) (hash(x)=44502074) -42300 val loss 5.8243 -42300 val perplexity 338.4359 -42300 train 5.965167 (lr=7.6027e-06) (hash(x)=56922131) -3100 val loss 7.0511 -3100 val perplexity 1154.1840 -3100 train 7.331172 (lr=6.9474e-05) (hash(x)=44479330) -3200 val loss 7.0365 -3200 val perplexity 1137.3547 -3200 train 7.123116 (lr=6.9438e-05) (hash(x)=54593096) -3700 val loss 7.0926 -3700 val perplexity 1202.9988 -3700 train 7.232045 (lr=9.8908e-05) (hash(x)=55388443) -42400 val loss 5.8262 -42400 val perplexity 339.0629 -42400 train 5.787869 (lr=7.5368e-06) (hash(x)=49004372) -3300 val loss 7.0396 -3300 val perplexity 1140.9235 -3300 train 6.961339 (lr=6.9400e-05) (hash(x)=45347643) -3800 val loss 7.0881 -3800 val perplexity 1197.5861 -3800 train 6.871030 (lr=9.8845e-05) (hash(x)=43790341) -42500 val loss 5.8224 -42500 val perplexity 337.7739 -42500 train 5.810077 (lr=7.4717e-06) (hash(x)=50651839) -3400 val loss 7.0278 -3400 val perplexity 1127.5500 -3400 train 7.154087 (lr=6.9360e-05) (hash(x)=47797247) -3500 val loss 7.0127 -3500 val perplexity 1110.6456 -3500 train 6.888947 (lr=6.9320e-05) (hash(x)=46115683) -3900 val loss 7.0927 -3900 val perplexity 1203.1331 -3900 train 7.089813 (lr=9.8780e-05) (hash(x)=50013318) -42600 val loss 5.8226 -42600 val perplexity 337.8404 -42600 train 6.146226 (lr=7.4074e-06) (hash(x)=50767721) -3600 val loss 6.9927 -3600 val perplexity 1088.6067 -3600 train 6.841239 (lr=6.9278e-05) (hash(x)=44502074) -4000 val loss 7.0888 -4000 val perplexity 1198.4447 -4000 train 7.173865 (lr=9.8713e-05) (hash(x)=51704787) -42700 val loss 5.8274 -42700 val perplexity 339.4667 -42700 train 5.559310 (lr=7.3440e-06) (hash(x)=49099183) -3700 val loss 6.9963 -3700 val perplexity 1092.6323 -3700 train 7.130051 (lr=6.9235e-05) (hash(x)=55388443) -3800 val loss 6.9835 -3800 val perplexity 1078.6748 -3800 train 6.767923 (lr=6.9191e-05) (hash(x)=43790341) -4100 val loss 7.0791 -4100 val perplexity 1186.8921 -4100 train 7.136350 (lr=9.8645e-05) (hash(x)=50821964) -42800 val loss 5.8205 -42800 val perplexity 337.1293 -42800 train 6.236982 (lr=7.2813e-06) (hash(x)=42272413) -3900 val loss 6.9793 -3900 val perplexity 1074.1350 -3900 train 6.971794 (lr=6.9146e-05) (hash(x)=50013318) -4200 val loss 7.1331 -4200 val perplexity 1252.7737 -4200 train 7.145490 (lr=9.8575e-05) (hash(x)=49675080) -42900 val loss 5.8239 -42900 val perplexity 338.2816 -42900 train 5.616502 (lr=7.2194e-06) (hash(x)=48582863) -4000 val loss 6.9700 -4000 val perplexity 1064.2661 -4000 train 7.038657 (lr=6.9099e-05) (hash(x)=51704787) -4100 val loss 6.9672 -4100 val perplexity 1061.2119 -4100 train 7.042075 (lr=6.9051e-05) (hash(x)=50821964) -4300 val loss 7.1120 -4300 val perplexity 1226.5977 -4300 train 6.766655 (lr=9.8503e-05) (hash(x)=43239281) -43000 val loss 5.8217 -43000 val perplexity 337.5590 -43000 train 5.796019 (lr=7.1583e-06) (hash(x)=48703446) -4200 val loss 6.9870 -4200 val perplexity 1082.4434 -4200 train 6.998270 (lr=6.9002e-05) (hash(x)=49675080) -4400 val loss 7.1208 -4400 val perplexity 1237.4618 -4400 train 6.820117 (lr=9.8430e-05) (hash(x)=45076737) -4300 val loss 6.9632 -4300 val perplexity 1056.9971 -4300 train 6.617874 (lr=6.8952e-05) (hash(x)=43239281) -43100 val loss 5.8203 -43100 val perplexity 337.0722 -43100 train 5.743421 (lr=7.0981e-06) (hash(x)=48730321) -4400 val loss 6.9505 -4400 val perplexity 1043.6844 -4400 train 6.636121 (lr=6.8901e-05) (hash(x)=45076737) -4500 val loss 7.1026 -4500 val perplexity 1215.1410 -4500 train 7.188966 (lr=9.8355e-05) (hash(x)=57930262) -43200 val loss 5.8184 -43200 val perplexity 336.4272 -43200 train 6.053661 (lr=7.0386e-06) (hash(x)=56536090) -4500 val loss 6.9386 -4500 val perplexity 1031.2871 -4500 train 7.046677 (lr=6.8848e-05) (hash(x)=57930262) -4600 val loss 7.0943 -4600 val perplexity 1205.0604 -4600 train 6.878264 (lr=9.8278e-05) (hash(x)=46721614) -4600 val loss 6.9329 -4600 val perplexity 1025.4572 -4600 train 6.712116 (lr=6.8794e-05) (hash(x)=46721614) -43300 val loss 5.8176 -43300 val perplexity 336.1527 -43300 train 5.962167 (lr=6.9800e-06) (hash(x)=54154116) -4700 val loss 6.9091 -4700 val perplexity 1001.3547 -4700 train 6.706449 (lr=6.8739e-05) (hash(x)=49837920) -4700 val loss 7.0600 -4700 val perplexity 1164.4540 -4700 train 6.883050 (lr=9.8199e-05) (hash(x)=49837920) -4800 val loss 6.8954 -4800 val perplexity 987.7471 -4800 train 7.025270 (lr=6.8683e-05) (hash(x)=48380045) -43400 val loss 5.8258 -43400 val perplexity 338.9396 -43400 train 5.723001 (lr=6.9222e-06) (hash(x)=50058055) -4800 val loss 7.0530 -4800 val perplexity 1156.3136 -4900 val loss 6.8825 -4900 val perplexity 975.0323 -4800 train 7.163837 (lr=9.8119e-05) (hash(x)=48380045) -4900 train 6.764085 (lr=6.8626e-05) (hash(x)=44202577) -43500 val loss 5.8194 -43500 val perplexity 336.7863 -43500 train 5.749609 (lr=6.8652e-06) (hash(x)=48743802) -5000 val loss 6.8525 -5000 val perplexity 946.2589 -5000 train 6.858996 (lr=6.8567e-05) (hash(x)=52038024) -4900 val loss 7.0500 -4900 val perplexity 1152.8193 -4900 train 6.937270 (lr=9.8036e-05) (hash(x)=44202577) -5100 val loss 6.8317 -5100 val perplexity 926.8068 -5100 train 7.012638 (lr=6.8507e-05) (hash(x)=53700038) -43600 val loss 5.8188 -43600 val perplexity 336.5771 -43600 train 5.458486 (lr=6.8090e-06) (hash(x)=42792886) -5200 val loss 6.8274 -5200 val perplexity 922.8022 -5200 train 6.814481 (lr=6.8446e-05) (hash(x)=48137625) -5000 val loss 7.0187 -5000 val perplexity 1117.3700 -5000 train 7.044736 (lr=9.7953e-05) (hash(x)=52038024) -43700 val loss 5.8163 -43700 val perplexity 335.7397 -43700 train 6.229331 (lr=6.7537e-06) (hash(x)=56446070) -5300 val loss 6.8135 -5300 val perplexity 910.0942 -5300 train 6.670180 (lr=6.8384e-05) (hash(x)=43161573) -5100 val loss 7.0031 -5100 val perplexity 1100.0636 -5100 train 7.189824 (lr=9.7867e-05) (hash(x)=53700038) -5400 val loss 6.8088 -5400 val perplexity 905.7951 -5400 train 7.007874 (lr=6.8320e-05) (hash(x)=56673322) -43800 val loss 5.8178 -43800 val perplexity 336.2392 -43800 train 5.653908 (lr=6.6992e-06) (hash(x)=45584354) -5500 val loss 6.8175 -5500 val perplexity 913.7433 -5500 train 6.971771 (lr=6.8256e-05) (hash(x)=53468295) -5200 val loss 6.9942 -5200 val perplexity 1090.2700 -5200 train 6.960814 (lr=9.7780e-05) (hash(x)=48137625) -43900 val loss 5.8161 -43900 val perplexity 335.6672 -43900 train 5.664525 (lr=6.6455e-06) (hash(x)=49339253) -5600 val loss 6.7675 -5600 val perplexity 869.1533 -5600 train 7.060785 (lr=6.8190e-05) (hash(x)=59287280) -5300 val loss 6.9828 -5300 val perplexity 1077.8917 -5300 train 6.854216 (lr=9.7691e-05) (hash(x)=43161573) -5700 val loss 6.7728 -5700 val perplexity 873.7137 -5700 train 7.033890 (lr=6.8123e-05) (hash(x)=57575806) -44000 val loss 5.8158 -44000 val perplexity 335.5459 -44000 train 5.635589 (lr=6.5926e-06) (hash(x)=46183203) -5800 val loss 6.7736 -5800 val perplexity 874.4873 -5800 train 6.723006 (lr=6.8055e-05) (hash(x)=46897279) -5400 val loss 6.9617 -5400 val perplexity 1055.4479 -5400 train 7.139781 (lr=9.7600e-05) (hash(x)=56673322) -44100 val loss 5.8135 -44100 val perplexity 334.7792 -44100 train 5.802389 (lr=6.5406e-06) (hash(x)=47849630) -5900 val loss 6.7703 -5900 val perplexity 871.5712 -5900 train 6.636249 (lr=6.7985e-05) (hash(x)=47565679) -5500 val loss 6.9375 -5500 val perplexity 1030.2126 -5500 train 7.112161 (lr=9.7508e-05) (hash(x)=53468295) -6000 val loss 6.7804 -6000 val perplexity 880.4198 -6000 train 6.648599 (lr=6.7915e-05) (hash(x)=51590090) -44200 val loss 5.8127 -44200 val perplexity 334.5341 -44200 train 6.146719 (lr=6.4894e-06) (hash(x)=49834275) -6100 val loss 6.7618 -6100 val perplexity 864.2370 -6100 train 7.137344 (lr=6.7843e-05) (hash(x)=59732271) -5600 val loss 6.9085 -5600 val perplexity 1000.7457 -5600 train 7.224924 (lr=9.7414e-05) (hash(x)=59287280) -44300 val loss 5.8095 -44300 val perplexity 333.4532 -44300 train 6.254009 (lr=6.4390e-06) (hash(x)=62535257) -6200 val loss 6.7642 -6200 val perplexity 866.2368 -6200 train 6.741118 (lr=6.7770e-05) (hash(x)=46394422) -5700 val loss 6.9043 -5700 val perplexity 996.5828 -5700 train 7.165204 (lr=9.7318e-05) (hash(x)=57575806) -6300 val loss 6.7599 -6300 val perplexity 862.5737 -6300 train 6.768298 (lr=6.7696e-05) (hash(x)=53748145) -44400 val loss 5.8078 -44400 val perplexity 332.8983 -44400 train 5.866293 (lr=6.3895e-06) (hash(x)=49253957) -6400 val loss 6.7398 -6400 val perplexity 845.4134 -6400 train 6.611226 (lr=6.7621e-05) (hash(x)=46054751) -5800 val loss 6.8937 -5800 val perplexity 986.0107 -5800 train 6.843620 (lr=9.7221e-05) (hash(x)=46897279) -6500 val loss 6.7447 -6500 val perplexity 849.5231 -6500 train 7.008954 (lr=6.7545e-05) (hash(x)=51816809) -44500 val loss 5.8086 -44500 val perplexity 333.1573 -44500 train 5.787062 (lr=6.3408e-06) (hash(x)=55368339) -5900 val loss 6.8839 -5900 val perplexity 976.4327 -5900 train 6.751657 (lr=9.7122e-05) (hash(x)=47565679) -6600 val loss 6.7294 -6600 val perplexity 836.6632 -6600 train 6.619945 (lr=6.7467e-05) (hash(x)=52453336) -44600 val loss 5.8085 -44600 val perplexity 333.1111 -44600 train 5.920068 (lr=6.2929e-06) (hash(x)=47098476) -6700 val loss 6.7275 -6700 val perplexity 835.0275 -6700 train 6.752491 (lr=6.7389e-05) (hash(x)=49108775) -6000 val loss 6.8816 -6000 val perplexity 974.1851 -6000 train 6.755291 (lr=9.7021e-05) (hash(x)=51590090) -6800 val loss 6.7240 -6800 val perplexity 832.1128 -6800 train 6.615003 (lr=6.7309e-05) (hash(x)=46745396) -44700 val loss 5.8119 -44700 val perplexity 334.2415 -44700 train 5.744551 (lr=6.2459e-06) (hash(x)=48280562) -6100 val loss 6.8433 -6100 val perplexity 937.6213 -6100 train 7.231963 (lr=9.6919e-05) (hash(x)=59732271) -6900 val loss 6.7129 -6900 val perplexity 822.9876 -6900 train 6.874073 (lr=6.7228e-05) (hash(x)=46534986) -44800 val loss 5.8111 -44800 val perplexity 333.9803 -44800 train 6.081460 (lr=6.1998e-06) (hash(x)=55591638) -7000 val loss 6.7105 -7000 val perplexity 821.0043 -7000 train 7.102075 (lr=6.7146e-05) (hash(x)=49317888) -6200 val loss 6.8340 -6200 val perplexity 928.8552 -6200 train 6.807478 (lr=9.6815e-05) (hash(x)=46394422) -7100 val loss 6.7149 -7100 val perplexity 824.5883 -7100 train 6.716461 (lr=6.7063e-05) (hash(x)=50360484) -44900 val loss 5.8106 -44900 val perplexity 333.8185 -44900 train 5.963002 (lr=6.1545e-06) (hash(x)=53757748) -6300 val loss 6.8275 -6300 val perplexity 922.9268 -6300 train 6.854854 (lr=9.6709e-05) (hash(x)=53748145) -7200 val loss 6.7289 -7200 val perplexity 836.1858 -7200 train 6.611360 (lr=6.6978e-05) (hash(x)=49515094) -45000 val loss 5.8083 -45000 val perplexity 333.0545 -45000 train 5.965748 (lr=6.1100e-06) (hash(x)=51685087) -7300 val loss 6.7329 -7300 val perplexity 839.6050 -7300 train 6.724388 (lr=6.6893e-05) (hash(x)=51546861) -6400 val loss 6.8335 -6400 val perplexity 928.4501 -6400 train 6.675262 (lr=9.6602e-05) (hash(x)=46054751) -7400 val loss 6.7554 -7400 val perplexity 858.6415 -7400 train 6.747338 (lr=6.6806e-05) (hash(x)=48320948) -45100 val loss 5.8066 -45100 val perplexity 332.4754 -45100 train 5.804672 (lr=6.0664e-06) (hash(x)=50093774) -6500 val loss 6.8217 -6500 val perplexity 917.5349 -6500 train 7.011214 (lr=9.6493e-05) (hash(x)=51816809) -7500 val loss 6.7581 -7500 val perplexity 861.0138 -7500 train 6.601406 (lr=6.6718e-05) (hash(x)=40167457) -45200 val loss 5.8079 -45200 val perplexity 332.9253 -45200 train 5.531459 (lr=6.0237e-06) (hash(x)=43460450) -7600 val loss 6.7487 -7600 val perplexity 852.9794 -7600 train 6.694093 (lr=6.6630e-05) (hash(x)=49942165) -6600 val loss 6.8168 -6600 val perplexity 913.0578 -6600 train 6.709125 (lr=9.6382e-05) (hash(x)=52453336) -7700 val loss 6.7365 -7700 val perplexity 842.6336 -7700 train 6.467313 (lr=6.6540e-05) (hash(x)=48853311) -45300 val loss 5.8075 -45300 val perplexity 332.7967 -45300 train 6.013202 (lr=5.9818e-06) (hash(x)=49935488) -6700 val loss 6.7775 -6700 val perplexity 877.8773 -6700 train 6.801557 (lr=9.6270e-05) (hash(x)=49108775) -7800 val loss 6.6977 -7800 val perplexity 810.5665 -7800 train 6.587886 (lr=6.6448e-05) (hash(x)=48510117) -7900 val loss 6.6830 -7900 val perplexity 798.6885 -7900 train 6.674704 (lr=6.6356e-05) (hash(x)=48339781) -45400 val loss 5.8084 -45400 val perplexity 333.0822 -45400 train 5.796069 (lr=5.9407e-06) (hash(x)=49447929) -6800 val loss 6.7620 -6800 val perplexity 864.3433 -6800 train 6.641977 (lr=9.6156e-05) (hash(x)=46745396) -8000 val loss 6.6692 -8000 val perplexity 787.8011 -8000 train 6.791812 (lr=6.6263e-05) (hash(x)=54927320) -45500 val loss 5.8073 -45500 val perplexity 332.7321 -45500 train 5.758606 (lr=5.9005e-06) (hash(x)=50713904) -8100 val loss 6.6685 -8100 val perplexity 787.1943 -8100 train 6.383808 (lr=6.6169e-05) (hash(x)=46461786) -6900 val loss 6.7623 -6900 val perplexity 864.6273 -6900 train 6.878799 (lr=9.6040e-05) (hash(x)=46534986) -8200 val loss 6.6542 -8200 val perplexity 776.0621 -8200 train 6.567985 (lr=6.6073e-05) (hash(x)=51536260) -45600 val loss 5.8281 -45600 val perplexity 339.7057 -45600 train 5.737741 (lr=5.8612e-06) (hash(x)=47674606) -7000 val loss 6.7450 -7000 val perplexity 849.7970 -7000 train 7.090025 (lr=9.5923e-05) (hash(x)=49317888) -8300 val loss 6.6445 -8300 val perplexity 768.5745 -8300 train 6.385070 (lr=6.5976e-05) (hash(x)=44770722) -45700 val loss 5.8069 -45700 val perplexity 332.6016 -45700 train 5.718114 (lr=5.8227e-06) (hash(x)=51539617) -8400 val loss 6.6291 -8400 val perplexity 756.7707 -8400 train 6.630779 (lr=6.5879e-05) (hash(x)=50104957) -7100 val loss 6.7150 -7100 val perplexity 824.7158 -7100 train 6.732314 (lr=9.5804e-05) (hash(x)=50360484) -8500 val loss 6.6186 -8500 val perplexity 748.8944 -8500 train 6.730096 (lr=6.5780e-05) (hash(x)=50132971) -45800 val loss 5.8064 -45800 val perplexity 332.4121 -45800 train 5.777891 (lr=5.7851e-06) (hash(x)=44448785) -7200 val loss 6.7163 -7200 val perplexity 825.7302 -7200 train 6.610737 (lr=9.5683e-05) (hash(x)=49515094) -8600 val loss 6.6052 -8600 val perplexity 738.9441 -8600 train 6.559481 (lr=6.5680e-05) (hash(x)=52193699) -45900 val loss 5.8067 -45900 val perplexity 332.5286 -45900 train 5.633348 (lr=5.7484e-06) (hash(x)=51499105) -8700 val loss 6.5946 -8700 val perplexity 731.1027 -8700 train 6.601634 (lr=6.5579e-05) (hash(x)=47902319) -7300 val loss 6.7274 -7300 val perplexity 834.9750 -7300 train 6.727911 (lr=9.5561e-05) (hash(x)=51546861) -8800 val loss 6.5905 -8800 val perplexity 728.1722 -8800 train 6.874655 (lr=6.5477e-05) (hash(x)=54904230) -46000 val loss 5.8019 -46000 val perplexity 330.9422 -46000 train 5.717524 (lr=5.7125e-06) (hash(x)=48359464) -8900 val loss 6.5658 -8900 val perplexity 710.3594 -7400 val loss 6.7125 -7400 val perplexity 822.6318 -8900 train 6.456809 (lr=6.5374e-05) (hash(x)=46311615) -7400 train 6.713262 (lr=9.5437e-05) (hash(x)=48320948) -46100 val loss 5.8034 -46100 val perplexity 331.4209 -46100 train 5.856997 (lr=5.6775e-06) (hash(x)=51885986) -9000 val loss 6.5762 -9000 val perplexity 717.8407 -9000 train 6.467047 (lr=6.5270e-05) (hash(x)=48535188) -7500 val loss 6.7149 -7500 val perplexity 824.6104 -7500 train 6.552715 (lr=9.5312e-05) (hash(x)=40167457) -9100 val loss 6.5482 -9100 val perplexity 697.9618 -9100 train 6.596004 (lr=6.5164e-05) (hash(x)=51757372) -46200 val loss 5.8014 -46200 val perplexity 330.7549 -46200 train 6.350838 (lr=5.6434e-06) (hash(x)=65186615) -9200 val loss 6.5645 -9200 val perplexity 709.4841 -9200 train 6.451322 (lr=6.5058e-05) (hash(x)=51131708) -7600 val loss 6.6876 -7600 val perplexity 802.4150 -7600 train 6.612953 (lr=9.5185e-05) (hash(x)=49942165) -9300 val loss 6.5492 -9300 val perplexity 698.7104 -9300 train 6.486413 (lr=6.4951e-05) (hash(x)=44784276) -46300 val loss 5.8012 -46300 val perplexity 330.6895 -46300 train 5.776049 (lr=5.6101e-06) (hash(x)=49626999) -7700 val loss 6.6817 -7700 val perplexity 797.6483 -7700 train 6.419989 (lr=9.5057e-05) (hash(x)=48853311) -9400 val loss 6.5449 -9400 val perplexity 695.6748 -9400 train 6.656165 (lr=6.4842e-05) (hash(x)=51981169) -46400 val loss 5.7988 -46400 val perplexity 329.9163 -46400 train 5.685073 (lr=5.5777e-06) (hash(x)=43325701) -9500 val loss 6.5243 -9500 val perplexity 681.5016 -9500 train 6.455208 (lr=6.4733e-05) (hash(x)=47232936) -7800 val loss 6.7101 -7800 val perplexity 820.6340 -7800 train 6.605823 (lr=9.4926e-05) (hash(x)=48510117) -9600 val loss 6.5283 -9600 val perplexity 684.2013 -9600 train 6.555980 (lr=6.4622e-05) (hash(x)=53800450) -46500 val loss 5.8014 -46500 val perplexity 330.7682 -46500 train 6.095033 (lr=5.5462e-06) (hash(x)=54028595) -7900 val loss 6.6982 -7900 val perplexity 810.9125 -7900 train 6.697021 (lr=9.4795e-05) (hash(x)=48339781) -9700 val loss 6.5110 -9700 val perplexity 672.4990 -9700 train 6.611804 (lr=6.4511e-05) (hash(x)=55768123) -46600 val loss 5.8004 -46600 val perplexity 330.4477 -46600 train 5.646887 (lr=5.5156e-06) (hash(x)=44519175) -9800 val loss 6.5201 -9800 val perplexity 678.6589 -9800 train 6.437235 (lr=6.4398e-05) (hash(x)=47745177) -8000 val loss 6.7015 -8000 val perplexity 813.6622 -8000 train 6.827817 (lr=9.4661e-05) (hash(x)=54927320) -46700 val loss 5.8020 -46700 val perplexity 330.9630 -46700 train 5.859873 (lr=5.4858e-06) (hash(x)=48357998) -9900 val loss 6.5059 -9900 val perplexity 669.1097 -9900 train 6.771001 (lr=6.4284e-05) (hash(x)=56592246) -8100 val loss 6.7125 -8100 val perplexity 822.6165 -8100 train 6.440390 (lr=9.4526e-05) (hash(x)=46461786) -10000 val loss 6.5050 -10000 val perplexity 668.4611 -10000 train 6.518945 (lr=6.4170e-05) (hash(x)=51655963) -46800 val loss 5.8007 -46800 val perplexity 330.5271 -46800 train 5.972873 (lr=5.4569e-06) (hash(x)=55911353) -10100 val loss 6.5228 -10100 val perplexity 680.4901 -10100 train 6.419588 (lr=6.4054e-05) (hash(x)=49809511) -8200 val loss 6.6683 -8200 val perplexity 787.0919 -8200 train 6.587510 (lr=9.4390e-05) (hash(x)=51536260) -46900 val loss 5.8022 -46900 val perplexity 331.0330 -46900 train 5.741434 (lr=5.4289e-06) (hash(x)=47897187) -10200 val loss 6.5020 -10200 val perplexity 666.4471 -10200 train 6.091907 (lr=6.3937e-05) (hash(x)=42297812) -8300 val loss 6.7038 -8300 val perplexity 815.4675 -8300 train 6.438375 (lr=9.4252e-05) (hash(x)=44770722) -10300 val loss 6.4973 -10300 val perplexity 663.3268 -10300 train 6.167974 (lr=6.3820e-05) (hash(x)=55529820) -47000 val loss 5.7987 -47000 val perplexity 329.8616 -47000 train 5.359689 (lr=5.4017e-06) (hash(x)=43196571) -10400 val loss 6.5084 -10400 val perplexity 670.7539 -10400 train 6.523131 (lr=6.3701e-05) (hash(x)=53255684) -8400 val loss 6.6674 -8400 val perplexity 786.3558 -8400 train 6.660437 (lr=9.4112e-05) (hash(x)=50104957) -10500 val loss 6.4665 -10500 val perplexity 643.2431 -10500 train 6.590051 (lr=6.3581e-05) (hash(x)=54306191) -47100 val loss 5.7992 -47100 val perplexity 330.0356 -47100 train 5.657062 (lr=5.3755e-06) (hash(x)=51224987) -8500 val loss 6.6906 -8500 val perplexity 804.8203 -8500 train 6.818645 (lr=9.3971e-05) (hash(x)=50132971) -10600 val loss 6.4734 -10600 val perplexity 647.6887 -10600 train 6.714907 (lr=6.3460e-05) (hash(x)=60130567) -47200 val loss 5.8016 -47200 val perplexity 330.8251 -47200 train 5.508888 (lr=5.3501e-06) (hash(x)=47943697) -10700 val loss 6.4934 -10700 val perplexity 660.7467 -10700 train 6.420184 (lr=6.3339e-05) (hash(x)=50074737) -8600 val loss 6.6548 -8600 val perplexity 776.4918 -8600 train 6.595765 (lr=9.3828e-05) (hash(x)=52193699) -10800 val loss 6.4459 -10800 val perplexity 630.1255 -10800 train 6.556093 (lr=6.3216e-05) (hash(x)=51547220) -47300 val loss 5.8012 -47300 val perplexity 330.6942 -47300 train 5.713035 (lr=5.3256e-06) (hash(x)=47351003) -10900 val loss 6.4832 -10900 val perplexity 654.0450 -10900 train 6.565746 (lr=6.3092e-05) (hash(x)=55943981) -8700 val loss 6.6440 -8700 val perplexity 768.1542 -8700 train 6.659471 (lr=9.3684e-05) (hash(x)=47902319) -47400 val loss 5.8019 -47400 val perplexity 330.9364 -47400 train 5.875019 (lr=5.3020e-06) (hash(x)=55562243) -11000 val loss 6.4434 -11000 val perplexity 628.5189 -11000 train 6.483435 (lr=6.2968e-05) (hash(x)=46444570) -8800 val loss 6.6390 -8800 val perplexity 764.3397 -8800 train 6.929801 (lr=9.3538e-05) (hash(x)=54904230) -11100 val loss 6.4500 -11100 val perplexity 632.7070 -11100 train 6.459686 (lr=6.2842e-05) (hash(x)=49589063) -47500 val loss 5.8001 -47500 val perplexity 330.3456 -47500 train 5.675170 (lr=5.2792e-06) (hash(x)=53544850) -11200 val loss 6.4506 -11200 val perplexity 633.0538 -11200 train 6.457446 (lr=6.2715e-05) (hash(x)=51392283) -8900 val loss 6.6381 -8900 val perplexity 763.6228 -8900 train 6.531604 (lr=9.3391e-05) (hash(x)=46311615) -47600 val loss 5.8024 -47600 val perplexity 331.0871 -47600 train 5.985342 (lr=5.2574e-06) (hash(x)=43634907) -11300 val loss 6.4364 -11300 val perplexity 624.1540 -11300 train 6.397132 (lr=6.2588e-05) (hash(x)=45081133) -9000 val loss 6.6386 -9000 val perplexity 764.0009 -9000 train 6.519309 (lr=9.3242e-05) (hash(x)=48535188) -11400 val loss 6.4289 -11400 val perplexity 619.4864 -11400 train 6.524812 (lr=6.2459e-05) (hash(x)=53700397) -47700 val loss 5.7999 -47700 val perplexity 330.2570 -47700 train 5.776142 (lr=5.2364e-06) (hash(x)=47909383) -11500 val loss 6.4207 -11500 val perplexity 614.4413 -11500 train 6.142918 (lr=6.2330e-05) (hash(x)=43839088) -9100 val loss 6.6102 -9100 val perplexity 742.6683 -9100 train 6.649930 (lr=9.3092e-05) (hash(x)=51757372) -47800 val loss 5.7953 -47800 val perplexity 328.7346 -47800 train 5.674577 (lr=5.2163e-06) (hash(x)=45871079) -11600 val loss 6.4324 -11600 val perplexity 621.6376 -11600 train 6.465709 (lr=6.2199e-05) (hash(x)=48088111) -9200 val loss 6.6127 -9200 val perplexity 744.4720 -9200 train 6.500169 (lr=9.2940e-05) (hash(x)=51131708) -11700 val loss 6.4299 -11700 val perplexity 620.0857 -11700 train 6.745090 (lr=6.2068e-05) (hash(x)=55108226) -47900 val loss 5.7951 -47900 val perplexity 328.6861 -47900 train 5.784526 (lr=5.1972e-06) (hash(x)=47333324) -11800 val loss 6.4176 -11800 val perplexity 612.5077 -11800 train 6.684676 (lr=6.1936e-05) (hash(x)=58524839) -9300 val loss 6.5997 -9300 val perplexity 734.8544 -9300 train 6.554953 (lr=9.2786e-05) (hash(x)=44784276) -48000 val loss 5.7948 -48000 val perplexity 328.5829 -48000 train 5.876556 (lr=5.1788e-06) (hash(x)=52758020) -11900 val loss 6.4150 -11900 val perplexity 610.9667 -11900 train 6.081831 (lr=6.1802e-05) (hash(x)=43864078) -9400 val loss 6.5917 -9400 val perplexity 729.0008 -9400 train 6.715466 (lr=9.2632e-05) (hash(x)=51981169) -12000 val loss 6.4093 -12000 val perplexity 607.4685 -12000 train 6.095185 (lr=6.1668e-05) (hash(x)=43448544) -48100 val loss 5.7937 -48100 val perplexity 328.2226 -48100 train 5.802242 (lr=5.1614e-06) (hash(x)=49806349) -12100 val loss 6.3978 -12100 val perplexity 600.5402 -12100 train 6.396197 (lr=6.1533e-05) (hash(x)=55200399) -9500 val loss 6.6014 -9500 val perplexity 736.1489 -9500 train 6.523407 (lr=9.2475e-05) (hash(x)=47232936) -12200 val loss 6.4067 -12200 val perplexity 605.8679 -48200 val loss 5.7914 -48200 val perplexity 327.4593 -12200 train 6.546592 (lr=6.1397e-05) (hash(x)=57627314) -48200 train 5.868467 (lr=5.1449e-06) (hash(x)=53220839) -9600 val loss 6.5844 -9600 val perplexity 723.6813 -9600 train 6.593096 (lr=9.2317e-05) (hash(x)=53800450) -12300 val loss 6.4145 -12300 val perplexity 610.6370 -12300 train 6.693074 (lr=6.1260e-05) (hash(x)=53617087) -48300 val loss 5.7940 -48300 val perplexity 328.3305 -48300 train 5.822805 (lr=5.1293e-06) (hash(x)=56052541) -12400 val loss 6.4177 -12400 val perplexity 612.5649 -12400 train 6.262563 (lr=6.1122e-05) (hash(x)=51135678) -9700 val loss 6.5799 -9700 val perplexity 720.4314 -9700 train 6.675194 (lr=9.2158e-05) (hash(x)=55768123) -48400 val loss 5.7905 -48400 val perplexity 327.1818 -48400 train 5.670732 (lr=5.1145e-06) (hash(x)=44482356) -12500 val loss 6.4237 -12500 val perplexity 616.2940 -12500 train 6.428000 (lr=6.0984e-05) (hash(x)=48025130) -12600 val loss 6.4127 -12600 val perplexity 609.5201 -9800 val loss 6.5909 -9800 val perplexity 728.4636 -12600 train 6.470107 (lr=6.0844e-05) (hash(x)=52135695) -9800 train 6.517842 (lr=9.1997e-05) (hash(x)=47745177) -48500 val loss 5.7897 -48500 val perplexity 326.9112 -48500 train 5.520173 (lr=5.1007e-06) (hash(x)=45714818) -12700 val loss 6.4382 -12700 val perplexity 625.2830 -12700 train 6.303355 (lr=6.0703e-05) (hash(x)=51888613) -9900 val loss 6.5974 -9900 val perplexity 733.2065 -9900 train 6.867185 (lr=9.1835e-05) (hash(x)=56592246) -48600 val loss 5.7913 -48600 val perplexity 327.4302 -48600 train 5.555061 (lr=5.0877e-06) (hash(x)=49476556) -12800 val loss 6.4182 -12800 val perplexity 612.8937 -12800 train 6.351943 (lr=6.0562e-05) (hash(x)=50418818) -12900 val loss 6.4053 -12900 val perplexity 605.0304 -12900 train 7.257672 (lr=6.0420e-05) (hash(x)=58649585) -10000 val loss 6.6087 -10000 val perplexity 741.5094 -10000 train 6.601171 (lr=9.1671e-05) (hash(x)=51655963) -48700 val loss 5.7916 -48700 val perplexity 327.5210 -48700 train 5.483535 (lr=5.0756e-06) (hash(x)=42508579) -13000 val loss 6.3998 -13000 val perplexity 601.7260 -13000 train 6.663928 (lr=6.0277e-05) (hash(x)=54567307) -10100 val loss 6.5917 -10100 val perplexity 728.9938 -10100 train 6.492918 (lr=9.1506e-05) (hash(x)=49809511) -13100 val loss 6.3887 -13100 val perplexity 595.0941 -13100 train 6.565659 (lr=6.0133e-05) (hash(x)=52071473) -48800 val loss 5.7920 -48800 val perplexity 327.6678 -48800 train 6.004168 (lr=5.0644e-06) (hash(x)=52737449) -13200 val loss 6.3823 -13200 val perplexity 591.2615 -13200 train 6.148626 (lr=5.9988e-05) (hash(x)=46293092) -10200 val loss 6.5790 -10200 val perplexity 719.8549 -10200 train 6.172993 (lr=9.1339e-05) (hash(x)=42297812) -48900 val loss 5.7912 -48900 val perplexity 327.4165 -48900 train 5.636753 (lr=5.0542e-06) (hash(x)=47057569) -13300 val loss 6.3882 -13300 val perplexity 594.8053 -13300 train 6.513732 (lr=5.9842e-05) (hash(x)=56511467) -10300 val loss 6.5718 -10300 val perplexity 714.6194 -10300 train 6.288669 (lr=9.1171e-05) (hash(x)=55529820) -13400 val loss 6.3714 -13400 val perplexity 584.8615 -13400 train 6.486885 (lr=5.9695e-05) (hash(x)=54753763) -49000 val loss 5.7940 -49000 val perplexity 328.3136 -49000 train 5.770644 (lr=5.0448e-06) (hash(x)=49908975) -13500 val loss 6.3630 -13500 val perplexity 579.9861 -13500 train 6.510321 (lr=5.9548e-05) (hash(x)=53610247) -10400 val loss 6.5765 -10400 val perplexity 717.9875 -10400 train 6.559584 (lr=9.1001e-05) (hash(x)=53255684) -49100 val loss 5.7920 -49100 val perplexity 327.6564 -49100 train 5.614700 (lr=5.0363e-06) (hash(x)=48427414) -13600 val loss 6.3478 -13600 val perplexity 571.2138 -13600 train 6.225275 (lr=5.9400e-05) (hash(x)=47526249) -10500 val loss 6.5377 -10500 val perplexity 690.7025 -10500 train 6.650892 (lr=9.0830e-05) (hash(x)=54306191) -13700 val loss 6.3612 -13700 val perplexity 578.9458 -13700 train 6.217256 (lr=5.9251e-05) (hash(x)=51185517) -49200 val loss 5.7897 -49200 val perplexity 326.9223 -49200 train 5.618961 (lr=5.0286e-06) (hash(x)=50246074) -13800 val loss 6.3604 -13800 val perplexity 578.4911 -13800 train 6.032686 (lr=5.9101e-05) (hash(x)=45953529) -10600 val loss 6.5441 -10600 val perplexity 695.1240 -10600 train 6.779467 (lr=9.0658e-05) (hash(x)=60130567) -49300 val loss 5.7902 -49300 val perplexity 327.0729 -49300 train 5.844873 (lr=5.0219e-06) (hash(x)=47715359) -13900 val loss 6.3799 -13900 val perplexity 589.8560 -13900 train 6.307417 (lr=5.8950e-05) (hash(x)=47238157) -14000 val loss 6.3578 -14000 val perplexity 577.0012 -14000 train 6.338602 (lr=5.8799e-05) (hash(x)=54250750) -10700 val loss 6.5191 -10700 val perplexity 677.9451 -10700 train 6.443255 (lr=9.0484e-05) (hash(x)=50074737) -49400 val loss 5.7951 -49400 val perplexity 328.7013 -49400 train 5.789733 (lr=5.0161e-06) (hash(x)=50175867) -14100 val loss 6.3644 -14100 val perplexity 580.7674 -14100 train 6.351201 (lr=5.8646e-05) (hash(x)=48198552) -10800 val loss 6.5320 -10800 val perplexity 686.7475 -10800 train 6.645676 (lr=9.0308e-05) (hash(x)=51547220) -14200 val loss 6.3626 -14200 val perplexity 579.7433 -49500 val loss 5.7927 -49500 val perplexity 327.8968 -14200 train 6.497223 (lr=5.8493e-05) (hash(x)=52020690) -49500 train 5.631345 (lr=5.0112e-06) (hash(x)=49336040) -14300 val loss 6.3541 -14300 val perplexity 574.8251 -14300 train 6.156728 (lr=5.8339e-05) (hash(x)=45165483) -10900 val loss 6.5470 -10900 val perplexity 697.1835 -10900 train 6.629424 (lr=9.0132e-05) (hash(x)=55943981) -49600 val loss 5.7945 -49600 val perplexity 328.5037 -49600 train 5.834286 (lr=5.0072e-06) (hash(x)=52039357) -14400 val loss 6.3595 -14400 val perplexity 577.9749 -14400 train 6.475995 (lr=5.8184e-05) (hash(x)=52184072) -11000 val loss 6.5333 -11000 val perplexity 687.6558 -11000 train 6.565971 (lr=8.9954e-05) (hash(x)=46444570) -14500 val loss 6.3595 -14500 val perplexity 577.9520 -14500 train 6.706707 (lr=5.8029e-05) (hash(x)=56046436) -49700 val loss 5.7948 -49700 val perplexity 328.5858 -49700 train 5.590436 (lr=5.0040e-06) (hash(x)=47568707) -14600 val loss 6.3534 -14600 val perplexity 574.4607 -14600 train 6.230230 (lr=5.7872e-05) (hash(x)=52029694) -11100 val loss 6.5385 -11100 val perplexity 691.2264 -11100 train 6.547019 (lr=8.9774e-05) (hash(x)=49589063) -49800 val loss 5.7962 -49800 val perplexity 329.0321 -49800 train 5.719766 (lr=5.0018e-06) (hash(x)=48451274) -14700 val loss 6.3699 -14700 val perplexity 584.0244 -14700 train 6.211578 (lr=5.7715e-05) (hash(x)=50258224) -11200 val loss 6.5242 -11200 val perplexity 681.4029 -11200 train 6.523566 (lr=8.9593e-05) (hash(x)=51392283) -14800 val loss 6.3676 -14800 val perplexity 582.6866 -14800 train 5.692614 (lr=5.7558e-05) (hash(x)=42112262) -49900 val loss 5.7984 -49900 val perplexity 329.7762 -49900 train 5.486496 (lr=5.0004e-06) (hash(x)=44523603) -14900 val loss 6.3608 -14900 val perplexity 578.7267 -14900 train 6.101720 (lr=5.7399e-05) (hash(x)=47219933) -11300 val loss 6.5394 -11300 val perplexity 691.8585 -11300 train 6.495648 (lr=8.9411e-05) (hash(x)=45081133) -49999 val loss 5.7909 -49999 val perplexity 327.3206 -15000 val loss 6.4009 -15000 val perplexity 602.3866 -15000 train 6.002666 (lr=5.7240e-05) (hash(x)=58309309) -11400 val loss 6.5171 -11400 val perplexity 676.6274 -11400 train 6.611968 (lr=8.9227e-05) (hash(x)=53700397) -15100 val loss 6.3498 -15100 val perplexity 572.3857 -15100 train 6.254695 (lr=5.7079e-05) (hash(x)=48756049) -15200 val loss 6.3577 -15200 val perplexity 576.9346 -15200 train 6.047930 (lr=5.6919e-05) (hash(x)=49791737) -11500 val loss 6.5199 -11500 val perplexity 678.5172 -11500 train 6.260948 (lr=8.9043e-05) (hash(x)=43839088) -15300 val loss 6.3336 -15300 val perplexity 563.2054 -15300 train 6.302131 (lr=5.6757e-05) (hash(x)=53084126) -11600 val loss 6.5099 -11600 val perplexity 671.7664 -11600 train 6.543172 (lr=8.8856e-05) (hash(x)=48088111) -15400 val loss 6.3310 -15400 val perplexity 561.6938 -15400 train 6.422942 (lr=5.6595e-05) (hash(x)=55041679) -15500 val loss 6.3303 -15500 val perplexity 561.3511 -15500 train 6.124517 (lr=5.6432e-05) (hash(x)=43249867) -11700 val loss 6.5147 -11700 val perplexity 674.9705 -11700 train 6.829165 (lr=8.8668e-05) (hash(x)=55108226) -15600 val loss 6.3211 -15600 val perplexity 556.1926 -15600 train 6.312618 (lr=5.6268e-05) (hash(x)=49006517) -15700 val loss 6.3168 -15700 val perplexity 553.8163 -15700 train 6.645555 (lr=5.6104e-05) (hash(x)=60986839) -11800 val loss 6.5320 -11800 val perplexity 686.7426 -11800 train 6.764431 (lr=8.8479e-05) (hash(x)=58524839) -15800 val loss 6.3371 -15800 val perplexity 565.1663 -15800 train 6.246183 (lr=5.5938e-05) (hash(x)=48354906) -11900 val loss 6.5044 -11900 val perplexity 668.0739 -11900 train 6.170547 (lr=8.8289e-05) (hash(x)=43864078) -15900 val loss 6.3422 -15900 val perplexity 568.0605 -15900 train 6.207072 (lr=5.5773e-05) (hash(x)=52679780) -16000 val loss 6.3104 -16000 val perplexity 550.2414 -16000 train 6.475496 (lr=5.5606e-05) (hash(x)=58049587) -12000 val loss 6.5055 -12000 val perplexity 668.7924 -12000 train 6.222349 (lr=8.8097e-05) (hash(x)=43448544) +300 val loss 8.6686 +300 val perplexity 5817.5054 +300 train 8.811249 (lr=3.0000e-05) (hash(x)=57250808) +33100 val loss 5.8459 +33100 val perplexity 345.8036 +33100 train 5.538730 (lr=4.9862e-06) (hash(x)=45424347) +400 val loss 8.0384 +400 val perplexity 3097.6018 +400 train 8.543970 (lr=2.9999e-05) (hash(x)=62519858) +33200 val loss 5.8466 +33200 val perplexity 346.0492 +33200 train 5.920062 (lr=4.9490e-06) (hash(x)=50207861) +0 train 11.272942 (lr=1.0000e-07) (hash(x)=57791809) +500 val loss 7.7802 +500 val perplexity 2392.6609 +500 train 7.703458 (lr=2.9998e-05) (hash(x)=47226806) +33300 val loss 5.8450 +33300 val perplexity 345.4918 +33300 train 5.760426 (lr=4.9119e-06) (hash(x)=45847922) +100 val loss 10.1555 +100 val perplexity 25732.9141 +100 train 10.156497 (lr=1.0100e-05) (hash(x)=48211824) +200 val loss 9.6899 +200 val perplexity 16153.4971 +200 train 9.682636 (lr=2.0000e-05) (hash(x)=50375849) +600 val loss 7.6837 +600 val perplexity 2172.7268 +600 train 7.721937 (lr=2.9996e-05) (hash(x)=51149322) +33400 val loss 5.8434 +33400 val perplexity 344.9362 +33400 train 5.692331 (lr=4.8750e-06) (hash(x)=44723617) +300 val loss 9.1789 +300 val perplexity 9690.8086 +300 train 9.290282 (lr=2.0000e-05) (hash(x)=57250808) +400 val loss 8.6497 +400 val perplexity 5708.3970 +400 train 9.058089 (lr=1.9999e-05) (hash(x)=62519858) +33500 val loss 5.8414 +33500 val perplexity 344.2636 +33500 train 5.946756 (lr=4.8382e-06) (hash(x)=54058855) +700 val loss 7.6160 +700 val perplexity 2030.3364 +700 train 7.619907 (lr=2.9993e-05) (hash(x)=51564551) +500 val loss 8.2863 +500 val perplexity 3969.3169 +500 train 8.230556 (lr=1.9998e-05) (hash(x)=47226806) +33600 val loss 5.8350 +33600 val perplexity 342.0563 +33600 train 5.830905 (lr=4.8015e-06) (hash(x)=53052866) +800 val loss 7.5799 +800 val perplexity 1958.3860 +800 train 7.364426 (lr=2.9990e-05) (hash(x)=45093459) +600 val loss 8.0756 +600 val perplexity 3215.1313 +600 train 8.109509 (lr=1.9997e-05) (hash(x)=51149322) +700 val loss 7.9214 +700 val perplexity 2755.6343 +700 train 7.899722 (lr=1.9996e-05) (hash(x)=51564551) +33700 val loss 5.8361 +33700 val perplexity 342.4419 +33700 train 5.771416 (lr=4.7650e-06) (hash(x)=50458295) +900 val loss 7.5575 +900 val perplexity 1915.1384 +900 train 7.854511 (lr=2.9987e-05) (hash(x)=54988361) +800 val loss 7.8110 +800 val perplexity 2467.5693 +800 train 7.656053 (lr=1.9994e-05) (hash(x)=45093459) +33800 val loss 5.8396 +33800 val perplexity 343.6276 +33800 train 6.001922 (lr=4.7286e-06) (hash(x)=58204533) +900 val loss 7.7303 +900 val perplexity 2276.1753 +900 train 7.996741 (lr=1.9991e-05) (hash(x)=54988361) +1000 val loss 7.5298 +1000 val perplexity 1862.8225 +1000 train 7.405553 (lr=2.9983e-05) (hash(x)=47588648) +1000 val loss 7.6711 +1000 val perplexity 2145.5459 +1000 train 7.529509 (lr=1.9989e-05) (hash(x)=47588648) +33900 val loss 5.8331 +33900 val perplexity 341.4266 +33900 train 5.973702 (lr=4.6923e-06) (hash(x)=52246445) +1100 val loss 7.5124 +1100 val perplexity 1830.6561 +1100 train 7.129356 (lr=2.9978e-05) (hash(x)=37984588) +1100 val loss 7.6313 +1100 val perplexity 2061.7136 +1100 train 7.311824 (lr=1.9985e-05) (hash(x)=37984588) +34000 val loss 5.8314 +34000 val perplexity 340.8391 +34000 train 5.773145 (lr=4.6562e-06) (hash(x)=49316883) +1200 val loss 7.6030 +1200 val perplexity 2004.1549 +1200 train 7.762239 (lr=1.9982e-05) (hash(x)=56333817) +1200 val loss 7.4873 +1200 val perplexity 1785.2194 +1200 train 7.662006 (lr=2.9973e-05) (hash(x)=56333817) +1300 val loss 7.5799 +1300 val perplexity 1958.3683 +1300 train 7.718001 (lr=1.9978e-05) (hash(x)=53454056) +34100 val loss 5.8307 +34100 val perplexity 340.5963 +34100 train 5.476385 (lr=4.6202e-06) (hash(x)=42991129) +1300 val loss 7.4784 +1300 val perplexity 1769.3606 +1300 train 7.629707 (lr=2.9968e-05) (hash(x)=53454056) +1400 val loss 7.5505 +1400 val perplexity 1901.6605 +1400 train 7.637989 (lr=1.9974e-05) (hash(x)=55284163) +34200 val loss 5.8298 +34200 val perplexity 340.2947 +34200 train 5.460163 (lr=4.5844e-06) (hash(x)=45199853) +1500 val loss 7.5291 +1500 val perplexity 1861.4205 +1500 train 7.397888 (lr=1.9970e-05) (hash(x)=48162598) +1400 val loss 7.4401 +1400 val perplexity 1702.9590 +1400 train 7.535004 (lr=2.9961e-05) (hash(x)=55284163) +1600 val loss 7.5044 +1600 val perplexity 1816.0538 +1600 train 7.550734 (lr=1.9965e-05) (hash(x)=54214535) +34300 val loss 5.8327 +34300 val perplexity 341.2793 +34300 train 5.719717 (lr=4.5487e-06) (hash(x)=48404386) +1700 val loss 7.4875 +1700 val perplexity 1785.4952 +1700 train 7.715155 (lr=1.9960e-05) (hash(x)=53525003) +1500 val loss 7.4127 +1500 val perplexity 1656.9574 +1500 train 7.241348 (lr=2.9955e-05) (hash(x)=48162598) +1800 val loss 7.4678 +1800 val perplexity 1750.7803 +1800 train 7.550333 (lr=1.9954e-05) (hash(x)=51848994) +34400 val loss 5.8297 +34400 val perplexity 340.2561 +34400 train 5.632356 (lr=4.5132e-06) (hash(x)=45614997) +1600 val loss 7.3951 +1600 val perplexity 1628.0189 +1600 train 7.461652 (lr=2.9947e-05) (hash(x)=54214535) +1900 val loss 7.4561 +1900 val perplexity 1730.3905 +1900 train 7.264217 (lr=1.9948e-05) (hash(x)=48405987) +34500 val loss 5.8305 +34500 val perplexity 340.5349 +34500 train 5.872010 (lr=4.4778e-06) (hash(x)=53157684) +2000 val loss 7.4352 +2000 val perplexity 1694.5630 +2000 train 7.796566 (lr=1.9942e-05) (hash(x)=58592291) +1700 val loss 7.3813 +1700 val perplexity 1605.7224 +1700 train 7.606174 (lr=2.9940e-05) (hash(x)=53525003) +2100 val loss 7.4240 +2100 val perplexity 1675.7378 +2100 train 7.526247 (lr=1.9935e-05) (hash(x)=51167081) +34600 val loss 5.8340 +34600 val perplexity 341.7344 +34600 train 6.056926 (lr=4.4426e-06) (hash(x)=57152287) +1800 val loss 7.3526 +1800 val perplexity 1560.2057 +1800 train 7.430679 (lr=2.9931e-05) (hash(x)=51848994) +2200 val loss 7.4054 +2200 val perplexity 1644.8823 +2200 train 7.406776 (lr=1.9928e-05) (hash(x)=47994988) +34700 val loss 5.8303 +34700 val perplexity 340.4497 +34700 train 5.276961 (lr=4.4075e-06) (hash(x)=46088213) +2300 val loss 7.3923 +2300 val perplexity 1623.3632 +2300 train 7.365356 (lr=1.9921e-05) (hash(x)=47377604) +1900 val loss 7.3380 +1900 val perplexity 1537.6453 +1900 train 7.149599 (lr=2.9922e-05) (hash(x)=48405987) +2400 val loss 7.3787 +2400 val perplexity 1601.5023 +2400 train 7.435731 (lr=1.9913e-05) (hash(x)=53554323) +34800 val loss 5.8292 +34800 val perplexity 340.0785 +34800 train 5.568765 (lr=4.3725e-06) (hash(x)=47098930) +2000 val loss 7.3143 +2000 val perplexity 1501.6104 +2000 train 7.660333 (lr=2.9913e-05) (hash(x)=58592291) +2500 val loss 7.3715 +2500 val perplexity 1589.9791 +2500 train 7.390099 (lr=1.9905e-05) (hash(x)=50780417) +34900 val loss 5.8313 +34900 val perplexity 340.8132 +34900 train 5.860092 (lr=4.3378e-06) (hash(x)=51033119) +2600 val loss 7.3618 +2600 val perplexity 1574.7303 +2600 train 7.284196 (lr=1.9897e-05) (hash(x)=46453562) +2100 val loss 7.3109 +2100 val perplexity 1496.5294 +2100 train 7.405928 (lr=2.9903e-05) (hash(x)=51167081) +2700 val loss 7.3551 +2700 val perplexity 1564.2192 +2700 train 7.365285 (lr=1.9888e-05) (hash(x)=54404221) +35000 val loss 5.8286 +35000 val perplexity 339.8669 +35000 train 5.496274 (lr=4.3031e-06) (hash(x)=47111881) +2800 val loss 7.3490 +2800 val perplexity 1554.6641 +2800 train 8.134493 (lr=1.9879e-05) (hash(x)=59318895) +2200 val loss 7.2868 +2200 val perplexity 1460.8304 +2200 train 7.274078 (lr=2.9893e-05) (hash(x)=47994988) +35100 val loss 5.8280 +35100 val perplexity 339.6801 +35100 train 5.868427 (lr=4.2687e-06) (hash(x)=52899306) +2900 val loss 7.3397 +2900 val perplexity 1540.2253 +2900 train 7.320337 (lr=1.9870e-05) (hash(x)=47845760) +2300 val loss 7.2803 +2300 val perplexity 1451.3661 +2300 train 7.264668 (lr=2.9882e-05) (hash(x)=47377604) +3000 val loss 7.3285 +3000 val perplexity 1523.0798 +3000 train 7.034776 (lr=1.9860e-05) (hash(x)=44336167) +35200 val loss 5.8254 +35200 val perplexity 338.7896 +35200 train 5.771688 (lr=4.2344e-06) (hash(x)=49124834) +3100 val loss 7.3222 +3100 val perplexity 1513.5110 +3100 train 7.609055 (lr=1.9850e-05) (hash(x)=44479330) +2400 val loss 7.2494 +2400 val perplexity 1407.2020 +2400 train 7.294912 (lr=2.9870e-05) (hash(x)=53554323) +35300 val loss 5.8276 +35300 val perplexity 339.5558 +35300 train 5.974386 (lr=4.2002e-06) (hash(x)=54127282) +3200 val loss 7.3135 +3200 val perplexity 1500.4652 +3200 train 7.381842 (lr=1.9839e-05) (hash(x)=54593096) +2500 val loss 7.2397 +2500 val perplexity 1393.6290 +2500 train 7.271302 (lr=2.9858e-05) (hash(x)=50780417) +3300 val loss 7.3023 +3300 val perplexity 1483.6639 +3300 train 7.198156 (lr=1.9828e-05) (hash(x)=45347643) +35400 val loss 5.8248 +35400 val perplexity 338.5892 +35400 train 5.699027 (lr=4.1662e-06) (hash(x)=51154858) +3400 val loss 7.2806 +3400 val perplexity 1451.8507 +3400 train 7.417652 (lr=1.9817e-05) (hash(x)=47797247) +2600 val loss 7.2269 +2600 val perplexity 1375.9912 +2600 train 7.140221 (lr=2.9846e-05) (hash(x)=46453562) +35500 val loss 5.8204 +35500 val perplexity 337.0939 +35500 train 5.878022 (lr=4.1324e-06) (hash(x)=51585534) +3500 val loss 7.2721 +3500 val perplexity 1439.6101 +3500 train 7.162816 (lr=1.9806e-05) (hash(x)=46115683) +2700 val loss 7.2100 +2700 val perplexity 1352.9432 +2700 train 7.207460 (lr=2.9832e-05) (hash(x)=54404221) +3600 val loss 7.2564 +3600 val perplexity 1417.1553 +3600 train 7.097079 (lr=1.9794e-05) (hash(x)=44502074) +35600 val loss 5.8185 +35600 val perplexity 336.4744 +35600 train 5.610509 (lr=4.0988e-06) (hash(x)=46228695) +3700 val loss 7.2436 +3700 val perplexity 1399.0702 +3700 train 7.389739 (lr=1.9782e-05) (hash(x)=55388443) +2800 val loss 7.1962 +2800 val perplexity 1334.3281 +2800 train 7.947353 (lr=2.9819e-05) (hash(x)=59318895) +35700 val loss 5.8184 +35700 val perplexity 336.4351 +35700 train 5.795092 (lr=4.0653e-06) (hash(x)=49080481) +3800 val loss 7.2379 +3800 val perplexity 1391.1320 +3800 train 7.018186 (lr=1.9769e-05) (hash(x)=43790341) +2900 val loss 7.1879 +2900 val perplexity 1323.3826 +2900 train 7.176562 (lr=2.9805e-05) (hash(x)=47845760) +3900 val loss 7.2271 +3900 val perplexity 1376.2458 +3900 train 7.203982 (lr=1.9756e-05) (hash(x)=50013318) +35800 val loss 5.8166 +35800 val perplexity 335.8297 +35800 train 5.977071 (lr=4.0319e-06) (hash(x)=56743084) +4000 val loss 7.2128 +4000 val perplexity 1356.7031 +4000 train 7.283560 (lr=1.9743e-05) (hash(x)=51704787) +3000 val loss 7.1691 +3000 val perplexity 1298.7339 +3000 train 6.876387 (lr=2.9790e-05) (hash(x)=44336167) +35900 val loss 5.8175 +35900 val perplexity 336.1277 +35900 train 5.935955 (lr=3.9988e-06) (hash(x)=53068229) +4100 val loss 7.2006 +4100 val perplexity 1340.2764 +4100 train 7.256206 (lr=1.9729e-05) (hash(x)=50821964) +4200 val loss 7.1969 +4200 val perplexity 1335.2313 +4200 train 7.215014 (lr=1.9715e-05) (hash(x)=49675080) +3100 val loss 7.1660 +3100 val perplexity 1294.5950 +3100 train 7.463683 (lr=2.9775e-05) (hash(x)=44479330) +36000 val loss 5.8160 +36000 val perplexity 335.6121 +36000 train 5.995482 (lr=3.9658e-06) (hash(x)=54363049) +4300 val loss 7.1760 +4300 val perplexity 1307.6105 +4300 train 6.828897 (lr=1.9701e-05) (hash(x)=43239281) +3200 val loss 7.1636 +3200 val perplexity 1291.5645 +3200 train 7.230880 (lr=2.9759e-05) (hash(x)=54593096) +4400 val loss 7.1704 +4400 val perplexity 1300.3605 +4400 train 6.848389 (lr=1.9686e-05) (hash(x)=45076737) +36100 val loss 5.8180 +36100 val perplexity 336.2846 +36100 train 5.735805 (lr=3.9330e-06) (hash(x)=50555767) +4500 val loss 7.1575 +4500 val perplexity 1283.7329 +4500 train 7.241147 (lr=1.9671e-05) (hash(x)=57930262) +3300 val loss 7.1228 +3300 val perplexity 1239.9019 +3300 train 7.037287 (lr=2.9743e-05) (hash(x)=45347643) +36200 val loss 5.8150 +36200 val perplexity 335.2841 +36200 train 5.976042 (lr=3.9003e-06) (hash(x)=52103710) +4600 val loss 7.1467 +4600 val perplexity 1269.9493 +4600 train 6.930983 (lr=1.9656e-05) (hash(x)=46721614) +3400 val loss 7.0961 +3400 val perplexity 1207.2466 +3400 train 7.252370 (lr=2.9726e-05) (hash(x)=47797247) +4700 val loss 7.1317 +4700 val perplexity 1250.9512 +4700 train 6.944176 (lr=1.9640e-05) (hash(x)=49837920) +36300 val loss 5.8144 +36300 val perplexity 335.0915 +36300 train 5.603258 (lr=3.8678e-06) (hash(x)=43347374) +4800 val loss 7.1179 +4800 val perplexity 1233.8871 +4800 train 7.237657 (lr=1.9624e-05) (hash(x)=48380045) +3500 val loss 7.0749 +3500 val perplexity 1181.8708 +3500 train 6.955238 (lr=2.9709e-05) (hash(x)=46115683) +36400 val loss 5.8118 +36400 val perplexity 334.2235 +36400 train 5.872618 (lr=3.8355e-06) (hash(x)=52661268) +4900 val loss 7.1065 +4900 val perplexity 1219.8602 +4900 train 6.969951 (lr=1.9607e-05) (hash(x)=44202577) +3600 val loss 7.0600 +3600 val perplexity 1164.4590 +3600 train 6.914934 (lr=2.9691e-05) (hash(x)=44502074) +5000 val loss 7.0871 +5000 val perplexity 1196.4263 +5000 train 7.121143 (lr=1.9591e-05) (hash(x)=52038024) +36500 val loss 5.8145 +36500 val perplexity 335.1328 +36500 train 5.718210 (lr=3.8034e-06) (hash(x)=47765383) +5100 val loss 7.0726 +5100 val perplexity 1179.1919 +5100 train 7.260921 (lr=1.9573e-05) (hash(x)=53700038) +3700 val loss 7.0414 +3700 val perplexity 1142.9490 +3700 train 7.190182 (lr=2.9672e-05) (hash(x)=55388443) +36600 val loss 5.8136 +36600 val perplexity 334.8369 +36600 train 5.661536 (lr=3.7715e-06) (hash(x)=49302411) +5200 val loss 7.0641 +5200 val perplexity 1169.2306 +5200 train 7.102770 (lr=1.9556e-05) (hash(x)=48137625) +3800 val loss 7.0332 +3800 val perplexity 1133.6306 +5300 val loss 7.0548 +5300 val perplexity 1158.3484 +5300 train 6.934961 (lr=1.9538e-05) (hash(x)=43161573) +3800 train 6.808227 (lr=2.9653e-05) (hash(x)=43790341) +36700 val loss 5.8184 +36700 val perplexity 336.4338 +36700 train 5.359552 (lr=3.7397e-06) (hash(x)=44113066) +5400 val loss 7.0451 +5400 val perplexity 1147.1997 +5400 train 7.218289 (lr=1.9520e-05) (hash(x)=56673322) +3900 val loss 7.0231 +3900 val perplexity 1122.2799 +3900 train 7.005631 (lr=2.9634e-05) (hash(x)=50013318) +5500 val loss 7.0312 +5500 val perplexity 1131.3496 +5500 train 7.210893 (lr=1.9502e-05) (hash(x)=53468295) +36800 val loss 5.8139 +36800 val perplexity 334.9363 +36800 train 5.685965 (lr=3.7081e-06) (hash(x)=55896927) +5600 val loss 7.0204 +5600 val perplexity 1119.2513 +5600 train 7.352832 (lr=1.9483e-05) (hash(x)=59287280) +4000 val loss 7.0029 +4000 val perplexity 1099.8507 +4000 train 7.073129 (lr=2.9614e-05) (hash(x)=51704787) +36900 val loss 5.8147 +36900 val perplexity 335.1872 +36900 train 5.603320 (lr=3.6767e-06) (hash(x)=49812554) +5700 val loss 7.0072 +5700 val perplexity 1104.5830 +5700 train 7.265706 (lr=1.9464e-05) (hash(x)=57575806) +4100 val loss 6.9945 +4100 val perplexity 1090.6444 +4100 train 7.052852 (lr=2.9593e-05) (hash(x)=50821964) +5800 val loss 7.0005 +5800 val perplexity 1097.1431 +5800 train 6.920443 (lr=1.9444e-05) (hash(x)=46897279) +37000 val loss 5.8132 +37000 val perplexity 334.6940 +37000 train 5.648577 (lr=3.6455e-06) (hash(x)=49391771) +5900 val loss 6.9976 +5900 val perplexity 1094.0217 +5900 train 6.839458 (lr=1.9424e-05) (hash(x)=47565679) +4200 val loss 6.9858 +4200 val perplexity 1081.1940 +4200 train 7.026465 (lr=2.9572e-05) (hash(x)=49675080) +37100 val loss 5.8089 +37100 val perplexity 333.2560 +6000 val loss 6.9768 +6000 val perplexity 1071.4912 +6000 train 6.860244 (lr=1.9404e-05) (hash(x)=51590090) +37100 train 5.749371 (lr=3.6144e-06) (hash(x)=50515328) +6100 val loss 6.9704 +6100 val perplexity 1064.6779 +6100 train 7.354523 (lr=1.9384e-05) (hash(x)=59732271) +4300 val loss 6.9686 +4300 val perplexity 1062.7574 +4300 train 6.627209 (lr=2.9551e-05) (hash(x)=43239281) +37200 val loss 5.8097 +37200 val perplexity 333.5259 +37200 train 6.032040 (lr=3.5836e-06) (hash(x)=57788407) +6200 val loss 6.9578 +6200 val perplexity 1051.3651 +6200 train 6.920312 (lr=1.9363e-05) (hash(x)=46394422) +4400 val loss 6.9667 +4400 val perplexity 1060.7450 +4400 train 6.640711 (lr=2.9529e-05) (hash(x)=45076737) +6300 val loss 6.9557 +6300 val perplexity 1049.0770 +6300 train 6.991278 (lr=1.9342e-05) (hash(x)=53748145) +37300 val loss 5.8055 +37300 val perplexity 332.1282 +37300 train 5.562133 (lr=3.5529e-06) (hash(x)=46333526) +6400 val loss 6.9372 +6400 val perplexity 1029.8702 +6400 train 6.829610 (lr=1.9320e-05) (hash(x)=46054751) +4500 val loss 6.9472 +4500 val perplexity 1040.2020 +4500 train 7.041008 (lr=2.9506e-05) (hash(x)=57930262) +37400 val loss 5.8042 +37400 val perplexity 331.6800 +37400 train 5.616745 (lr=3.5224e-06) (hash(x)=45862038) +6500 val loss 6.9281 +6500 val perplexity 1020.5649 +6500 train 7.145925 (lr=1.9299e-05) (hash(x)=51816809) +4600 val loss 6.9405 +4600 val perplexity 1033.3013 +4600 train 6.719579 (lr=2.9483e-05) (hash(x)=46721614) +6600 val loss 6.9224 +6600 val perplexity 1014.7181 +6600 train 6.844716 (lr=1.9276e-05) (hash(x)=52453336) +37500 val loss 5.8044 +37500 val perplexity 331.7577 +37500 train 5.506943 (lr=3.4921e-06) (hash(x)=46543051) +6700 val loss 6.9084 +6700 val perplexity 1000.5973 +6700 train 6.949422 (lr=1.9254e-05) (hash(x)=49108775) +4700 val loss 6.9305 +4700 val perplexity 1023.0546 +4700 train 6.752384 (lr=2.9460e-05) (hash(x)=49837920) +37600 val loss 5.8024 +37600 val perplexity 331.0776 +37600 train 5.762861 (lr=3.4620e-06) (hash(x)=51314450) +6800 val loss 6.8938 +6800 val perplexity 986.1386 +6800 train 6.821076 (lr=1.9231e-05) (hash(x)=46745396) +4800 val loss 6.9201 +4800 val perplexity 1012.4470 +4800 train 7.052123 (lr=2.9436e-05) (hash(x)=48380045) +6900 val loss 6.8826 +6900 val perplexity 975.1379 +6900 train 7.041726 (lr=1.9208e-05) (hash(x)=46534986) +37700 val loss 5.8027 +37700 val perplexity 331.1888 +37700 train 5.653619 (lr=3.4321e-06) (hash(x)=46706763) +7000 val loss 6.8713 +7000 val perplexity 964.2391 +7000 train 7.242034 (lr=1.9185e-05) (hash(x)=49317888) +4900 val loss 6.9119 +4900 val perplexity 1004.1332 +4900 train 6.763124 (lr=2.9411e-05) (hash(x)=44202577) +7100 val loss 6.8579 +7100 val perplexity 951.4022 +7100 train 6.858851 (lr=1.9161e-05) (hash(x)=50360484) +37800 val loss 5.8043 +37800 val perplexity 331.7183 +37800 train 5.649731 (lr=3.4024e-06) (hash(x)=45793085) +5000 val loss 6.8899 +5000 val perplexity 982.2875 +5000 train 6.911688 (lr=2.9386e-05) (hash(x)=52038024) +7200 val loss 6.8598 +7200 val perplexity 953.2181 +7200 train 6.732186 (lr=1.9137e-05) (hash(x)=49515094) +37900 val loss 5.8032 +37900 val perplexity 331.3444 +37900 train 5.673090 (lr=3.3728e-06) (hash(x)=49969493) +7300 val loss 6.8645 +7300 val perplexity 957.6605 +7300 train 6.862384 (lr=1.9112e-05) (hash(x)=51546861) +5100 val loss 6.8705 +5100 val perplexity 963.3829 +5100 train 7.051399 (lr=2.9360e-05) (hash(x)=53700038) +7400 val loss 6.8287 +7400 val perplexity 924.0242 +7400 train 6.833217 (lr=1.9087e-05) (hash(x)=48320948) +38000 val loss 5.8018 +38000 val perplexity 330.9039 +38000 train 5.591085 (lr=3.3435e-06) (hash(x)=47184489) +7500 val loss 6.8510 +7500 val perplexity 944.8021 +7500 train 6.681520 (lr=1.9062e-05) (hash(x)=40167457) +5200 val loss 6.8629 +5200 val perplexity 956.1302 +5200 train 6.891701 (lr=2.9334e-05) (hash(x)=48137625) +7600 val loss 6.8041 +7600 val perplexity 901.5132 +7600 train 6.739791 (lr=1.9037e-05) (hash(x)=49942165) +38100 val loss 5.8007 +38100 val perplexity 330.5446 +38100 train 6.253730 (lr=3.3143e-06) (hash(x)=54905742) +5300 val loss 6.8480 +5300 val perplexity 941.9955 +5300 train 6.710324 (lr=2.9307e-05) (hash(x)=43161573) +7700 val loss 6.7932 +7700 val perplexity 891.7384 +7700 train 6.525459 (lr=1.9011e-05) (hash(x)=48853311) +38200 val loss 5.8005 +38200 val perplexity 330.4677 +38200 train 6.053580 (lr=3.2854e-06) (hash(x)=48831719) +7800 val loss 6.7823 +7800 val perplexity 882.1234 +7800 train 6.677845 (lr=1.8985e-05) (hash(x)=48510117) +5400 val loss 6.8356 +5400 val perplexity 930.3694 +5400 train 7.028541 (lr=2.9280e-05) (hash(x)=56673322) +7900 val loss 6.7722 +7900 val perplexity 873.2110 +7900 train 6.772038 (lr=1.8959e-05) (hash(x)=48339781) +38300 val loss 5.8004 +38300 val perplexity 330.4370 +38300 train 5.658151 (lr=3.2566e-06) (hash(x)=48944700) +5500 val loss 6.8360 +5500 val perplexity 930.7462 +5500 train 7.027546 (lr=2.9252e-05) (hash(x)=53468295) +8000 val loss 6.7581 +8000 val perplexity 861.0417 +8000 train 6.875800 (lr=1.8932e-05) (hash(x)=54927320) +8100 val loss 6.7527 +8100 val perplexity 856.3400 +8100 train 6.470116 (lr=1.8905e-05) (hash(x)=46461786) +38400 val loss 5.8003 +38400 val perplexity 330.3956 +38400 train 5.438194 (lr=3.2281e-06) (hash(x)=46504848) +5600 val loss 6.8136 +5600 val perplexity 910.1801 +5600 train 7.126484 (lr=2.9224e-05) (hash(x)=59287280) +8200 val loss 6.7415 +8200 val perplexity 846.8405 +8200 train 6.675226 (lr=1.8878e-05) (hash(x)=51536260) +38500 val loss 5.7998 +38500 val perplexity 330.2318 +38500 train 5.547457 (lr=3.1997e-06) (hash(x)=50148948) +5700 val loss 6.7996 +5700 val perplexity 897.5067 +5700 train 7.042775 (lr=2.9196e-05) (hash(x)=57575806) +8300 val loss 6.7343 +8300 val perplexity 840.7861 +8300 train 6.471604 (lr=1.8850e-05) (hash(x)=44770722) +8400 val loss 6.7183 +8400 val perplexity 827.4392 +8400 train 6.711835 (lr=1.8822e-05) (hash(x)=50104957) +38600 val loss 5.8004 +38600 val perplexity 330.4203 +38600 train 6.575957 (lr=3.1716e-06) (hash(x)=60660913) +5800 val loss 6.7894 +5800 val perplexity 888.3729 +5800 train 6.765071 (lr=2.9166e-05) (hash(x)=46897279) +8500 val loss 6.7127 +8500 val perplexity 822.7887 +8500 train 6.833847 (lr=1.8794e-05) (hash(x)=50132971) +38700 val loss 5.8003 +38700 val perplexity 330.4113 +38700 train 5.610072 (lr=3.1436e-06) (hash(x)=50256252) +8600 val loss 6.7029 +8600 val perplexity 814.7958 +8600 train 6.661052 (lr=1.8766e-05) (hash(x)=52193699) +5900 val loss 6.7793 +5900 val perplexity 879.4363 +5900 train 6.645764 (lr=2.9137e-05) (hash(x)=47565679) +8700 val loss 6.6874 +8700 val perplexity 802.2065 +8700 train 6.693353 (lr=1.8737e-05) (hash(x)=47902319) +38800 val loss 5.7988 +38800 val perplexity 329.9174 +38800 train 5.884041 (lr=3.1159e-06) (hash(x)=52592415) +6000 val loss 6.7548 +6000 val perplexity 858.1721 +6000 train 6.641681 (lr=2.9106e-05) (hash(x)=51590090) +8800 val loss 6.6733 +8800 val perplexity 791.0112 +8800 train 6.962318 (lr=1.8708e-05) (hash(x)=54904230) +38900 val loss 5.7962 +38900 val perplexity 329.0420 +38900 train 5.753305 (lr=3.0883e-06) (hash(x)=51945430) +8900 val loss 6.6708 +8900 val perplexity 789.0576 +8900 train 6.564697 (lr=1.8678e-05) (hash(x)=46311615) +6100 val loss 6.7350 +6100 val perplexity 841.3701 +6100 train 7.148363 (lr=2.9076e-05) (hash(x)=59732271) +9000 val loss 6.6680 +9000 val perplexity 786.8581 +9000 train 6.570761 (lr=1.8648e-05) (hash(x)=48535188) +39000 val loss 5.7962 +39000 val perplexity 329.0483 +39000 train 6.038047 (lr=3.0610e-06) (hash(x)=51838869) +6200 val loss 6.7218 +6200 val perplexity 830.3090 +6200 train 6.730457 (lr=2.9044e-05) (hash(x)=46394422) +9100 val loss 6.6423 +9100 val perplexity 766.8331 +9100 train 6.699732 (lr=1.8618e-05) (hash(x)=51757372) +9200 val loss 6.6322 +9200 val perplexity 759.1351 +39100 val loss 5.7910 +39100 val perplexity 327.3528 +9200 train 6.498425 (lr=1.8588e-05) (hash(x)=51131708) +39100 train 5.949833 (lr=3.0339e-06) (hash(x)=54556054) +6300 val loss 6.7129 +6300 val perplexity 822.9562 +6300 train 6.745351 (lr=2.9013e-05) (hash(x)=53748145) +9300 val loss 6.6233 +9300 val perplexity 752.4560 +9300 train 6.583850 (lr=1.8557e-05) (hash(x)=44784276) +39200 val loss 5.7916 +39200 val perplexity 327.5486 +39200 train 6.159741 (lr=3.0069e-06) (hash(x)=48915509) +6400 val loss 6.6963 +6400 val perplexity 809.3851 +6400 train 6.565305 (lr=2.8981e-05) (hash(x)=46054751) +9400 val loss 6.6140 +9400 val perplexity 745.4399 +9400 train 6.713509 (lr=1.8526e-05) (hash(x)=51981169) +9500 val loss 6.6029 +9500 val perplexity 737.2364 +9500 train 6.548369 (lr=1.8495e-05) (hash(x)=47232936) +6500 val loss 6.6803 +6500 val perplexity 796.5932 +6500 train 6.894208 (lr=2.8948e-05) (hash(x)=51816809) +39300 val loss 5.7952 +39300 val perplexity 328.7202 +39300 train 5.599744 (lr=2.9802e-06) (hash(x)=46121589) +9600 val loss 6.6007 +9600 val perplexity 735.5938 +9600 train 6.618612 (lr=1.8463e-05) (hash(x)=53800450) +6600 val loss 6.6689 +6600 val perplexity 787.5052 +6600 train 6.546772 (lr=2.8915e-05) (hash(x)=52453336) +39400 val loss 5.7933 +39400 val perplexity 328.0949 +39400 train 5.319389 (lr=2.9537e-06) (hash(x)=45556201) +9700 val loss 6.5856 +9700 val perplexity 724.5635 +9700 train 6.718117 (lr=1.8432e-05) (hash(x)=55768123) +9800 val loss 6.5824 +9800 val perplexity 722.2406 +9800 train 6.485301 (lr=1.8399e-05) (hash(x)=47745177) +6700 val loss 6.6587 +6700 val perplexity 779.5637 +6700 train 6.687222 (lr=2.8881e-05) (hash(x)=49108775) +39500 val loss 5.7926 +39500 val perplexity 327.8751 +39500 train 5.831804 (lr=2.9274e-06) (hash(x)=47014749) +9900 val loss 6.5768 +9900 val perplexity 718.2352 +9900 train 6.853314 (lr=1.8367e-05) (hash(x)=56592246) +6800 val loss 6.6496 +6800 val perplexity 772.4988 +6800 train 6.554808 (lr=2.8847e-05) (hash(x)=46745396) +10000 val loss 6.5695 +10000 val perplexity 712.9823 +39600 val loss 5.7950 +39600 val perplexity 328.6419 +39600 train 5.597177 (lr=2.9014e-06) (hash(x)=47395114) +10000 train 6.598790 (lr=1.8334e-05) (hash(x)=51655963) +10100 val loss 6.5608 +10100 val perplexity 706.8427 +10100 train 6.472806 (lr=1.8301e-05) (hash(x)=49809511) +6900 val loss 6.6302 +6900 val perplexity 757.6134 +6900 train 6.795412 (lr=2.8812e-05) (hash(x)=46534986) +39700 val loss 5.7966 +39700 val perplexity 329.1624 +39700 train 5.933496 (lr=2.8755e-06) (hash(x)=50858008) +10200 val loss 6.5506 +10200 val perplexity 699.6796 +10200 train 6.151064 (lr=1.8268e-05) (hash(x)=42297812) +7000 val loss 6.6220 +7000 val perplexity 751.4176 +7000 train 6.984348 (lr=2.8777e-05) (hash(x)=49317888) +10300 val loss 6.5493 +10300 val perplexity 698.7657 +10300 train 6.291403 (lr=1.8234e-05) (hash(x)=55529820) +39800 val loss 5.7927 +39800 val perplexity 327.8849 +39800 train 5.518047 (lr=2.8498e-06) (hash(x)=44096434) +10400 val loss 6.5355 +10400 val perplexity 689.1912 +10400 train 6.555791 (lr=1.8200e-05) (hash(x)=53255684) +7100 val loss 6.6047 +7100 val perplexity 738.5612 +7100 train 6.615108 (lr=2.8741e-05) (hash(x)=50360484) +39900 val loss 5.7905 +39900 val perplexity 327.1691 +39900 train 5.525070 (lr=2.8244e-06) (hash(x)=46902151) +10500 val loss 6.5225 +10500 val perplexity 680.2932 +10500 train 6.646883 (lr=1.8166e-05) (hash(x)=54306191) +7200 val loss 6.5970 +7200 val perplexity 732.8769 +7200 train 6.459873 (lr=2.8705e-05) (hash(x)=49515094) +10600 val loss 6.5141 +10600 val perplexity 674.6191 +10600 train 6.772575 (lr=1.8132e-05) (hash(x)=60130567) +40000 val loss 5.7907 +40000 val perplexity 327.2545 +40000 train 5.579343 (lr=2.7992e-06) (hash(x)=46735297) +10700 val loss 6.5082 +10700 val perplexity 670.5950 +10700 train 6.441580 (lr=1.8097e-05) (hash(x)=50074737) +7300 val loss 6.5970 +7300 val perplexity 732.8573 +7300 train 6.591541 (lr=2.8668e-05) (hash(x)=51546861) +10800 val loss 6.4942 +10800 val perplexity 661.3207 +40100 val loss 5.7918 +40100 val perplexity 327.5977 +10800 train 6.606314 (lr=1.8062e-05) (hash(x)=51547220) +40100 train 5.736505 (lr=2.7742e-06) (hash(x)=48035170) +7400 val loss 6.5846 +7400 val perplexity 723.8818 +7400 train 6.580837 (lr=2.8631e-05) (hash(x)=48320948) +10900 val loss 6.4899 +10900 val perplexity 658.4739 +10900 train 6.576293 (lr=1.8026e-05) (hash(x)=55943981) +40200 val loss 5.7910 +40200 val perplexity 327.3313 +40200 train 5.612554 (lr=2.7494e-06) (hash(x)=46651322) +11000 val loss 6.4768 +11000 val perplexity 649.9118 +11000 train 6.490388 (lr=1.7991e-05) (hash(x)=46444570) +7500 val loss 6.6144 +7500 val perplexity 745.7269 +7500 train 6.426689 (lr=2.8594e-05) (hash(x)=40167457) +11100 val loss 6.4695 +11100 val perplexity 645.1491 +11100 train 6.479462 (lr=1.7955e-05) (hash(x)=49589063) +40300 val loss 5.7905 +40300 val perplexity 327.1910 +40300 train 5.638853 (lr=2.7248e-06) (hash(x)=46378099) +7600 val loss 6.5656 +7600 val perplexity 710.2097 +7600 train 6.521790 (lr=2.8556e-05) (hash(x)=49942165) +11200 val loss 6.4586 +11200 val perplexity 638.1451 +11200 train 6.440417 (lr=1.7919e-05) (hash(x)=51392283) +40400 val loss 5.7877 +40400 val perplexity 326.2671 +40400 train 5.623124 (lr=2.7004e-06) (hash(x)=46495438) +11300 val loss 6.4543 +11300 val perplexity 635.4151 +11300 train 6.374045 (lr=1.7882e-05) (hash(x)=45081133) +7700 val loss 6.5547 +7700 val perplexity 702.5410 +7700 train 6.279706 (lr=2.8517e-05) (hash(x)=48853311) +11400 val loss 6.4466 +11400 val perplexity 630.5587 +11400 train 6.576564 (lr=1.7845e-05) (hash(x)=53700397) +40500 val loss 5.7852 +40500 val perplexity 325.4369 +40500 train 5.965801 (lr=2.6763e-06) (hash(x)=52059147) +7800 val loss 6.5460 +7800 val perplexity 696.4794 +7800 train 6.435349 (lr=2.8478e-05) (hash(x)=48510117) +11500 val loss 6.4391 +11500 val perplexity 625.8223 +11500 train 6.141678 (lr=1.7809e-05) (hash(x)=43839088) +11600 val loss 6.4333 +11600 val perplexity 622.2242 +11600 train 6.462057 (lr=1.7771e-05) (hash(x)=48088111) +40600 val loss 5.7852 +40600 val perplexity 325.4349 +40600 train 5.944921 (lr=2.6524e-06) (hash(x)=54885045) +7900 val loss 6.5335 +7900 val perplexity 687.8083 +7900 train 6.534476 (lr=2.8438e-05) (hash(x)=48339781) +11700 val loss 6.4256 +11700 val perplexity 617.4771 +11700 train 6.758646 (lr=1.7734e-05) (hash(x)=55108226) +40700 val loss 5.7832 +40700 val perplexity 324.8008 +40700 train 6.190995 (lr=2.6287e-06) (hash(x)=53213971) +8000 val loss 6.5302 +8000 val perplexity 685.5157 +8000 train 6.631647 (lr=2.8398e-05) (hash(x)=54927320) +11800 val loss 6.4160 +11800 val perplexity 611.5540 +11800 train 6.693108 (lr=1.7696e-05) (hash(x)=58524839) +11900 val loss 6.4109 +11900 val perplexity 608.4536 +11900 train 6.069282 (lr=1.7658e-05) (hash(x)=43864078) +40800 val loss 5.7815 +40800 val perplexity 324.2373 +40800 train 5.414309 (lr=2.6053e-06) (hash(x)=45133794) +8100 val loss 6.5352 +8100 val perplexity 688.9568 +8100 train 6.259796 (lr=2.8358e-05) (hash(x)=46461786) +12000 val loss 6.4053 +12000 val perplexity 605.0189 +12000 train 6.088469 (lr=1.7619e-05) (hash(x)=43448544) +40900 val loss 5.7804 +40900 val perplexity 323.8820 +40900 train 5.920214 (lr=2.5820e-06) (hash(x)=56546547) +12100 val loss 6.4010 +12100 val perplexity 602.4222 +12100 train 6.389794 (lr=1.7581e-05) (hash(x)=55200399) +8200 val loss 6.5172 +8200 val perplexity 676.7072 +8200 train 6.438480 (lr=2.8317e-05) (hash(x)=51536260) +12200 val loss 6.3976 +12200 val perplexity 600.3896 +12200 train 6.555463 (lr=1.7542e-05) (hash(x)=57627314) +41000 val loss 5.7812 +41000 val perplexity 324.1488 +41000 train 6.175647 (lr=2.5590e-06) (hash(x)=49552270) +8300 val loss 6.5100 +8300 val perplexity 671.8131 +8300 train 6.225160 (lr=2.8276e-05) (hash(x)=44770722) +12300 val loss 6.3920 +12300 val perplexity 597.0750 +12300 train 6.686368 (lr=1.7503e-05) (hash(x)=53617087) +12400 val loss 6.3915 +12400 val perplexity 596.7524 +12400 train 6.238227 (lr=1.7464e-05) (hash(x)=51135678) +41100 val loss 5.7820 +41100 val perplexity 324.4072 +41100 train 6.016449 (lr=2.5362e-06) (hash(x)=51222370) +8400 val loss 6.4976 +8400 val perplexity 663.5226 +8400 train 6.498326 (lr=2.8234e-05) (hash(x)=50104957) +12500 val loss 6.3961 +12500 val perplexity 599.5096 +12500 train 6.379384 (lr=1.7424e-05) (hash(x)=48025130) +41200 val loss 5.7835 +41200 val perplexity 324.8964 +41200 train 5.817165 (lr=2.5137e-06) (hash(x)=50883608) +8500 val loss 6.4963 +8500 val perplexity 662.6705 +8500 train 6.602393 (lr=2.8191e-05) (hash(x)=50132971) +12600 val loss 6.3819 +12600 val perplexity 591.0383 +12600 train 6.455917 (lr=1.7384e-05) (hash(x)=52135695) +12700 val loss 6.3759 +12700 val perplexity 587.5368 +12700 train 6.225551 (lr=1.7344e-05) (hash(x)=51888613) +41300 val loss 5.7828 +41300 val perplexity 324.6816 +41300 train 6.032202 (lr=2.4913e-06) (hash(x)=52996828) +8600 val loss 6.4881 +8600 val perplexity 657.2643 +8600 train 6.454095 (lr=2.8149e-05) (hash(x)=52193699) +12800 val loss 6.3716 +12800 val perplexity 585.0143 +12800 train 6.316040 (lr=1.7303e-05) (hash(x)=50418818) +41400 val loss 5.7820 +41400 val perplexity 324.4106 +41400 train 5.739006 (lr=2.4692e-06) (hash(x)=48822716) +12900 val loss 6.3516 +12900 val perplexity 573.3994 +12900 train 7.288659 (lr=1.7263e-05) (hash(x)=58649585) +8700 val loss 6.4808 +8700 val perplexity 652.5202 +8700 train 6.485448 (lr=2.8105e-05) (hash(x)=47902319) +13000 val loss 6.3470 +13000 val perplexity 570.8048 +13000 train 6.643065 (lr=1.7222e-05) (hash(x)=54567307) +41500 val loss 5.7837 +41500 val perplexity 324.9680 +41500 train 5.764733 (lr=2.4474e-06) (hash(x)=50936577) +8800 val loss 6.4688 +8800 val perplexity 644.7336 +8800 train 6.766957 (lr=2.8062e-05) (hash(x)=54904230) +13100 val loss 6.3356 +13100 val perplexity 564.2836 +13100 train 6.514795 (lr=1.7181e-05) (hash(x)=52071473) +13200 val loss 6.3275 +13200 val perplexity 559.7623 +13200 train 6.100046 (lr=1.7139e-05) (hash(x)=46293092) +41600 val loss 5.7844 +41600 val perplexity 325.1906 +41600 train 5.826467 (lr=2.4257e-06) (hash(x)=44375046) +8900 val loss 6.4769 +8900 val perplexity 649.9738 +8900 train 6.363889 (lr=2.8017e-05) (hash(x)=46311615) +13300 val loss 6.3229 +13300 val perplexity 557.1670 +13300 train 6.464483 (lr=1.7098e-05) (hash(x)=56511467) +41700 val loss 5.7814 +41700 val perplexity 324.2136 +41700 train 5.541641 (lr=2.4043e-06) (hash(x)=44060021) +9000 val loss 6.4852 +9000 val perplexity 655.3868 +9000 train 6.376635 (lr=2.7973e-05) (hash(x)=48535188) +13400 val loss 6.3110 +13400 val perplexity 550.5897 +13400 train 6.438718 (lr=1.7056e-05) (hash(x)=54753763) +13500 val loss 6.3099 +13500 val perplexity 550.0142 +13500 train 6.481706 (lr=1.7014e-05) (hash(x)=53610247) +41800 val loss 5.7825 +41800 val perplexity 324.5732 +41800 train 6.109643 (lr=2.3832e-06) (hash(x)=57765221) +9100 val loss 6.4464 +9100 val perplexity 630.4104 +9100 train 6.492523 (lr=2.7928e-05) (hash(x)=51757372) +13600 val loss 6.3025 +13600 val perplexity 545.9620 +13600 train 6.168638 (lr=1.6971e-05) (hash(x)=47526249) +9200 val loss 6.4428 +9200 val perplexity 628.1381 +9200 train 6.286029 (lr=2.7882e-05) (hash(x)=51131708) +13700 val loss 6.3015 +13700 val perplexity 545.3768 +13700 train 6.149080 (lr=1.6929e-05) (hash(x)=51185517) +9300 val loss 6.4476 +9300 val perplexity 631.1780 +9300 train 6.408109 (lr=2.7836e-05) (hash(x)=44784276) +13800 val loss 6.2877 +13800 val perplexity 537.9174 +13800 train 5.962605 (lr=1.6886e-05) (hash(x)=45953529) +9400 val loss 6.4365 +9400 val perplexity 624.2275 +9400 train 6.551546 (lr=2.7789e-05) (hash(x)=51981169) +13900 val loss 6.2884 +13900 val perplexity 538.3025 +13900 train 6.209026 (lr=1.6843e-05) (hash(x)=47238157) +9500 val loss 6.4304 +9500 val perplexity 620.3918 +9500 train 6.377679 (lr=2.7743e-05) (hash(x)=47232936) +14000 val loss 6.2766 +14000 val perplexity 531.9893 +14000 train 6.250593 (lr=1.6800e-05) (hash(x)=54250750) +9600 val loss 6.4265 +9600 val perplexity 618.0176 +9600 train 6.447754 (lr=2.7695e-05) (hash(x)=53800450) +14100 val loss 6.2778 +14100 val perplexity 532.6421 +14100 train 6.283087 (lr=1.6756e-05) (hash(x)=48198552) +9700 val loss 6.4206 +9700 val perplexity 614.3608 +9700 train 6.535484 (lr=2.7647e-05) (hash(x)=55768123) +14200 val loss 6.2684 +14200 val perplexity 527.6479 +14200 train 6.400944 (lr=1.6712e-05) (hash(x)=52020690) +9800 val loss 6.4211 +9800 val perplexity 614.6752 +9800 train 6.329356 (lr=2.7599e-05) (hash(x)=47745177) +14300 val loss 6.2700 +14300 val perplexity 528.4673 +14300 train 6.068441 (lr=1.6668e-05) (hash(x)=45165483) +9900 val loss 6.4165 +9900 val perplexity 611.8605 +9900 train 6.684052 (lr=2.7550e-05) (hash(x)=56592246) +14400 val loss 6.2605 +14400 val perplexity 523.4636 +14400 train 6.381829 (lr=1.6624e-05) (hash(x)=52184072) +10000 val loss 6.4127 +10000 val perplexity 609.5571 +10000 train 6.423335 (lr=2.7501e-05) (hash(x)=51655963) +14500 val loss 6.2595 +14500 val perplexity 522.9387 +14500 train 6.625453 (lr=1.6580e-05) (hash(x)=56046436) +10100 val loss 6.4063 +10100 val perplexity 605.6565 +10100 train 6.316051 (lr=2.7452e-05) (hash(x)=49809511) +14600 val loss 6.2506 +14600 val perplexity 518.3021 +14600 train 6.077270 (lr=1.6535e-05) (hash(x)=52029694) +10200 val loss 6.4073 +10200 val perplexity 606.2531 +10200 train 6.003387 (lr=2.7402e-05) (hash(x)=42297812) +10300 val loss 6.4008 +10300 val perplexity 602.3458 +10300 train 6.107283 (lr=2.7351e-05) (hash(x)=55529820) +14700 val loss 6.2547 +14700 val perplexity 520.4769 +14700 train 6.127725 (lr=1.6490e-05) (hash(x)=50258224) +10400 val loss 6.3959 +10400 val perplexity 599.3727 +10400 train 6.402360 (lr=2.7300e-05) (hash(x)=53255684) +14800 val loss 6.2477 +14800 val perplexity 516.8441 +14800 train 5.555237 (lr=1.6445e-05) (hash(x)=42112262) +10500 val loss 6.3880 +10500 val perplexity 594.6732 +10500 train 6.511672 (lr=2.7249e-05) (hash(x)=54306191) +14900 val loss 6.2454 +14900 val perplexity 515.6106 +14900 train 5.989807 (lr=1.6400e-05) (hash(x)=47219933) +10600 val loss 6.3645 +10600 val perplexity 580.8281 +10600 train 6.605578 (lr=2.7197e-05) (hash(x)=60130567) +15000 val loss 6.2683 +15000 val perplexity 527.5637 +15000 train 5.984100 (lr=1.6354e-05) (hash(x)=58309309) +10700 val loss 6.3720 +10700 val perplexity 585.2172 +10700 train 6.300022 (lr=2.7145e-05) (hash(x)=50074737) +15100 val loss 6.2403 +15100 val perplexity 513.0330 +15100 train 6.133285 (lr=1.6308e-05) (hash(x)=48756049) +10800 val loss 6.3473 +10800 val perplexity 570.9641 +10800 train 6.463106 (lr=2.7093e-05) (hash(x)=51547220) +15200 val loss 6.2448 +15200 val perplexity 515.3191 +15200 train 5.959652 (lr=1.6262e-05) (hash(x)=49791737) +10900 val loss 6.3492 +10900 val perplexity 572.0287 +10900 train 6.445091 (lr=2.7040e-05) (hash(x)=55943981) +15300 val loss 6.2314 +15300 val perplexity 508.4584 +15300 train 6.199980 (lr=1.6216e-05) (hash(x)=53084126) +11000 val loss 6.3456 +11000 val perplexity 569.9709 +11000 train 6.380251 (lr=2.6986e-05) (hash(x)=46444570) +15400 val loss 6.2193 +15400 val perplexity 502.3578 +15400 train 6.332304 (lr=1.6170e-05) (hash(x)=55041679) +11100 val loss 6.3319 +11100 val perplexity 562.2263 +11100 train 6.358819 (lr=2.6932e-05) (hash(x)=49589063) +15500 val loss 6.2079 +15500 val perplexity 496.6724 +15500 train 6.005895 (lr=1.6123e-05) (hash(x)=43249867) +11200 val loss 6.3314 +11200 val perplexity 561.9272 +11200 train 6.321804 (lr=2.6878e-05) (hash(x)=51392283) +11300 val loss 6.3328 +11300 val perplexity 562.7238 +15600 val loss 6.2097 +15600 val perplexity 497.5314 +11300 train 6.258594 (lr=2.6823e-05) (hash(x)=45081133) +15600 train 6.188116 (lr=1.6077e-05) (hash(x)=49006517) +11400 val loss 6.3132 +11400 val perplexity 551.8087 +11400 train 6.428157 (lr=2.6768e-05) (hash(x)=53700397) +15700 val loss 6.1972 +15700 val perplexity 491.3583 +15700 train 6.533650 (lr=1.6030e-05) (hash(x)=60986839) +11500 val loss 6.3073 +11500 val perplexity 548.5825 +11500 train 6.028381 (lr=2.6713e-05) (hash(x)=43839088) +15800 val loss 6.1905 +15800 val perplexity 488.0716 +15800 train 6.113477 (lr=1.5982e-05) (hash(x)=48354906) +11600 val loss 6.3009 +11600 val perplexity 545.0706 +11600 train 6.311018 (lr=2.6657e-05) (hash(x)=48088111) +15900 val loss 6.1867 +15900 val perplexity 486.2625 +15900 train 6.046615 (lr=1.5935e-05) (hash(x)=52679780) +11700 val loss 6.2979 +11700 val perplexity 543.4081 +11700 train 6.627666 (lr=2.6601e-05) (hash(x)=55108226) +16000 val loss 6.1875 +16000 val perplexity 486.6359 +16000 train 6.343832 (lr=1.5887e-05) (hash(x)=58049587) +11800 val loss 6.2948 +11800 val perplexity 541.7360 +11800 train 6.570461 (lr=2.6544e-05) (hash(x)=58524839) +16100 val loss 6.1768 +16100 val perplexity 481.4322 +16100 train 5.772702 (lr=1.5840e-05) (hash(x)=39795825) +11900 val loss 6.2895 +11900 val perplexity 538.8598 +11900 train 5.951778 (lr=2.6487e-05) (hash(x)=43864078) +16200 val loss 6.1829 +16200 val perplexity 484.4030 +16200 train 6.145854 (lr=1.5792e-05) (hash(x)=49498689) +12000 val loss 6.2874 +12000 val perplexity 537.7761 +12000 train 5.976575 (lr=2.6429e-05) (hash(x)=43448544) +16300 val loss 6.1814 +16300 val perplexity 483.6715 +16300 train 6.457381 (lr=1.5744e-05) (hash(x)=51235194) +12100 val loss 6.2875 +12100 val perplexity 537.8207 +12100 train 6.261235 (lr=2.6371e-05) (hash(x)=55200399) +16400 val loss 6.1773 +16400 val perplexity 481.6779 +16400 train 6.038260 (lr=1.5695e-05) (hash(x)=47789417) +12200 val loss 6.2775 +12200 val perplexity 532.4796 +12200 train 6.425420 (lr=2.6313e-05) (hash(x)=57627314) +16500 val loss 6.1644 +16500 val perplexity 475.5187 +12300 val loss 6.2805 +12300 val perplexity 534.0385 +16500 train 6.734771 (lr=1.5647e-05) (hash(x)=59665889) +12300 train 6.582706 (lr=2.6254e-05) (hash(x)=53617087) +12400 val loss 6.2833 +12400 val perplexity 535.5295 +12400 train 6.139731 (lr=2.6195e-05) (hash(x)=51135678) +16600 val loss 6.1633 +16600 val perplexity 475.0009 +16600 train 6.050817 (lr=1.5598e-05) (hash(x)=49034893) +12500 val loss 6.2874 +12500 val perplexity 537.7396 +12500 train 6.300069 (lr=2.6136e-05) (hash(x)=48025130) +16700 val loss 6.1563 +16700 val perplexity 471.6913 +16700 train 5.994668 (lr=1.5549e-05) (hash(x)=45685955) +12600 val loss 6.2811 +12600 val perplexity 534.3832 +12600 train 6.340992 (lr=2.6076e-05) (hash(x)=52135695) +16800 val loss 6.1524 +16800 val perplexity 469.8439 +16800 train 5.953376 (lr=1.5500e-05) (hash(x)=49589835) +12700 val loss 6.2673 +12700 val perplexity 527.0472 +12700 train 6.130489 (lr=2.6016e-05) (hash(x)=51888613) +16900 val loss 6.1502 +16900 val perplexity 468.7939 +16900 train 5.562690 (lr=1.5451e-05) (hash(x)=38655565) +12800 val loss 6.2615 +12800 val perplexity 523.9930 +12800 train 6.195919 (lr=2.5955e-05) (hash(x)=50418818) +17000 val loss 6.1429 +17000 val perplexity 465.3944 +17000 train 6.074118 (lr=1.5401e-05) (hash(x)=46730732) +12900 val loss 6.2405 +12900 val perplexity 513.1074 +12900 train 7.155957 (lr=2.5894e-05) (hash(x)=58649585) +17100 val loss 6.1497 +17100 val perplexity 468.5907 +17100 train 5.775746 (lr=1.5352e-05) (hash(x)=43843722) +13000 val loss 6.2417 +13000 val perplexity 513.7241 +13000 train 6.543957 (lr=2.5833e-05) (hash(x)=54567307) +17200 val loss 6.1450 +17200 val perplexity 466.3618 +17200 train 5.907995 (lr=1.5302e-05) (hash(x)=50198978) +13100 val loss 6.2277 +13100 val perplexity 506.5857 +13100 train 6.391589 (lr=2.5771e-05) (hash(x)=52071473) +17300 val loss 6.1456 +17300 val perplexity 466.6370 +17300 train 6.146933 (lr=1.5252e-05) (hash(x)=52976276) +13200 val loss 6.2173 +13200 val perplexity 501.3470 +13200 train 5.986736 (lr=2.5709e-05) (hash(x)=46293092) +17400 val loss 6.1382 +17400 val perplexity 463.2163 +17400 train 5.974552 (lr=1.5202e-05) (hash(x)=52224275) +13300 val loss 6.2114 +13300 val perplexity 498.4130 +13300 train 6.343679 (lr=2.5647e-05) (hash(x)=56511467) +17500 val loss 6.1466 +17500 val perplexity 467.1130 +17500 train 5.944997 (lr=1.5152e-05) (hash(x)=49055112) +13400 val loss 6.2102 +13400 val perplexity 497.8014 +13400 train 6.326076 (lr=2.5584e-05) (hash(x)=54753763) +13500 val loss 6.2064 +13500 val perplexity 495.9179 +13500 train 6.362766 (lr=2.5521e-05) (hash(x)=53610247) +17600 val loss 6.1389 +17600 val perplexity 463.5208 +17600 train 6.177176 (lr=1.5101e-05) (hash(x)=49996498) +13600 val loss 6.2035 +13600 val perplexity 494.4572 +13600 train 6.070455 (lr=2.5457e-05) (hash(x)=47526249) +17700 val loss 6.1283 +17700 val perplexity 458.6779 +17700 train 6.124263 (lr=1.5050e-05) (hash(x)=50526074) +13700 val loss 6.1924 +13700 val perplexity 489.0071 +13700 train 6.040852 (lr=2.5393e-05) (hash(x)=51185517) +17800 val loss 6.1163 +17800 val perplexity 453.1730 +17800 train 5.990662 (lr=1.5000e-05) (hash(x)=48410966) +13800 val loss 6.1885 +13800 val perplexity 487.1234 +13800 train 5.862451 (lr=2.5329e-05) (hash(x)=45953529) +17900 val loss 6.1149 +17900 val perplexity 452.5472 +17900 train 6.019664 (lr=1.4949e-05) (hash(x)=51654550) +13900 val loss 6.1897 +13900 val perplexity 487.7038 +13900 train 6.112508 (lr=2.5264e-05) (hash(x)=47238157) +18000 val loss 6.1058 +18000 val perplexity 448.4368 +18000 train 6.192564 (lr=1.4898e-05) (hash(x)=44227619) +14000 val loss 6.1817 +14000 val perplexity 483.8129 +14000 train 6.152245 (lr=2.5199e-05) (hash(x)=54250750) +18100 val loss 6.1046 +18100 val perplexity 447.9284 +18100 train 5.876388 (lr=1.4846e-05) (hash(x)=45771090) +14100 val loss 6.1829 +14100 val perplexity 484.4113 +14100 train 6.178734 (lr=2.5134e-05) (hash(x)=48198552) +18200 val loss 6.0999 +18200 val perplexity 445.8280 +18200 train 5.709765 (lr=1.4795e-05) (hash(x)=37914739) +14200 val loss 6.1799 +14200 val perplexity 482.9359 +14200 train 6.291629 (lr=2.5068e-05) (hash(x)=52020690) +18300 val loss 6.0939 +18300 val perplexity 443.1320 +18300 train 6.563478 (lr=1.4743e-05) (hash(x)=68784189) +14300 val loss 6.1738 +14300 val perplexity 479.9911 +14300 train 5.991513 (lr=2.5002e-05) (hash(x)=45165483) +18400 val loss 6.0904 +18400 val perplexity 441.6196 +18400 train 6.196808 (lr=1.4692e-05) (hash(x)=55006900) +14400 val loss 6.1680 +14400 val perplexity 477.2226 +14400 train 6.264584 (lr=2.4936e-05) (hash(x)=52184072) +18500 val loss 6.0847 +18500 val perplexity 439.0848 +18500 train 6.512639 (lr=1.4640e-05) (hash(x)=56001248) +14500 val loss 6.1665 +14500 val perplexity 476.5056 +14500 train 6.543195 (lr=2.4869e-05) (hash(x)=56046436) +14600 val loss 6.1706 +14600 val perplexity 478.4710 +14600 train 5.988442 (lr=2.4802e-05) (hash(x)=52029694) +18600 val loss 6.0894 +18600 val perplexity 441.1639 +18600 train 6.263895 (lr=1.4588e-05) (hash(x)=51339352) +14700 val loss 6.1733 +14700 val perplexity 479.7799 +14700 train 6.055110 (lr=2.4735e-05) (hash(x)=50258224) +18700 val loss 6.0824 +18700 val perplexity 438.0617 +18700 train 6.221242 (lr=1.4536e-05) (hash(x)=53256371) +14800 val loss 6.1702 +14800 val perplexity 478.3008 +14800 train 5.483133 (lr=2.4668e-05) (hash(x)=42112262) +18800 val loss 6.0742 +18800 val perplexity 434.5176 +18800 train 6.043338 (lr=1.4483e-05) (hash(x)=51580096) +14900 val loss 6.1602 +14900 val perplexity 473.5176 +14900 train 5.922615 (lr=2.4600e-05) (hash(x)=47219933) +18900 val loss 6.0767 +18900 val perplexity 435.5679 +18900 train 6.144569 (lr=1.4431e-05) (hash(x)=53001073) +15000 val loss 6.1880 +15000 val perplexity 486.8824 +15000 train 5.891795 (lr=2.4531e-05) (hash(x)=58309309) +19000 val loss 6.0745 +19000 val perplexity 434.6262 +19000 train 5.917216 (lr=1.4378e-05) (hash(x)=51531298) +15100 val loss 6.1635 +15100 val perplexity 475.0865 +15100 train 6.054174 (lr=2.4463e-05) (hash(x)=48756049) +19100 val loss 6.0718 +19100 val perplexity 433.4719 +19100 train 5.860352 (lr=1.4326e-05) (hash(x)=48047406) +15200 val loss 6.1626 +15200 val perplexity 474.6823 +15200 train 5.871546 (lr=2.4394e-05) (hash(x)=49791737) +19200 val loss 6.0651 +19200 val perplexity 430.5753 +19200 train 5.862375 (lr=1.4273e-05) (hash(x)=44939100) +15300 val loss 6.1614 +15300 val perplexity 474.0892 +15300 train 6.114204 (lr=2.4324e-05) (hash(x)=53084126) +19300 val loss 6.0640 +19300 val perplexity 430.0791 +19300 train 5.928426 (lr=1.4220e-05) (hash(x)=50157419) +15400 val loss 6.1452 +15400 val perplexity 466.4908 +15400 train 6.246527 (lr=2.4255e-05) (hash(x)=55041679) +19400 val loss 6.0605 +19400 val perplexity 428.5787 +19400 train 5.877625 (lr=1.4167e-05) (hash(x)=46973593) +15500 val loss 6.1330 +15500 val perplexity 460.8046 +15500 train 5.945592 (lr=2.4185e-05) (hash(x)=43249867) +19500 val loss 6.0585 +19500 val perplexity 427.7506 +19500 train 5.880248 (lr=1.4114e-05) (hash(x)=44056368) +15600 val loss 6.1364 +15600 val perplexity 462.3803 +15600 train 6.115905 (lr=2.4115e-05) (hash(x)=49006517) +15700 val loss 6.1257 +15700 val perplexity 457.4771 +15700 train 6.449719 (lr=2.4044e-05) (hash(x)=60986839) +19600 val loss 6.0533 +19600 val perplexity 425.5116 +19600 train 6.380017 (lr=1.4060e-05) (hash(x)=58496552) +15800 val loss 6.1231 +15800 val perplexity 456.2701 +15800 train 6.061267 (lr=2.3974e-05) (hash(x)=48354906) +19700 val loss 6.0488 +19700 val perplexity 423.6030 +19700 train 5.907188 (lr=1.4007e-05) (hash(x)=47753232) +15900 val loss 6.1177 +15900 val perplexity 453.8130 +15900 train 5.972078 (lr=2.3903e-05) (hash(x)=52679780) +19800 val loss 6.0437 +19800 val perplexity 421.4333 +19800 train 5.866465 (lr=1.3953e-05) (hash(x)=49031231) +16000 val loss 6.1136 +16000 val perplexity 451.9574 +16000 train 6.252616 (lr=2.3831e-05) (hash(x)=58049587) +19900 val loss 6.0425 +19900 val perplexity 420.9256 +19900 train 6.074161 (lr=1.3900e-05) (hash(x)=54352451) +16100 val loss 6.1053 +16100 val perplexity 448.2053 +16100 train 5.723023 (lr=2.3760e-05) (hash(x)=39795825) +20000 val loss 6.0371 +20000 val perplexity 418.6718 +20000 train 5.848784 (lr=1.3846e-05) (hash(x)=45789621) +16200 val loss 6.1046 +16200 val perplexity 447.9215 +16200 train 6.088348 (lr=2.3688e-05) (hash(x)=49498689) +20100 val loss 6.0351 +20100 val perplexity 417.8513 +20100 train 6.016353 (lr=1.3792e-05) (hash(x)=45096720) +16300 val loss 6.1108 +16300 val perplexity 450.6829 +16300 train 6.431908 (lr=2.3615e-05) (hash(x)=51235194) +20200 val loss 6.0350 +20200 val perplexity 417.7830 +20200 train 6.111966 (lr=1.3738e-05) (hash(x)=49377287) +16400 val loss 6.1213 +16400 val perplexity 455.4350 +16400 train 5.985395 (lr=2.3543e-05) (hash(x)=47789417) +20300 val loss 6.0290 +20300 val perplexity 415.3068 +20300 train 5.838013 (lr=1.3684e-05) (hash(x)=49454716) +16500 val loss 6.0975 +16500 val perplexity 444.7294 +16500 train 6.658655 (lr=2.3470e-05) (hash(x)=59665889) +20400 val loss 6.0270 +20400 val perplexity 414.4816 +20400 train 5.529781 (lr=1.3630e-05) (hash(x)=38699401) +16600 val loss 6.0940 +16600 val perplexity 443.1777 +16600 train 6.001691 (lr=2.3397e-05) (hash(x)=49034893) +20500 val loss 6.0256 +20500 val perplexity 413.9007 +20500 train 6.123451 (lr=1.3575e-05) (hash(x)=53266981) +16700 val loss 6.0890 +16700 val perplexity 440.9671 +16700 train 5.932962 (lr=2.3324e-05) (hash(x)=45685955) +20600 val loss 6.0224 +20600 val perplexity 412.5876 +20600 train 5.985712 (lr=1.3521e-05) (hash(x)=51783804) +16800 val loss 6.0897 +16800 val perplexity 441.2748 +16800 train 5.888167 (lr=2.3250e-05) (hash(x)=49589835) +20700 val loss 6.0232 +20700 val perplexity 412.8987 +20700 train 5.945555 (lr=1.3466e-05) (hash(x)=51388467) +16900 val loss 6.0878 +16900 val perplexity 440.4301 +16900 train 5.503837 (lr=2.3176e-05) (hash(x)=38655565) +20800 val loss 6.0170 +20800 val perplexity 410.3474 +20800 train 5.763718 (lr=1.3412e-05) (hash(x)=45324903) +17000 val loss 6.0790 +17000 val perplexity 436.6061 +17000 train 6.026536 (lr=2.3102e-05) (hash(x)=46730732) +17100 val loss 6.0846 +17100 val perplexity 439.0268 +17100 train 5.723007 (lr=2.3028e-05) (hash(x)=43843722) +20900 val loss 6.0137 +20900 val perplexity 408.9897 +20900 train 5.663091 (lr=1.3357e-05) (hash(x)=45972467) +17200 val loss 6.0808 +17200 val perplexity 437.3953 +17200 train 5.851749 (lr=2.2953e-05) (hash(x)=50198978) +21000 val loss 6.0099 +21000 val perplexity 407.4515 +21000 train 5.966038 (lr=1.3302e-05) (hash(x)=46754507) +17300 val loss 6.0828 +17300 val perplexity 438.2347 +17300 train 6.101326 (lr=2.2878e-05) (hash(x)=52976276) +21100 val loss 6.0061 +21100 val perplexity 405.9081 +21100 train 6.342019 (lr=1.3247e-05) (hash(x)=59913667) +17400 val loss 6.0808 +17400 val perplexity 437.3946 +17400 train 5.918645 (lr=2.2803e-05) (hash(x)=52224275) +21200 val loss 6.0074 +21200 val perplexity 406.4099 +21200 train 5.992761 (lr=1.3192e-05) (hash(x)=46487759) +17500 val loss 6.0912 +17500 val perplexity 441.9562 +17500 train 5.885674 (lr=2.2727e-05) (hash(x)=49055112) +21300 val loss 6.0012 +21300 val perplexity 403.9193 +21300 train 5.781599 (lr=1.3137e-05) (hash(x)=45535495) +17600 val loss 6.0834 +17600 val perplexity 438.5305 +17600 train 6.120990 (lr=2.2652e-05) (hash(x)=49996498) +21400 val loss 5.9985 +21400 val perplexity 402.8162 +21400 train 6.021730 (lr=1.3082e-05) (hash(x)=54559895) +17700 val loss 6.0735 +17700 val perplexity 434.2019 +17700 train 6.088815 (lr=2.2576e-05) (hash(x)=50526074) +21500 val loss 5.9918 +21500 val perplexity 400.1187 +21500 train 5.835318 (lr=1.3026e-05) (hash(x)=48474176) +17800 val loss 6.0650 +17800 val perplexity 430.5317 +17800 train 5.957778 (lr=2.2500e-05) (hash(x)=48410966) +21600 val loss 5.9877 +21600 val perplexity 398.4915 +21600 train 5.955646 (lr=1.2971e-05) (hash(x)=47915584) +17900 val loss 6.0574 +17900 val perplexity 427.2758 +17900 train 5.958517 (lr=2.2423e-05) (hash(x)=51654550) +21700 val loss 5.9848 +21700 val perplexity 397.3612 +21700 train 5.740158 (lr=1.2916e-05) (hash(x)=45846696) +18000 val loss 6.0496 +18000 val perplexity 423.9271 +18000 train 6.091769 (lr=2.2346e-05) (hash(x)=44227619) +21800 val loss 5.9832 +21800 val perplexity 396.7269 +21800 train 5.746738 (lr=1.2860e-05) (hash(x)=47201143) +18100 val loss 6.0494 +18100 val perplexity 423.8782 +18100 train 5.845154 (lr=2.2270e-05) (hash(x)=45771090) +21900 val loss 5.9805 +21900 val perplexity 395.6386 +21900 train 5.920254 (lr=1.2805e-05) (hash(x)=50028395) +18200 val loss 6.0442 +18200 val perplexity 421.6790 +18200 train 5.670067 (lr=2.2192e-05) (hash(x)=37914739) +22000 val loss 5.9758 +22000 val perplexity 393.7770 +22000 train 5.972282 (lr=1.2749e-05) (hash(x)=50273781) +18300 val loss 6.0389 +18300 val perplexity 419.4438 +18300 train 6.493767 (lr=2.2115e-05) (hash(x)=68784189) +22100 val loss 5.9787 +22100 val perplexity 394.9455 +22100 train 5.922459 (lr=1.2693e-05) (hash(x)=47859152) +18400 val loss 6.0414 +18400 val perplexity 420.4859 +18400 train 6.166500 (lr=2.2038e-05) (hash(x)=55006900) +18500 val loss 6.0341 +18500 val perplexity 417.4431 +18500 train 6.481091 (lr=2.1960e-05) (hash(x)=56001248) +22200 val loss 5.9770 +22200 val perplexity 394.2448 +22200 train 5.806381 (lr=1.2637e-05) (hash(x)=47820296) +18600 val loss 6.0434 +18600 val perplexity 421.3397 +18600 train 6.216574 (lr=2.1882e-05) (hash(x)=51339352) +22300 val loss 5.9711 +22300 val perplexity 391.9242 +22300 train 5.694263 (lr=1.2581e-05) (hash(x)=51741508) +18700 val loss 6.0332 +18700 val perplexity 417.0340 +18700 train 6.174406 (lr=2.1804e-05) (hash(x)=53256371) +22400 val loss 5.9756 +22400 val perplexity 393.7068 +22400 train 5.555858 (lr=1.2526e-05) (hash(x)=44585679) +18800 val loss 6.0260 +18800 val perplexity 414.0494 +18800 train 5.994168 (lr=2.1725e-05) (hash(x)=51580096) +22500 val loss 5.9712 +22500 val perplexity 391.9800 +22500 train 6.050448 (lr=1.2470e-05) (hash(x)=48753202) +18900 val loss 6.0367 +18900 val perplexity 418.5020 +18900 train 6.105165 (lr=2.1647e-05) (hash(x)=53001073) +22600 val loss 5.9684 +22600 val perplexity 390.8818 +22600 train 5.555081 (lr=1.2414e-05) (hash(x)=45351902) +19000 val loss 6.0285 +19000 val perplexity 415.0747 +19000 train 5.876069 (lr=2.1568e-05) (hash(x)=51531298) +22700 val loss 5.9661 +22700 val perplexity 389.9694 +22700 train 6.446915 (lr=1.2357e-05) (hash(x)=59263323) +19100 val loss 6.0270 +19100 val perplexity 414.4510 +19100 train 5.812671 (lr=2.1489e-05) (hash(x)=48047406) +22800 val loss 5.9661 +22800 val perplexity 389.9673 +22800 train 5.517475 (lr=1.2301e-05) (hash(x)=44873637) +19200 val loss 6.0160 +19200 val perplexity 409.9267 +19200 train 5.828525 (lr=2.1409e-05) (hash(x)=44939100) +22900 val loss 5.9566 +22900 val perplexity 386.2760 +22900 train 6.126547 (lr=1.2245e-05) (hash(x)=58208930) +19300 val loss 6.0167 +19300 val perplexity 410.2137 +19300 train 5.882833 (lr=2.1330e-05) (hash(x)=50157419) +19400 val loss 6.0138 +19400 val perplexity 409.0353 +23000 val loss 5.9509 +23000 val perplexity 384.0978 +19400 train 5.840789 (lr=2.1250e-05) (hash(x)=46973593) +23000 train 5.938713 (lr=1.2189e-05) (hash(x)=50248612) +19500 val loss 6.0150 +19500 val perplexity 409.5139 +19500 train 5.830599 (lr=2.1171e-05) (hash(x)=44056368) +23100 val loss 5.9617 +23100 val perplexity 388.2698 +23100 train 5.866947 (lr=1.2133e-05) (hash(x)=47902220) +19600 val loss 6.0093 +19600 val perplexity 407.2095 +19600 train 6.286026 (lr=2.1091e-05) (hash(x)=58496552) +23200 val loss 5.9456 +23200 val perplexity 382.0500 +23200 train 5.819086 (lr=1.2076e-05) (hash(x)=45606218) +19700 val loss 6.0027 +19700 val perplexity 404.5251 +19700 train 5.873129 (lr=2.1010e-05) (hash(x)=47753232) +23300 val loss 5.9452 +23300 val perplexity 381.9192 +23300 train 6.057199 (lr=1.2020e-05) (hash(x)=55596334) +19800 val loss 5.9962 +19800 val perplexity 401.8822 +19800 train 5.818875 (lr=2.0930e-05) (hash(x)=49031231) +23400 val loss 5.9433 +23400 val perplexity 381.2005 +23400 train 5.974837 (lr=1.1963e-05) (hash(x)=59736985) +19900 val loss 5.9990 +19900 val perplexity 403.0160 +19900 train 6.032429 (lr=2.0850e-05) (hash(x)=54352451) +23500 val loss 5.9384 +23500 val perplexity 379.3220 +23500 train 5.880207 (lr=1.1907e-05) (hash(x)=48935694) +20000 val loss 5.9902 +20000 val perplexity 399.4755 +20000 train 5.821084 (lr=2.0769e-05) (hash(x)=45789621) +23600 val loss 5.9435 +23600 val perplexity 381.2614 +23600 train 5.748632 (lr=1.1850e-05) (hash(x)=53977113) +20100 val loss 5.9872 +20100 val perplexity 398.3021 +20100 train 5.979465 (lr=2.0688e-05) (hash(x)=45096720) +23700 val loss 5.9416 +23700 val perplexity 380.5549 +23700 train 5.571252 (lr=1.1794e-05) (hash(x)=45466992) +20200 val loss 5.9899 +20200 val perplexity 399.3816 +20200 train 6.072440 (lr=2.0607e-05) (hash(x)=49377287) +23800 val loss 5.9330 +23800 val perplexity 377.2707 +23800 train 5.722865 (lr=1.1737e-05) (hash(x)=45830878) +20300 val loss 5.9853 +20300 val perplexity 397.5463 +20300 train 5.777882 (lr=2.0526e-05) (hash(x)=49454716) +23900 val loss 5.9316 +23900 val perplexity 376.7677 +23900 train 5.815681 (lr=1.1681e-05) (hash(x)=48268066) +20400 val loss 5.9874 +20400 val perplexity 398.3587 +20400 train 5.489896 (lr=2.0444e-05) (hash(x)=38699401) +24000 val loss 5.9324 +24000 val perplexity 377.0595 +24000 train 5.763990 (lr=1.1624e-05) (hash(x)=51062942) +20500 val loss 5.9861 +20500 val perplexity 397.8539 +20500 train 6.082136 (lr=2.0363e-05) (hash(x)=53266981) +24100 val loss 5.9271 +24100 val perplexity 375.0807 +24100 train 5.821166 (lr=1.1567e-05) (hash(x)=51090246) +20600 val loss 5.9795 +20600 val perplexity 395.2306 +20600 train 5.927663 (lr=2.0281e-05) (hash(x)=51783804) +24200 val loss 5.9325 +24200 val perplexity 377.1023 +24200 train 5.644570 (lr=1.1511e-05) (hash(x)=48686225) +20700 val loss 5.9787 +20700 val perplexity 394.9346 +20700 train 5.894282 (lr=2.0199e-05) (hash(x)=51388467) +24300 val loss 5.9228 +24300 val perplexity 373.4643 +24300 train 5.596101 (lr=1.1454e-05) (hash(x)=43604814) +20800 val loss 5.9755 +20800 val perplexity 393.6576 +20800 train 5.733730 (lr=2.0117e-05) (hash(x)=45324903) +24400 val loss 5.9207 +24400 val perplexity 372.6592 +24400 train 5.720966 (lr=1.1397e-05) (hash(x)=48143655) +20900 val loss 5.9748 +20900 val perplexity 393.4082 +20900 train 5.639361 (lr=2.0035e-05) (hash(x)=45972467) +24500 val loss 5.9223 +24500 val perplexity 373.2679 +24500 train 6.007196 (lr=1.1341e-05) (hash(x)=58317397) +21000 val loss 5.9768 +21000 val perplexity 394.1708 +21000 train 5.946597 (lr=1.9953e-05) (hash(x)=46754507) +24600 val loss 5.9150 +24600 val perplexity 370.5453 +24600 train 6.046793 (lr=1.1284e-05) (hash(x)=53855425) +21100 val loss 5.9712 +21100 val perplexity 391.9688 +21100 train 6.286791 (lr=1.9871e-05) (hash(x)=59913667) +24700 val loss 5.9188 +24700 val perplexity 371.9542 +24700 train 5.737185 (lr=1.1227e-05) (hash(x)=49829785) +21200 val loss 5.9707 +21200 val perplexity 391.7911 +21200 train 5.959611 (lr=1.9788e-05) (hash(x)=46487759) +24800 val loss 5.9066 +24800 val perplexity 367.4540 +24800 train 5.966415 (lr=1.1170e-05) (hash(x)=51337298) +21300 val loss 5.9670 +21300 val perplexity 390.3396 +21300 train 5.756226 (lr=1.9705e-05) (hash(x)=45535495) +24900 val loss 5.9111 +24900 val perplexity 369.1197 +24900 train 5.948283 (lr=1.1114e-05) (hash(x)=47385279) +21400 val loss 5.9625 +21400 val perplexity 388.5822 +21400 train 5.974060 (lr=1.9623e-05) (hash(x)=54559895) +25000 val loss 5.9024 +25000 val perplexity 365.8966 +25000 train 6.364158 (lr=1.1057e-05) (hash(x)=53631975) +21500 val loss 5.9543 +21500 val perplexity 385.4239 +21500 train 5.809160 (lr=1.9540e-05) (hash(x)=48474176) +25100 val loss 5.8961 +25100 val perplexity 363.6031 +25100 train 5.909728 (lr=1.1000e-05) (hash(x)=52991852) +21600 val loss 5.9506 +21600 val perplexity 383.9960 +21600 train 5.916177 (lr=1.9457e-05) (hash(x)=47915584) +25200 val loss 5.8960 +25200 val perplexity 363.5669 +25200 train 6.299012 (lr=1.0943e-05) (hash(x)=51863290) +21700 val loss 5.9525 +21700 val perplexity 384.7280 +21700 train 5.705974 (lr=1.9373e-05) (hash(x)=45846696) +25300 val loss 5.8955 +25300 val perplexity 363.4105 +25300 train 5.676990 (lr=1.0886e-05) (hash(x)=46898810) +21800 val loss 5.9531 +21800 val perplexity 384.9510 +21800 train 5.724777 (lr=1.9290e-05) (hash(x)=47201143) +25400 val loss 5.8900 +25400 val perplexity 361.3885 +25400 train 5.981023 (lr=1.0830e-05) (hash(x)=50297951) +21900 val loss 5.9472 +21900 val perplexity 382.6832 +21900 train 5.879095 (lr=1.9207e-05) (hash(x)=50028395) +25500 val loss 5.8939 +25500 val perplexity 362.7997 +25500 train 5.731374 (lr=1.0773e-05) (hash(x)=49942718) +22000 val loss 5.9441 +22000 val perplexity 381.5149 +22000 train 5.940948 (lr=1.9123e-05) (hash(x)=50273781) +25600 val loss 5.8891 +25600 val perplexity 361.0835 +22100 val loss 5.9443 +22100 val perplexity 381.5700 +25600 train 5.774881 (lr=1.0716e-05) (hash(x)=46944979) +22100 train 5.887846 (lr=1.9040e-05) (hash(x)=47859152) +25700 val loss 5.8881 +25700 val perplexity 360.7173 +22200 val loss 5.9424 +22200 val perplexity 380.8488 +25700 train 6.158441 (lr=1.0659e-05) (hash(x)=48116500) +22200 train 5.783244 (lr=1.8956e-05) (hash(x)=47820296) +25800 val loss 5.8825 +25800 val perplexity 358.7125 +25800 train 6.141578 (lr=1.0603e-05) (hash(x)=59750893) +22300 val loss 5.9399 +22300 val perplexity 379.8859 +22300 train 5.649718 (lr=1.8872e-05) (hash(x)=51741508) +25900 val loss 5.8847 +25900 val perplexity 359.4806 +25900 train 5.587708 (lr=1.0546e-05) (hash(x)=49479938) +22400 val loss 5.9418 +22400 val perplexity 380.6195 +22400 train 5.514085 (lr=1.8788e-05) (hash(x)=44585679) +26000 val loss 5.8847 +26000 val perplexity 359.5068 +26000 train 6.015381 (lr=1.0489e-05) (hash(x)=51134577) +22500 val loss 5.9416 +22500 val perplexity 380.5318 +22500 train 6.013096 (lr=1.8704e-05) (hash(x)=48753202) +26100 val loss 5.8801 +26100 val perplexity 357.8409 +26100 train 5.793796 (lr=1.0433e-05) (hash(x)=53677080) +22600 val loss 5.9344 +22600 val perplexity 377.8289 +22600 train 5.539925 (lr=1.8620e-05) (hash(x)=45351902) +26200 val loss 5.8802 +26200 val perplexity 357.8738 +26200 train 5.769418 (lr=1.0376e-05) (hash(x)=48885091) +22700 val loss 5.9339 +22700 val perplexity 377.6348 +22700 train 6.382971 (lr=1.8536e-05) (hash(x)=59263323) +26300 val loss 5.8802 +26300 val perplexity 357.8973 +26300 train 5.970710 (lr=1.0319e-05) (hash(x)=57217953) +22800 val loss 5.9362 +22800 val perplexity 378.4873 +22800 train 5.511533 (lr=1.8452e-05) (hash(x)=44873637) +26400 val loss 5.8744 +26400 val perplexity 355.8020 +26400 train 5.687129 (lr=1.0263e-05) (hash(x)=48504109) +22900 val loss 5.9266 +22900 val perplexity 374.8903 +22900 train 6.089273 (lr=1.8368e-05) (hash(x)=58208930) +26500 val loss 5.8714 +26500 val perplexity 354.7402 +26500 train 6.492834 (lr=1.0206e-05) (hash(x)=55775918) +23000 val loss 5.9192 +23000 val perplexity 372.1071 +23000 train 5.920835 (lr=1.8283e-05) (hash(x)=50248612) +26600 val loss 5.8694 +26600 val perplexity 354.0326 +26600 train 5.856175 (lr=1.0150e-05) (hash(x)=51978817) +23100 val loss 5.9385 +23100 val perplexity 379.3830 +23100 train 5.831228 (lr=1.8199e-05) (hash(x)=47902220) +26700 val loss 5.8720 +26700 val perplexity 354.9430 +26700 train 5.674723 (lr=1.0093e-05) (hash(x)=50995814) +23200 val loss 5.9137 +23200 val perplexity 370.0719 +23200 train 5.790851 (lr=1.8114e-05) (hash(x)=45606218) +26800 val loss 5.8653 +26800 val perplexity 352.5807 +26800 train 5.775627 (lr=1.0037e-05) (hash(x)=49986643) +23300 val loss 5.9191 +23300 val perplexity 372.0711 +23300 train 6.040370 (lr=1.8030e-05) (hash(x)=55596334) +26900 val loss 5.8614 +26900 val perplexity 351.2286 +26900 train 5.932100 (lr=9.9802e-06) (hash(x)=51408788) +23400 val loss 5.9149 +23400 val perplexity 370.5341 +23400 train 5.936928 (lr=1.7945e-05) (hash(x)=59736985) +27000 val loss 5.8587 +27000 val perplexity 350.2613 +27000 train 5.618080 (lr=9.9238e-06) (hash(x)=49024980) +23500 val loss 5.9116 +23500 val perplexity 369.3007 +23500 train 5.853740 (lr=1.7860e-05) (hash(x)=48935694) +27100 val loss 5.8568 +27100 val perplexity 349.6035 +27100 train 6.146255 (lr=9.8675e-06) (hash(x)=45027677) +23600 val loss 5.9120 +23600 val perplexity 369.4622 +23600 train 5.701444 (lr=1.7776e-05) (hash(x)=53977113) +27200 val loss 5.8558 +27200 val perplexity 349.2647 +27200 train 5.934349 (lr=9.8112e-06) (hash(x)=49930937) +23700 val loss 5.9207 +23700 val perplexity 372.6779 +23700 train 5.541676 (lr=1.7691e-05) (hash(x)=45466992) +27300 val loss 5.8525 +27300 val perplexity 348.1022 +27300 train 5.779110 (lr=9.7549e-06) (hash(x)=51016030) +23800 val loss 5.9072 +23800 val perplexity 367.6752 +23800 train 5.699251 (lr=1.7606e-05) (hash(x)=45830878) +27400 val loss 5.8502 +27400 val perplexity 347.2988 +27400 train 6.109749 (lr=9.6987e-06) (hash(x)=48154914) +23900 val loss 5.9023 +23900 val perplexity 365.8708 +23900 train 5.800008 (lr=1.7521e-05) (hash(x)=48268066) +27500 val loss 5.8507 +27500 val perplexity 347.4839 +27500 train 5.962000 (lr=9.6426e-06) (hash(x)=53265478) +24000 val loss 5.8987 +24000 val perplexity 364.5544 +24000 train 5.722995 (lr=1.7436e-05) (hash(x)=51062942) +27600 val loss 5.8674 +27600 val perplexity 353.3300 +27600 train 5.699236 (lr=9.5865e-06) (hash(x)=39456920) +24100 val loss 5.9053 +24100 val perplexity 366.9691 +24100 train 5.797174 (lr=1.7351e-05) (hash(x)=51090246) +27700 val loss 5.8497 +27700 val perplexity 347.1281 +27700 train 6.052728 (lr=9.5304e-06) (hash(x)=55794844) +24200 val loss 5.9040 +24200 val perplexity 366.5029 +24200 train 5.629851 (lr=1.7266e-05) (hash(x)=48686225) +27800 val loss 5.8503 +27800 val perplexity 347.3496 +27800 train 5.772732 (lr=9.4745e-06) (hash(x)=49091362) +24300 val loss 5.8971 +24300 val perplexity 363.9623 +24300 train 5.569736 (lr=1.7181e-05) (hash(x)=43604814) +27900 val loss 5.8468 +27900 val perplexity 346.1378 +27900 train 5.671597 (lr=9.4185e-06) (hash(x)=49494590) +24400 val loss 5.8966 +24400 val perplexity 363.8072 +24400 train 5.702633 (lr=1.7096e-05) (hash(x)=48143655) +28000 val loss 5.8432 +28000 val perplexity 344.8905 +28000 train 5.238519 (lr=9.3627e-06) (hash(x)=42333356) +24500 val loss 5.8990 +24500 val perplexity 364.6841 +24500 train 5.981970 (lr=1.7011e-05) (hash(x)=58317397) +28100 val loss 5.8429 +28100 val perplexity 344.7632 +28100 train 6.091377 (lr=9.3069e-06) (hash(x)=49378958) +24600 val loss 5.8901 +24600 val perplexity 361.4359 +24600 train 6.016593 (lr=1.6926e-05) (hash(x)=53855425) +28200 val loss 5.8369 +28200 val perplexity 342.7185 +28200 train 6.032902 (lr=9.2511e-06) (hash(x)=56595308) +24700 val loss 5.9024 +24700 val perplexity 365.9040 +24700 train 5.708404 (lr=1.6841e-05) (hash(x)=49829785) +28300 val loss 5.8424 +28300 val perplexity 344.5992 +28300 train 6.435670 (lr=9.1955e-06) (hash(x)=52311024) +24800 val loss 5.8827 +24800 val perplexity 358.7901 +24800 train 5.948217 (lr=1.6755e-05) (hash(x)=51337298) +28400 val loss 5.8363 +28400 val perplexity 342.5097 +28400 train 5.790722 (lr=9.1399e-06) (hash(x)=47997766) +24900 val loss 5.8848 +24900 val perplexity 359.5360 +24900 train 5.926387 (lr=1.6670e-05) (hash(x)=47385279) +28500 val loss 5.8318 +28500 val perplexity 340.9885 +28500 train 5.800593 (lr=9.0844e-06) (hash(x)=46906589) +25000 val loss 5.8806 +25000 val perplexity 358.0083 +25000 train 6.329424 (lr=1.6585e-05) (hash(x)=53631975) +28600 val loss 5.8271 +28600 val perplexity 339.3832 +28600 train 5.635410 (lr=9.0290e-06) (hash(x)=44172312) +25100 val loss 5.8737 +25100 val perplexity 355.5683 +25100 train 5.890868 (lr=1.6500e-05) (hash(x)=52991852) +28700 val loss 5.8279 +28700 val perplexity 339.6490 +28700 train 5.914693 (lr=8.9736e-06) (hash(x)=50435687) +25200 val loss 5.8748 +25200 val perplexity 355.9369 +25200 train 6.262851 (lr=1.6415e-05) (hash(x)=51863290) +28800 val loss 5.8278 +28800 val perplexity 339.6092 +28800 train 5.766870 (lr=8.9183e-06) (hash(x)=53678477) +25300 val loss 5.8749 +25300 val perplexity 355.9756 +25300 train 5.655712 (lr=1.6330e-05) (hash(x)=46898810) +28900 val loss 5.8319 +28900 val perplexity 341.0195 +28900 train 5.668084 (lr=8.8631e-06) (hash(x)=47353092) +25400 val loss 5.8713 +25400 val perplexity 354.7136 +25400 train 5.961609 (lr=1.6245e-05) (hash(x)=50297951) +29000 val loss 5.8196 +29000 val perplexity 336.8515 +29000 train 6.044529 (lr=8.8080e-06) (hash(x)=56723899) +25500 val loss 5.8676 +25500 val perplexity 353.4156 +25500 train 5.712247 (lr=1.6159e-05) (hash(x)=49942718) +29100 val loss 5.8219 +29100 val perplexity 337.6258 +29100 train 5.860459 (lr=8.7530e-06) (hash(x)=56056600) +25600 val loss 5.8637 +25600 val perplexity 352.0114 +25600 train 5.759308 (lr=1.6074e-05) (hash(x)=46944979) +29200 val loss 5.8203 +29200 val perplexity 337.0878 +29200 train 5.774289 (lr=8.6981e-06) (hash(x)=50538322) +25700 val loss 5.8668 +25700 val perplexity 353.1143 +25700 train 6.134089 (lr=1.5989e-05) (hash(x)=48116500) +29300 val loss 5.8174 +29300 val perplexity 336.0903 +29300 train 5.952323 (lr=8.6432e-06) (hash(x)=57261305) +25800 val loss 5.8621 +25800 val perplexity 351.4714 +25800 train 6.118748 (lr=1.5904e-05) (hash(x)=59750893) +29400 val loss 5.8176 +29400 val perplexity 336.1701 +29400 train 5.653827 (lr=8.5885e-06) (hash(x)=49014077) +25900 val loss 5.8634 +25900 val perplexity 351.9072 +25900 train 5.604434 (lr=1.5819e-05) (hash(x)=49479938) +29500 val loss 5.8136 +29500 val perplexity 334.8350 +29500 train 5.560246 (lr=8.5338e-06) (hash(x)=48698820) +26000 val loss 5.8610 +26000 val perplexity 351.0605 +26000 train 5.995680 (lr=1.5734e-05) (hash(x)=51134577) +29600 val loss 5.8113 +29600 val perplexity 334.0497 +29600 train 5.671588 (lr=8.4793e-06) (hash(x)=48327445) +26100 val loss 5.8567 +26100 val perplexity 349.5738 +26100 train 5.762384 (lr=1.5649e-05) (hash(x)=53677080) +29700 val loss 5.8143 +29700 val perplexity 335.0526 +29700 train 5.950148 (lr=8.4248e-06) (hash(x)=52084503) +26200 val loss 5.8558 +26200 val perplexity 349.2388 +26200 train 5.745913 (lr=1.5564e-05) (hash(x)=48885091) +29800 val loss 5.8183 +29800 val perplexity 336.4054 +29800 train 5.637198 (lr=8.3705e-06) (hash(x)=44402411) +26300 val loss 5.8578 +26300 val perplexity 349.9416 +26300 train 5.937270 (lr=1.5479e-05) (hash(x)=57217953) +29900 val loss 5.8216 +29900 val perplexity 337.5128 +29900 train 5.490565 (lr=8.3162e-06) (hash(x)=43932550) +26400 val loss 5.8493 +26400 val perplexity 346.9803 +26400 train 5.636497 (lr=1.5394e-05) (hash(x)=48504109) +30000 val loss 5.8063 +30000 val perplexity 332.3831 +30000 train 5.663086 (lr=8.2621e-06) (hash(x)=48250169) +26500 val loss 5.8461 +26500 val perplexity 345.8705 +26500 train 6.452854 (lr=1.5309e-05) (hash(x)=55775918) +30100 val loss 5.8051 +30100 val perplexity 331.9740 +30100 train 5.903311 (lr=8.2080e-06) (hash(x)=46727253) +26600 val loss 5.8484 +26600 val perplexity 346.6755 +26600 train 5.833241 (lr=1.5224e-05) (hash(x)=51978817) +30200 val loss 5.7990 +30200 val perplexity 329.9583 +30200 train 5.835582 (lr=8.1541e-06) (hash(x)=52740773) +26700 val loss 5.8464 +26700 val perplexity 345.9814 +26700 train 5.652956 (lr=1.5140e-05) (hash(x)=50995814) +30300 val loss 5.7994 +30300 val perplexity 330.0945 +30300 train 5.852531 (lr=8.1003e-06) (hash(x)=52481310) +26800 val loss 5.8416 +26800 val perplexity 344.3252 +26800 train 5.760257 (lr=1.5055e-05) (hash(x)=49986643) +30400 val loss 5.8047 +30400 val perplexity 331.8644 +30400 train 5.745065 (lr=8.0466e-06) (hash(x)=44942827) +26900 val loss 5.8370 +26900 val perplexity 342.7338 +26900 train 5.931504 (lr=1.4970e-05) (hash(x)=51408788) +30500 val loss 5.7939 +30500 val perplexity 328.3029 +30500 train 5.695507 (lr=7.9931e-06) (hash(x)=53118706) +27000 val loss 5.8341 +27000 val perplexity 341.7437 +27000 train 5.587416 (lr=1.4886e-05) (hash(x)=49024980) +30600 val loss 5.7964 +30600 val perplexity 329.1194 +30600 train 5.512551 (lr=7.9396e-06) (hash(x)=46278699) +27100 val loss 5.8329 +27100 val perplexity 341.3468 +27100 train 6.119393 (lr=1.4801e-05) (hash(x)=45027677) +30700 val loss 5.7959 +30700 val perplexity 328.9410 +30700 train 5.768550 (lr=7.8863e-06) (hash(x)=51949086) +27200 val loss 5.8316 +27200 val perplexity 340.9177 +27200 train 5.909894 (lr=1.4717e-05) (hash(x)=49930937) +30800 val loss 5.7940 +30800 val perplexity 328.3263 +30800 train 5.815023 (lr=7.8331e-06) (hash(x)=50575709) +27300 val loss 5.8321 +27300 val perplexity 341.0592 +27300 train 5.749785 (lr=1.4632e-05) (hash(x)=51016030) +30900 val loss 5.7921 +30900 val perplexity 327.6891 +30900 train 5.623732 (lr=7.7800e-06) (hash(x)=53992116) +27400 val loss 5.8282 +27400 val perplexity 339.7318 +27400 train 6.084442 (lr=1.4548e-05) (hash(x)=48154914) +31000 val loss 5.7931 +31000 val perplexity 328.0347 +31000 train 5.610317 (lr=7.7270e-06) (hash(x)=44967737) +27500 val loss 5.8271 +27500 val perplexity 339.3614 +27500 train 5.934493 (lr=1.4464e-05) (hash(x)=53265478) +31100 val loss 5.7873 +31100 val perplexity 326.1444 +31100 train 5.495876 (lr=7.6742e-06) (hash(x)=41850245) +27600 val loss 5.8532 +27600 val perplexity 348.3305 +27600 train 5.673011 (lr=1.4380e-05) (hash(x)=39456920) +31200 val loss 5.7895 +31200 val perplexity 326.8626 +31200 train 5.649887 (lr=7.6215e-06) (hash(x)=49921656) +27700 val loss 5.8304 +27700 val perplexity 340.4877 +27700 train 6.012886 (lr=1.4296e-05) (hash(x)=55794844) +31300 val loss 5.7901 +31300 val perplexity 327.0437 +31300 train 5.638779 (lr=7.5690e-06) (hash(x)=47068006) +27800 val loss 5.8271 +27800 val perplexity 339.3675 +27800 train 5.719471 (lr=1.4212e-05) (hash(x)=49091362) +31400 val loss 5.7889 +31400 val perplexity 326.6466 +31400 train 5.854243 (lr=7.5165e-06) (hash(x)=52608679) +27900 val loss 5.8258 +27900 val perplexity 338.9278 +27900 train 5.662131 (lr=1.4128e-05) (hash(x)=49494590) +31500 val loss 5.7899 +31500 val perplexity 326.9759 +31500 train 5.740549 (lr=7.4643e-06) (hash(x)=51542050) +28000 val loss 5.8240 +28000 val perplexity 338.3145 +28000 train 5.203104 (lr=1.4044e-05) (hash(x)=42333356) +31600 val loss 5.7837 +31600 val perplexity 324.9550 +31600 train 5.683465 (lr=7.4121e-06) (hash(x)=49170260) +28100 val loss 5.8198 +28100 val perplexity 336.8958 +28100 train 6.023970 (lr=1.3960e-05) (hash(x)=49378958) +31700 val loss 5.7865 +31700 val perplexity 325.8862 +31700 train 5.794418 (lr=7.3601e-06) (hash(x)=57677051) +28200 val loss 5.8126 +28200 val perplexity 334.4845 +28200 train 6.005305 (lr=1.3877e-05) (hash(x)=56595308) +31800 val loss 5.7770 +31800 val perplexity 322.7915 +31800 train 5.951791 (lr=7.3083e-06) (hash(x)=52567943) +28300 val loss 5.8173 +28300 val perplexity 336.0581 +28300 train 6.452550 (lr=1.3793e-05) (hash(x)=52311024) +31900 val loss 5.7812 +31900 val perplexity 324.1601 +31900 train 5.979164 (lr=7.2566e-06) (hash(x)=51246750) +28400 val loss 5.8107 +28400 val perplexity 333.8564 +28400 train 5.784709 (lr=1.3710e-05) (hash(x)=47997766) +32000 val loss 5.7812 +32000 val perplexity 324.1569 +32000 train 6.239797 (lr=7.2050e-06) (hash(x)=63256210) +28500 val loss 5.8086 +28500 val perplexity 333.1524 +28500 train 5.772715 (lr=1.3627e-05) (hash(x)=46906589) +32100 val loss 5.7756 +32100 val perplexity 322.3339 +32100 train 5.996445 (lr=7.1536e-06) (hash(x)=52238178) +28600 val loss 5.8034 +28600 val perplexity 331.4316 +28600 train 5.617081 (lr=1.3543e-05) (hash(x)=44172312) +32200 val loss 5.7667 +32200 val perplexity 319.4783 +32200 train 6.019013 (lr=7.1024e-06) (hash(x)=56290278) +28700 val loss 5.8066 +28700 val perplexity 332.5017 +28700 train 5.899007 (lr=1.3460e-05) (hash(x)=50435687) +32300 val loss 5.7686 +32300 val perplexity 320.0852 +32300 train 5.821920 (lr=7.0513e-06) (hash(x)=48455511) +28800 val loss 5.8046 +28800 val perplexity 331.8189 +28800 train 5.720612 (lr=1.3377e-05) (hash(x)=53678477) +32400 val loss 5.7687 +32400 val perplexity 320.1248 +32400 train 5.572311 (lr=7.0003e-06) (hash(x)=42754343) +28900 val loss 5.8163 +28900 val perplexity 335.7429 +28900 train 5.639201 (lr=1.3295e-05) (hash(x)=47353092) +32500 val loss 5.7653 +32500 val perplexity 319.0257 +32500 train 5.743356 (lr=6.9495e-06) (hash(x)=51062421) +29000 val loss 5.7941 +29000 val perplexity 328.3599 +29000 train 6.015324 (lr=1.3212e-05) (hash(x)=56723899) +32600 val loss 5.7674 +32600 val perplexity 319.7191 +32600 train 5.579992 (lr=6.8989e-06) (hash(x)=43957725) +29100 val loss 5.8014 +29100 val perplexity 330.7534 +29100 train 5.835095 (lr=1.3129e-05) (hash(x)=56056600) +32700 val loss 5.7646 +32700 val perplexity 318.8035 +32700 train 5.816367 (lr=6.8485e-06) (hash(x)=55971557) +29200 val loss 5.7978 +29200 val perplexity 329.5623 +29200 train 5.749757 (lr=1.3047e-05) (hash(x)=50538322) +32800 val loss 5.7644 +32800 val perplexity 318.7572 +32800 train 5.801012 (lr=6.7982e-06) (hash(x)=49429775) +29300 val loss 5.7952 +29300 val perplexity 328.7195 +29300 train 5.871194 (lr=1.2965e-05) (hash(x)=57261305) +32900 val loss 5.7669 +32900 val perplexity 319.5330 +32900 train 5.646833 (lr=6.7480e-06) (hash(x)=49564803) +29400 val loss 5.8002 +29400 val perplexity 330.3651 +29400 train 5.639715 (lr=1.2883e-05) (hash(x)=49014077) +33000 val loss 5.7613 +33000 val perplexity 317.7588 +33000 train 5.621592 (lr=6.6981e-06) (hash(x)=46343736) +29500 val loss 5.7929 +29500 val perplexity 327.9675 +29500 train 5.534022 (lr=1.2801e-05) (hash(x)=48698820) +33100 val loss 5.7629 +33100 val perplexity 318.2655 +33100 train 5.465615 (lr=6.6483e-06) (hash(x)=45424347) +29600 val loss 5.7889 +29600 val perplexity 326.6507 +29600 train 5.633460 (lr=1.2719e-05) (hash(x)=48327445) +33200 val loss 5.7618 +33200 val perplexity 317.9326 +33200 train 5.821382 (lr=6.5987e-06) (hash(x)=50207861) +29700 val loss 5.7934 +29700 val perplexity 328.1129 +29700 train 5.949692 (lr=1.2637e-05) (hash(x)=52084503) +33300 val loss 5.7627 +33300 val perplexity 318.2148 +33300 train 5.666207 (lr=6.5493e-06) (hash(x)=45847922) +29800 val loss 5.7928 +29800 val perplexity 327.9302 +29800 train 5.602334 (lr=1.2556e-05) (hash(x)=44402411) +33400 val loss 5.7615 +33400 val perplexity 317.8156 +33400 train 5.625560 (lr=6.5000e-06) (hash(x)=44723617) +29900 val loss 5.7974 +29900 val perplexity 329.4268 +29900 train 5.472091 (lr=1.2474e-05) (hash(x)=43932550) +33500 val loss 5.7578 +33500 val perplexity 316.6458 +33500 train 5.868995 (lr=6.4509e-06) (hash(x)=54058855) +30000 val loss 5.7825 +30000 val perplexity 324.5602 +30000 train 5.627662 (lr=1.2393e-05) (hash(x)=48250169) +33600 val loss 5.7515 +33600 val perplexity 314.6570 +33600 train 5.742185 (lr=6.4020e-06) (hash(x)=53052866) +30100 val loss 5.7848 +30100 val perplexity 325.3123 +30100 train 5.877565 (lr=1.2312e-05) (hash(x)=46727253) +33700 val loss 5.7521 +33700 val perplexity 314.8468 +33700 train 5.696295 (lr=6.3533e-06) (hash(x)=50458295) +30200 val loss 5.7743 +30200 val perplexity 321.9292 +30200 train 5.830214 (lr=1.2231e-05) (hash(x)=52740773) +33800 val loss 5.7556 +33800 val perplexity 315.9640 +33800 train 5.922135 (lr=6.3048e-06) (hash(x)=58204533) +30300 val loss 5.7768 +30300 val perplexity 322.7395 +30300 train 5.825465 (lr=1.2150e-05) (hash(x)=52481310) +33900 val loss 5.7485 +33900 val perplexity 313.7070 +33900 train 5.896043 (lr=6.2564e-06) (hash(x)=52246445) +30400 val loss 5.7864 +30400 val perplexity 325.8445 +30400 train 5.725902 (lr=1.2070e-05) (hash(x)=44942827) +34000 val loss 5.7463 +34000 val perplexity 313.0226 +34000 train 5.689269 (lr=6.2083e-06) (hash(x)=49316883) +30500 val loss 5.7748 +30500 val perplexity 322.0930 +30500 train 5.652507 (lr=1.1990e-05) (hash(x)=53118706) +34100 val loss 5.7458 +34100 val perplexity 312.8819 +34100 train 5.401742 (lr=6.1603e-06) (hash(x)=42991129) +30600 val loss 5.7746 +30600 val perplexity 322.0168 +30600 train 5.487098 (lr=1.1909e-05) (hash(x)=46278699) +34200 val loss 5.7444 +34200 val perplexity 312.4464 +34200 train 5.374725 (lr=6.1125e-06) (hash(x)=45199853) +30700 val loss 5.7734 +30700 val perplexity 321.6157 +30700 train 5.740088 (lr=1.1829e-05) (hash(x)=51949086) +34300 val loss 5.7479 +34300 val perplexity 313.5393 +34300 train 5.656493 (lr=6.0650e-06) (hash(x)=48404386) +30800 val loss 5.7726 +30800 val perplexity 321.3735 +30800 train 5.779682 (lr=1.1750e-05) (hash(x)=50575709) +34400 val loss 5.7444 +34400 val perplexity 312.4415 +34400 train 5.554443 (lr=6.0176e-06) (hash(x)=45614997) +30900 val loss 5.7698 +30900 val perplexity 320.4705 +30900 train 5.606625 (lr=1.1670e-05) (hash(x)=53992116) +34500 val loss 5.7430 +34500 val perplexity 312.0042 +34500 train 5.804786 (lr=5.9704e-06) (hash(x)=53157684) +31000 val loss 5.7708 +31000 val perplexity 320.7980 +31000 train 5.580514 (lr=1.1591e-05) (hash(x)=44967737) +34600 val loss 5.7460 +34600 val perplexity 312.9328 +34600 train 5.944212 (lr=5.9234e-06) (hash(x)=57152287) +31100 val loss 5.7713 +31100 val perplexity 320.9418 +31100 train 5.473865 (lr=1.1511e-05) (hash(x)=41850245) +34700 val loss 5.7445 +34700 val perplexity 312.4519 +34700 train 5.200546 (lr=5.8766e-06) (hash(x)=46088213) +31200 val loss 5.7690 +31200 val perplexity 320.2273 +31200 train 5.626186 (lr=1.1432e-05) (hash(x)=49921656) +34800 val loss 5.7412 +34800 val perplexity 311.4271 +34800 train 5.489526 (lr=5.8301e-06) (hash(x)=47098930) +31300 val loss 5.7700 +31300 val perplexity 320.5365 +31300 train 5.617104 (lr=1.1353e-05) (hash(x)=47068006) +34900 val loss 5.7422 +34900 val perplexity 311.7371 +34900 train 5.772087 (lr=5.7837e-06) (hash(x)=51033119) +31400 val loss 5.7653 +31400 val perplexity 319.0359 +31400 train 5.857385 (lr=1.1275e-05) (hash(x)=52608679) +35000 val loss 5.7426 +35000 val perplexity 311.8882 +35000 train 5.422277 (lr=5.7375e-06) (hash(x)=47111881) +31500 val loss 5.7670 +31500 val perplexity 319.5877 +31500 train 5.695386 (lr=1.1196e-05) (hash(x)=51542050) +35100 val loss 5.7393 +35100 val perplexity 310.8597 +35100 train 5.785703 (lr=5.6916e-06) (hash(x)=52899306) +31600 val loss 5.7661 +31600 val perplexity 319.2773 +31600 train 5.645914 (lr=1.1118e-05) (hash(x)=49170260) +35200 val loss 5.7364 +35200 val perplexity 309.9406 +35200 train 5.680959 (lr=5.6458e-06) (hash(x)=49124834) +31700 val loss 5.7576 +31700 val perplexity 316.6001 +31700 train 5.762600 (lr=1.1040e-05) (hash(x)=57677051) +35300 val loss 5.7347 +35300 val perplexity 309.4172 +35300 train 5.892675 (lr=5.6003e-06) (hash(x)=54127282) +31800 val loss 5.7554 +31800 val perplexity 315.8761 +31800 train 5.918190 (lr=1.0962e-05) (hash(x)=52567943) +35400 val loss 5.7359 +35400 val perplexity 309.8057 +35400 train 5.620642 (lr=5.5550e-06) (hash(x)=51154858) +31900 val loss 5.7568 +31900 val perplexity 316.3446 +31900 train 5.969959 (lr=1.0885e-05) (hash(x)=51246750) +35500 val loss 5.7286 +35500 val perplexity 307.5465 +35500 train 5.816206 (lr=5.5099e-06) (hash(x)=51585534) +32000 val loss 5.7598 +32000 val perplexity 317.2903 +32000 train 6.202359 (lr=1.0808e-05) (hash(x)=63256210) +35600 val loss 5.7270 +35600 val perplexity 307.0443 +35600 train 5.533147 (lr=5.4650e-06) (hash(x)=46228695) +32100 val loss 5.7544 +32100 val perplexity 315.5753 +32100 train 5.969201 (lr=1.0730e-05) (hash(x)=52238178) +35700 val loss 5.7284 +35700 val perplexity 307.4717 +35700 train 5.687887 (lr=5.4204e-06) (hash(x)=49080481) +32200 val loss 5.7469 +32200 val perplexity 313.2162 +32200 train 5.999879 (lr=1.0654e-05) (hash(x)=56290278) +35800 val loss 5.7282 +35800 val perplexity 307.4243 +35800 train 5.903237 (lr=5.3759e-06) (hash(x)=56743084) +32300 val loss 5.7464 +32300 val perplexity 313.0519 +32300 train 5.808727 (lr=1.0577e-05) (hash(x)=48455511) +35900 val loss 5.7277 +35900 val perplexity 307.2527 +35900 train 5.862816 (lr=5.3317e-06) (hash(x)=53068229) +32400 val loss 5.7467 +32400 val perplexity 313.1571 +32400 train 5.567654 (lr=1.0500e-05) (hash(x)=42754343) +36000 val loss 5.7247 +36000 val perplexity 306.3492 +36000 train 5.909879 (lr=5.2877e-06) (hash(x)=54363049) +32500 val loss 5.7437 +32500 val perplexity 312.2157 +32500 train 5.747748 (lr=1.0424e-05) (hash(x)=51062421) +36100 val loss 5.7261 +36100 val perplexity 306.7718 +36100 train 5.648923 (lr=5.2440e-06) (hash(x)=50555767) +32600 val loss 5.7473 +32600 val perplexity 313.3425 +32600 train 5.549796 (lr=1.0348e-05) (hash(x)=43957725) +36200 val loss 5.7241 +36200 val perplexity 306.1559 +36200 train 5.901261 (lr=5.2004e-06) (hash(x)=52103710) +32700 val loss 5.7427 +32700 val perplexity 311.8927 +32700 train 5.779653 (lr=1.0273e-05) (hash(x)=55971557) +36300 val loss 5.7246 +36300 val perplexity 306.3088 +36300 train 5.545993 (lr=5.1571e-06) (hash(x)=43347374) +32800 val loss 5.7436 +32800 val perplexity 312.1911 +32800 train 5.759826 (lr=1.0197e-05) (hash(x)=49429775) +36400 val loss 5.7209 +36400 val perplexity 305.1776 +36400 train 5.777555 (lr=5.1141e-06) (hash(x)=52661268) +36500 val loss 5.7228 +36500 val perplexity 305.7747 +36500 train 5.641724 (lr=5.0712e-06) (hash(x)=47765383) +32900 val loss 5.7443 +32900 val perplexity 312.3911 +32900 train 5.617378 (lr=1.0122e-05) (hash(x)=49564803) +36600 val loss 5.7208 +36600 val perplexity 305.1348 +36600 train 5.589707 (lr=5.0286e-06) (hash(x)=49302411) +33000 val loss 5.7409 +33000 val perplexity 311.3361 +33000 train 5.619483 (lr=1.0047e-05) (hash(x)=46343736) +36700 val loss 5.7228 +36700 val perplexity 305.7614 +36700 train 5.295547 (lr=4.9863e-06) (hash(x)=44113066) +33100 val loss 5.7395 +33100 val perplexity 310.9238 +33100 train 5.428693 (lr=9.9725e-06) (hash(x)=45424347) +36800 val loss 5.7222 +36800 val perplexity 305.5866 +36800 train 5.613403 (lr=4.9441e-06) (hash(x)=55896927) +33200 val loss 5.7422 +33200 val perplexity 311.7471 +33200 train 5.810764 (lr=9.8980e-06) (hash(x)=50207861) +36900 val loss 5.7211 +36900 val perplexity 305.2525 +36900 train 5.530878 (lr=4.9023e-06) (hash(x)=49812554) +33300 val loss 5.7406 +33300 val perplexity 311.2474 +33300 train 5.652924 (lr=9.8239e-06) (hash(x)=45847922) +37000 val loss 5.7207 +37000 val perplexity 305.1256 +37000 train 5.581516 (lr=4.8606e-06) (hash(x)=49391771) +33400 val loss 5.7362 +33400 val perplexity 309.8747 +33400 train 5.596319 (lr=9.7500e-06) (hash(x)=44723617) +37100 val loss 5.7152 +37100 val perplexity 303.4531 +37100 train 5.666396 (lr=4.8192e-06) (hash(x)=50515328) +33500 val loss 5.7361 +33500 val perplexity 309.8577 +33500 train 5.847381 (lr=9.6764e-06) (hash(x)=54058855) +37200 val loss 5.7170 +37200 val perplexity 303.9865 +37200 train 5.918746 (lr=4.7781e-06) (hash(x)=57788407) +33600 val loss 5.7300 +33600 val perplexity 307.9594 +33600 train 5.706859 (lr=9.6030e-06) (hash(x)=53052866) +37300 val loss 5.7121 +37300 val perplexity 302.4960 +37300 train 5.471431 (lr=4.7372e-06) (hash(x)=46333526) +33700 val loss 5.7309 +33700 val perplexity 308.2530 +33700 train 5.661858 (lr=9.5300e-06) (hash(x)=50458295) +37400 val loss 5.7107 +37400 val perplexity 302.0852 +37400 train 5.544876 (lr=4.6965e-06) (hash(x)=45862038) +33800 val loss 5.7337 +33800 val perplexity 309.0993 +33800 train 5.901291 (lr=9.4572e-06) (hash(x)=58204533) +37500 val loss 5.7108 +37500 val perplexity 302.1224 +37500 train 5.429259 (lr=4.6561e-06) (hash(x)=46543051) +33900 val loss 5.7263 +33900 val perplexity 306.8349 +33900 train 5.889362 (lr=9.3846e-06) (hash(x)=52246445) +37600 val loss 5.7086 +37600 val perplexity 301.4545 +37600 train 5.666259 (lr=4.6160e-06) (hash(x)=51314450) +34000 val loss 5.7217 +34000 val perplexity 305.4088 +34000 train 5.668020 (lr=9.3124e-06) (hash(x)=49316883) +37700 val loss 5.7082 +37700 val perplexity 301.3421 +37700 train 5.569125 (lr=4.5761e-06) (hash(x)=46706763) +34100 val loss 5.7232 +34100 val perplexity 305.8786 +34100 train 5.371028 (lr=9.2405e-06) (hash(x)=42991129) +37800 val loss 5.7094 +37800 val perplexity 301.6876 +37800 train 5.588169 (lr=4.5365e-06) (hash(x)=45793085) +34200 val loss 5.7198 +34200 val perplexity 304.8291 +34200 train 5.360204 (lr=9.1688e-06) (hash(x)=45199853) +37900 val loss 5.7087 +37900 val perplexity 301.4929 +37900 train 5.575988 (lr=4.4971e-06) (hash(x)=49969493) +34300 val loss 5.7223 +34300 val perplexity 305.6010 +34300 train 5.615972 (lr=9.0974e-06) (hash(x)=48404386) +38000 val loss 5.7067 +38000 val perplexity 300.8715 +38000 train 5.509633 (lr=4.4580e-06) (hash(x)=47184489) +34400 val loss 5.7209 +34400 val perplexity 305.1680 +34400 train 5.532707 (lr=9.0264e-06) (hash(x)=45614997) +38100 val loss 5.7071 +38100 val perplexity 300.9880 +38100 train 6.141479 (lr=4.4191e-06) (hash(x)=54905742) +34500 val loss 5.7213 +34500 val perplexity 305.3103 +34500 train 5.766829 (lr=8.9556e-06) (hash(x)=53157684) +38200 val loss 5.7053 +38200 val perplexity 300.4463 +38200 train 5.980143 (lr=4.3805e-06) (hash(x)=48831719) +34600 val loss 5.7256 +34600 val perplexity 306.6061 +34600 train 5.929462 (lr=8.8851e-06) (hash(x)=57152287) +38300 val loss 5.7069 +38300 val perplexity 300.9468 +38300 train 5.585613 (lr=4.3422e-06) (hash(x)=48944700) +38400 val loss 5.7068 +38400 val perplexity 300.8973 +38400 train 5.328422 (lr=4.3041e-06) (hash(x)=46504848) +34700 val loss 5.7203 +34700 val perplexity 304.9876 +34700 train 5.166498 (lr=8.8150e-06) (hash(x)=46088213) +38500 val loss 5.7047 +38500 val perplexity 300.2755 +38500 train 5.464530 (lr=4.2663e-06) (hash(x)=50148948) +34800 val loss 5.7175 +34800 val perplexity 304.1377 +34800 train 5.473274 (lr=8.7451e-06) (hash(x)=47098930) +38600 val loss 5.7069 +38600 val perplexity 300.9461 +38600 train 6.515060 (lr=4.2288e-06) (hash(x)=60660913) +34900 val loss 5.7198 +34900 val perplexity 304.8422 +34900 train 5.725802 (lr=8.6755e-06) (hash(x)=51033119) +38700 val loss 5.7042 +38700 val perplexity 300.1372 +38700 train 5.524274 (lr=4.1915e-06) (hash(x)=50256252) +35000 val loss 5.7189 +35000 val perplexity 304.5674 +35000 train 5.393657 (lr=8.6063e-06) (hash(x)=47111881) +38800 val loss 5.7027 +38800 val perplexity 299.6721 +38800 train 5.762375 (lr=4.1545e-06) (hash(x)=52592415) +35100 val loss 5.7163 +35100 val perplexity 303.7693 +35100 train 5.743896 (lr=8.5374e-06) (hash(x)=52899306) +38900 val loss 5.7012 +38900 val perplexity 299.2174 +38900 train 5.667720 (lr=4.1178e-06) (hash(x)=51945430) +35200 val loss 5.7157 +35200 val perplexity 303.5977 +35200 train 5.660658 (lr=8.4688e-06) (hash(x)=49124834) +39000 val loss 5.6987 +39000 val perplexity 298.4764 +39000 train 5.957221 (lr=4.0813e-06) (hash(x)=51838869) +35300 val loss 5.7152 +35300 val perplexity 303.4402 +35300 train 5.881164 (lr=8.4005e-06) (hash(x)=54127282) +39100 val loss 5.6955 +39100 val perplexity 297.5403 +39100 train 5.859197 (lr=4.0452e-06) (hash(x)=54556054) +35400 val loss 5.7137 +35400 val perplexity 302.9942 +35400 train 5.595296 (lr=8.3325e-06) (hash(x)=51154858) +39200 val loss 5.6956 +39200 val perplexity 297.5639 +39200 train 6.051885 (lr=4.0093e-06) (hash(x)=48915509) +35500 val loss 5.7110 +35500 val perplexity 302.1741 +35500 train 5.770454 (lr=8.2648e-06) (hash(x)=51585534) +39300 val loss 5.6972 +39300 val perplexity 298.0456 +39300 train 5.509531 (lr=3.9736e-06) (hash(x)=46121589) +35600 val loss 5.7078 +35600 val perplexity 301.1989 +35600 train 5.506782 (lr=8.1975e-06) (hash(x)=46228695) +39400 val loss 5.6960 +39400 val perplexity 297.6610 +39400 train 5.247723 (lr=3.9383e-06) (hash(x)=45556201) +35700 val loss 5.7084 +35700 val perplexity 301.4003 +35700 train 5.678492 (lr=8.1305e-06) (hash(x)=49080481) +39500 val loss 5.6958 +39500 val perplexity 297.6191 +39500 train 5.748965 (lr=3.9032e-06) (hash(x)=47014749) +35800 val loss 5.7027 +35800 val perplexity 299.6689 +35800 train 5.873546 (lr=8.0639e-06) (hash(x)=56743084) +39600 val loss 5.6986 +39600 val perplexity 298.4635 +39600 train 5.512006 (lr=3.8685e-06) (hash(x)=47395114) +35900 val loss 5.7053 +35900 val perplexity 300.4590 +35900 train 5.825395 (lr=7.9976e-06) (hash(x)=53068229) +39700 val loss 5.6988 +39700 val perplexity 298.5094 +39700 train 5.838912 (lr=3.8340e-06) (hash(x)=50858008) +36000 val loss 5.7041 +36000 val perplexity 300.0976 +36000 train 5.896332 (lr=7.9316e-06) (hash(x)=54363049) +39800 val loss 5.6952 +39800 val perplexity 297.4215 +39800 train 5.438666 (lr=3.7998e-06) (hash(x)=44096434) +36100 val loss 5.7047 +36100 val perplexity 300.2639 +36100 train 5.629881 (lr=7.8659e-06) (hash(x)=50555767) +39900 val loss 5.6937 +39900 val perplexity 297.0037 +39900 train 5.444809 (lr=3.7659e-06) (hash(x)=46902151) +36200 val loss 5.7012 +36200 val perplexity 299.2270 +36200 train 5.885538 (lr=7.8006e-06) (hash(x)=52103710) +40000 val loss 5.6921 +40000 val perplexity 296.5062 +40000 train 5.521812 (lr=3.7322e-06) (hash(x)=46735297) +36300 val loss 5.7044 +36300 val perplexity 300.1891 +36300 train 5.518615 (lr=7.7357e-06) (hash(x)=43347374) +40100 val loss 5.6941 +40100 val perplexity 297.0978 +40100 train 5.652420 (lr=3.6989e-06) (hash(x)=48035170) +36400 val loss 5.6986 +36400 val perplexity 298.4595 +36400 train 5.739610 (lr=7.6711e-06) (hash(x)=52661268) +40200 val loss 5.6923 +40200 val perplexity 296.5833 +40200 train 5.530097 (lr=3.6658e-06) (hash(x)=46651322) +36500 val loss 5.7040 +36500 val perplexity 300.0547 +36500 train 5.621748 (lr=7.6068e-06) (hash(x)=47765383) +40300 val loss 5.6922 +40300 val perplexity 296.5413 +40300 train 5.567995 (lr=3.6331e-06) (hash(x)=46378099) +36600 val loss 5.7022 +36600 val perplexity 299.5312 +36600 train 5.556540 (lr=7.5429e-06) (hash(x)=49302411) +40400 val loss 5.6876 +40400 val perplexity 295.1811 +40400 train 5.546243 (lr=3.6006e-06) (hash(x)=46495438) +36700 val loss 5.7019 +36700 val perplexity 299.4502 +36700 train 5.268260 (lr=7.4794e-06) (hash(x)=44113066) +40500 val loss 5.6839 +40500 val perplexity 294.0993 +40500 train 5.864549 (lr=3.5684e-06) (hash(x)=52059147) +36800 val loss 5.6989 +36800 val perplexity 298.5392 +36800 train 5.538425 (lr=7.4162e-06) (hash(x)=55896927) +40600 val loss 5.6846 +40600 val perplexity 294.2888 +40600 train 5.845756 (lr=3.5365e-06) (hash(x)=54885045) +36900 val loss 5.6990 +36900 val perplexity 298.5684 +36900 train 5.504206 (lr=7.3534e-06) (hash(x)=49812554) +40700 val loss 5.6837 +40700 val perplexity 294.0358 +40700 train 6.034006 (lr=3.5050e-06) (hash(x)=53213971) +37000 val loss 5.7009 +37000 val perplexity 299.1427 +37000 train 5.553427 (lr=7.2909e-06) (hash(x)=49391771) +40800 val loss 5.6804 +40800 val perplexity 293.0755 +40800 train 5.355559 (lr=3.4737e-06) (hash(x)=45133794) +37100 val loss 5.6955 +37100 val perplexity 297.5160 +37100 train 5.644794 (lr=7.2289e-06) (hash(x)=50515328) +40900 val loss 5.6809 +40900 val perplexity 293.2118 +40900 train 5.816621 (lr=3.4427e-06) (hash(x)=56546547) +37200 val loss 5.6974 +37200 val perplexity 298.0884 +37200 train 5.905831 (lr=7.1671e-06) (hash(x)=57788407) +41000 val loss 5.6790 +41000 val perplexity 292.6622 +41000 train 6.063414 (lr=3.4120e-06) (hash(x)=49552270) +37300 val loss 5.6918 +37300 val perplexity 296.4171 +37300 train 5.450413 (lr=7.1058e-06) (hash(x)=46333526) +41100 val loss 5.6815 +41100 val perplexity 293.3907 +41100 train 5.909220 (lr=3.3816e-06) (hash(x)=51222370) +37400 val loss 5.6904 +37400 val perplexity 296.0257 +37400 train 5.508546 (lr=7.0448e-06) (hash(x)=45862038) +41200 val loss 5.6818 +41200 val perplexity 293.4826 +41200 train 5.702371 (lr=3.3516e-06) (hash(x)=50883608) +37500 val loss 5.6900 +37500 val perplexity 295.8817 +37500 train 5.394752 (lr=6.9842e-06) (hash(x)=46543051) +41300 val loss 5.6809 +41300 val perplexity 293.2199 +41300 train 5.950223 (lr=3.3218e-06) (hash(x)=52996828) +37600 val loss 5.6881 +37600 val perplexity 295.3298 +37600 train 5.648445 (lr=6.9240e-06) (hash(x)=51314450) +41400 val loss 5.6797 +41400 val perplexity 292.8702 +41400 train 5.635095 (lr=3.2923e-06) (hash(x)=48822716) +37700 val loss 5.6875 +37700 val perplexity 295.1576 +37700 train 5.557153 (lr=6.8642e-06) (hash(x)=46706763) +41500 val loss 5.6835 +41500 val perplexity 293.9640 +41500 train 5.678239 (lr=3.2632e-06) (hash(x)=50936577) +37800 val loss 5.6881 +37800 val perplexity 295.3330 +37800 train 5.559156 (lr=6.8047e-06) (hash(x)=45793085) +41600 val loss 5.6811 +41600 val perplexity 293.2666 +41600 train 5.753641 (lr=3.2343e-06) (hash(x)=44375046) +37900 val loss 5.6892 +37900 val perplexity 295.6505 +37900 train 5.532322 (lr=6.7456e-06) (hash(x)=49969493) +41700 val loss 5.6789 +41700 val perplexity 292.6385 +41700 train 5.455396 (lr=3.2058e-06) (hash(x)=44060021) +38000 val loss 5.6871 +38000 val perplexity 295.0320 +38000 train 5.484699 (lr=6.6870e-06) (hash(x)=47184489) +41800 val loss 5.6796 +41800 val perplexity 292.8382 +41800 train 6.004835 (lr=3.1775e-06) (hash(x)=57765221) +38100 val loss 5.6846 +38100 val perplexity 294.2978 +38100 train 6.139813 (lr=6.6287e-06) (hash(x)=54905742) +41900 val loss 5.6773 +41900 val perplexity 292.1704 +41900 train 5.459730 (lr=3.1496e-06) (hash(x)=46051470) +38200 val loss 5.6848 +38200 val perplexity 294.3666 +38200 train 5.964708 (lr=6.5708e-06) (hash(x)=48831719) +42000 val loss 5.6808 +42000 val perplexity 293.1853 +42000 train 5.906244 (lr=3.1220e-06) (hash(x)=52077616) +38300 val loss 5.6846 +38300 val perplexity 294.3119 +38300 train 5.530490 (lr=6.5133e-06) (hash(x)=48944700) +42100 val loss 5.6776 +42100 val perplexity 292.2441 +42100 train 5.537674 (lr=3.0947e-06) (hash(x)=47845199) +38400 val loss 5.6847 +38400 val perplexity 294.3256 +38400 train 5.301461 (lr=6.4562e-06) (hash(x)=46504848) +42200 val loss 5.6755 +42200 val perplexity 291.6210 +42200 train 5.981801 (lr=3.0677e-06) (hash(x)=51549823) +38500 val loss 5.6842 +38500 val perplexity 294.1839 +38500 train 5.425443 (lr=6.3995e-06) (hash(x)=50148948) +42300 val loss 5.6729 +42300 val perplexity 290.8887 +42300 train 5.833851 (lr=3.0411e-06) (hash(x)=56922131) +38600 val loss 5.6821 +38600 val perplexity 293.5729 +38600 train 6.499399 (lr=6.3431e-06) (hash(x)=60660913) +42400 val loss 5.6729 +42400 val perplexity 290.8855 +42400 train 5.627751 (lr=3.0147e-06) (hash(x)=49004372) +38700 val loss 5.6828 +38700 val perplexity 293.7617 +38700 train 5.483240 (lr=6.2872e-06) (hash(x)=50256252) +42500 val loss 5.6715 +42500 val perplexity 290.4590 +42500 train 5.661749 (lr=2.9887e-06) (hash(x)=50651839) +38800 val loss 5.6825 +38800 val perplexity 293.6813 +38800 train 5.756256 (lr=6.2318e-06) (hash(x)=52592415) +42600 val loss 5.6700 +42600 val perplexity 290.0420 +42600 train 5.865386 (lr=2.9630e-06) (hash(x)=50767721) +38900 val loss 5.6796 +38900 val perplexity 292.8368 +38900 train 5.647102 (lr=6.1767e-06) (hash(x)=51945430) +42700 val loss 5.6700 +42700 val perplexity 290.0218 +42700 train 5.382096 (lr=2.9376e-06) (hash(x)=49099183) +39000 val loss 5.6794 +39000 val perplexity 292.7779 +39000 train 5.944742 (lr=6.1220e-06) (hash(x)=51838869) +42800 val loss 5.6692 +42800 val perplexity 289.7902 +42800 train 5.861118 (lr=2.9125e-06) (hash(x)=42272413) +39100 val loss 5.6747 +39100 val perplexity 291.3890 +39100 train 5.826692 (lr=6.0677e-06) (hash(x)=54556054) +42900 val loss 5.6685 +42900 val perplexity 289.5898 +42900 train 5.485934 (lr=2.8878e-06) (hash(x)=48582863) +39200 val loss 5.6754 +39200 val perplexity 291.6077 +39200 train 6.062328 (lr=6.0139e-06) (hash(x)=48915509) +43000 val loss 5.6673 +43000 val perplexity 289.2658 +43000 train 5.683171 (lr=2.8633e-06) (hash(x)=48703446) +39300 val loss 5.6748 +39300 val perplexity 291.4175 +39300 train 5.504872 (lr=5.9605e-06) (hash(x)=46121589) +43100 val loss 5.6670 +43100 val perplexity 289.1513 +43100 train 5.588244 (lr=2.8392e-06) (hash(x)=48730321) +39400 val loss 5.6737 +39400 val perplexity 291.1223 +39400 train 5.219228 (lr=5.9075e-06) (hash(x)=45556201) +43200 val loss 5.6675 +43200 val perplexity 289.3157 +43200 train 5.942907 (lr=2.8155e-06) (hash(x)=56536090) +39500 val loss 5.6757 +39500 val perplexity 291.6861 +39500 train 5.727135 (lr=5.8549e-06) (hash(x)=47014749) +43300 val loss 5.6663 +43300 val perplexity 288.9573 +43300 train 5.837591 (lr=2.7920e-06) (hash(x)=54154116) +39600 val loss 5.6781 +39600 val perplexity 292.4022 +39600 train 5.483907 (lr=5.8027e-06) (hash(x)=47395114) +43400 val loss 5.6707 +43400 val perplexity 290.2275 +43400 train 5.563474 (lr=2.7689e-06) (hash(x)=50058055) +39700 val loss 5.6783 +39700 val perplexity 292.4517 +39700 train 5.815925 (lr=5.7510e-06) (hash(x)=50858008) +43500 val loss 5.6659 +43500 val perplexity 288.8568 +43500 train 5.581208 (lr=2.7461e-06) (hash(x)=48743802) +39800 val loss 5.6736 +39800 val perplexity 291.0880 +39800 train 5.404604 (lr=5.6997e-06) (hash(x)=44096434) +43600 val loss 5.6659 +43600 val perplexity 288.8497 +43600 train 5.308034 (lr=2.7236e-06) (hash(x)=42792886) +39900 val loss 5.6735 +39900 val perplexity 291.0581 +39900 train 5.423934 (lr=5.6488e-06) (hash(x)=46902151) +43700 val loss 5.6652 +43700 val perplexity 288.6570 +43700 train 6.046514 (lr=2.7015e-06) (hash(x)=56446070) +40000 val loss 5.6722 +40000 val perplexity 290.6612 +40000 train 5.499574 (lr=5.5983e-06) (hash(x)=46735297) +43800 val loss 5.6656 +43800 val perplexity 288.7483 +43800 train 5.503555 (lr=2.6797e-06) (hash(x)=45584354) +40100 val loss 5.6751 +40100 val perplexity 291.5037 +40100 train 5.643946 (lr=5.5483e-06) (hash(x)=48035170) +43900 val loss 5.6643 +43900 val perplexity 288.3962 +43900 train 5.533401 (lr=2.6582e-06) (hash(x)=49339253) +40200 val loss 5.6711 +40200 val perplexity 290.3429 +40200 train 5.499719 (lr=5.4987e-06) (hash(x)=46651322) +44000 val loss 5.6642 +44000 val perplexity 288.3560 +44000 train 5.504177 (lr=2.6370e-06) (hash(x)=46183203) +40300 val loss 5.6713 +40300 val perplexity 290.4194 +40300 train 5.530605 (lr=5.4496e-06) (hash(x)=46378099) +44100 val loss 5.6609 +44100 val perplexity 287.3964 +44100 train 5.657050 (lr=2.6162e-06) (hash(x)=47849630) +40400 val loss 5.6668 +40400 val perplexity 289.1123 +40400 train 5.513836 (lr=5.4009e-06) (hash(x)=46495438) +44200 val loss 5.6618 +44200 val perplexity 287.6685 +44200 train 6.005238 (lr=2.5957e-06) (hash(x)=49834275) +40500 val loss 5.6643 +40500 val perplexity 288.3794 +44300 val loss 5.6596 +44300 val perplexity 287.0203 +40500 train 5.840748 (lr=5.3526e-06) (hash(x)=52059147) +44300 train 6.130713 (lr=2.5756e-06) (hash(x)=62535257) +40600 val loss 5.6636 +40600 val perplexity 288.1708 +40600 train 5.816614 (lr=5.3048e-06) (hash(x)=54885045) +44400 val loss 5.6566 +44400 val perplexity 286.1799 +44400 train 5.708674 (lr=2.5558e-06) (hash(x)=49253957) +40700 val loss 5.6626 +40700 val perplexity 287.9027 +44500 val loss 5.6595 +44500 val perplexity 287.0050 +40700 train 6.054996 (lr=5.2574e-06) (hash(x)=53213971) +44500 train 5.652122 (lr=2.5363e-06) (hash(x)=55368339) +40800 val loss 5.6605 +40800 val perplexity 287.3036 +44600 val loss 5.6580 +44600 val perplexity 286.5688 +40800 train 5.333339 (lr=5.2105e-06) (hash(x)=45133794) +44600 train 5.710472 (lr=2.5172e-06) (hash(x)=47098476) +40900 val loss 5.6600 +40900 val perplexity 287.1600 +40900 train 5.805818 (lr=5.1641e-06) (hash(x)=56546547) +44700 val loss 5.6581 +44700 val perplexity 286.6038 +44700 train 5.560949 (lr=2.4984e-06) (hash(x)=48280562) +41000 val loss 5.6596 +41000 val perplexity 287.0235 +41000 train 6.051268 (lr=5.1180e-06) (hash(x)=49552270) +44800 val loss 5.6573 +44800 val perplexity 286.3790 +44800 train 5.986833 (lr=2.4799e-06) (hash(x)=55591638) +41100 val loss 5.6606 +41100 val perplexity 287.3127 +41100 train 5.906196 (lr=5.0725e-06) (hash(x)=51222370) +44900 val loss 5.6573 +44900 val perplexity 286.3782 +44900 train 5.797924 (lr=2.4618e-06) (hash(x)=53757748) +41200 val loss 5.6618 +41200 val perplexity 287.6550 +41200 train 5.683946 (lr=5.0273e-06) (hash(x)=50883608) +45000 val loss 5.6558 +45000 val perplexity 285.9333 +45000 train 5.785798 (lr=2.4440e-06) (hash(x)=51685087) +41300 val loss 5.6619 +41300 val perplexity 287.6934 +41300 train 5.912921 (lr=4.9827e-06) (hash(x)=52996828) +45100 val loss 5.6547 +45100 val perplexity 285.6379 +45100 train 5.680123 (lr=2.4266e-06) (hash(x)=50093774) +41400 val loss 5.6608 +41400 val perplexity 287.3891 +41400 train 5.608542 (lr=4.9385e-06) (hash(x)=48822716) +45200 val loss 5.6550 +45200 val perplexity 285.7112 +45200 train 5.412447 (lr=2.4095e-06) (hash(x)=43460450) +41500 val loss 5.6617 +41500 val perplexity 287.6374 +41500 train 5.648216 (lr=4.8947e-06) (hash(x)=50936577) +45300 val loss 5.6565 +45300 val perplexity 286.1377 +45300 train 5.858303 (lr=2.3927e-06) (hash(x)=49935488) +41600 val loss 5.6615 +41600 val perplexity 287.5863 +41600 train 5.730962 (lr=4.8515e-06) (hash(x)=44375046) +45400 val loss 5.6560 +45400 val perplexity 286.0119 +45400 train 5.647653 (lr=2.3763e-06) (hash(x)=49447929) +41700 val loss 5.6584 +41700 val perplexity 286.6945 +41700 train 5.431281 (lr=4.8087e-06) (hash(x)=44060021) +45500 val loss 5.6548 +45500 val perplexity 285.6649 +45500 train 5.651358 (lr=2.3602e-06) (hash(x)=50713904) +41800 val loss 5.6599 +41800 val perplexity 287.1296 +41800 train 5.983839 (lr=4.7663e-06) (hash(x)=57765221) +45600 val loss 5.6728 +45600 val perplexity 290.8596 +45600 train 5.614160 (lr=2.3445e-06) (hash(x)=47674606) +41900 val loss 5.6565 +41900 val perplexity 286.1500 +41900 train 5.436088 (lr=4.7244e-06) (hash(x)=46051470) +45700 val loss 5.6550 +45700 val perplexity 285.7089 +45700 train 5.576493 (lr=2.3291e-06) (hash(x)=51539617) +42000 val loss 5.6578 +42000 val perplexity 286.5139 +42000 train 5.869827 (lr=4.6830e-06) (hash(x)=52077616) +45800 val loss 5.6551 +45800 val perplexity 285.7513 +45800 train 5.579877 (lr=2.3141e-06) (hash(x)=44448785) +42100 val loss 5.6568 +42100 val perplexity 286.2314 +42100 train 5.500535 (lr=4.6421e-06) (hash(x)=47845199) +45900 val loss 5.6542 +45900 val perplexity 285.4986 +45900 train 5.488136 (lr=2.2994e-06) (hash(x)=51499105) +42200 val loss 5.6541 +42200 val perplexity 285.4507 +42200 train 5.937322 (lr=4.6016e-06) (hash(x)=51549823) +46000 val loss 5.6512 +46000 val perplexity 284.6299 +46000 train 5.581034 (lr=2.2850e-06) (hash(x)=48359464) +42300 val loss 5.6513 +42300 val perplexity 284.6518 +42300 train 5.802900 (lr=4.5616e-06) (hash(x)=56922131) +46100 val loss 5.6528 +46100 val perplexity 285.0784 +46100 train 5.752692 (lr=2.2710e-06) (hash(x)=51885986) +46200 val loss 5.6496 +46200 val perplexity 284.1775 +46200 train 6.221810 (lr=2.2574e-06) (hash(x)=65186615) +42400 val loss 5.6511 +42400 val perplexity 284.6078 +42400 train 5.610093 (lr=4.5221e-06) (hash(x)=49004372) +46300 val loss 5.6496 +46300 val perplexity 284.1659 +46300 train 5.643756 (lr=2.2441e-06) (hash(x)=49626999) +42500 val loss 5.6496 +42500 val perplexity 284.1840 +42500 train 5.639362 (lr=4.4830e-06) (hash(x)=50651839) +46400 val loss 5.6496 +46400 val perplexity 284.1683 +46400 train 5.571313 (lr=2.2311e-06) (hash(x)=43325701) +42600 val loss 5.6487 +42600 val perplexity 283.9241 +42600 train 5.890523 (lr=4.4445e-06) (hash(x)=50767721) +46500 val loss 5.6483 +46500 val perplexity 283.7958 +46500 train 5.964220 (lr=2.2185e-06) (hash(x)=54028595) +42700 val loss 5.6491 +42700 val perplexity 284.0499 +42700 train 5.349964 (lr=4.4064e-06) (hash(x)=49099183) +46600 val loss 5.6486 +46600 val perplexity 283.8989 +46600 train 5.479692 (lr=2.2062e-06) (hash(x)=44519175) +42800 val loss 5.6472 +42800 val perplexity 283.5035 +42800 train 5.959209 (lr=4.3688e-06) (hash(x)=42272413) +46700 val loss 5.6496 +46700 val perplexity 284.1763 +46700 train 5.672090 (lr=2.1943e-06) (hash(x)=48357998) +42900 val loss 5.6484 +42900 val perplexity 283.8392 +42900 train 5.443577 (lr=4.3316e-06) (hash(x)=48582863) +46800 val loss 5.6489 +46800 val perplexity 283.9762 +46800 train 5.872787 (lr=2.1828e-06) (hash(x)=55911353) +43000 val loss 5.6450 +43000 val perplexity 282.8856 +43000 train 5.648489 (lr=4.2950e-06) (hash(x)=48703446) +46900 val loss 5.6477 +46900 val perplexity 283.6317 +46900 train 5.590394 (lr=2.1716e-06) (hash(x)=47897187) +43100 val loss 5.6459 +43100 val perplexity 283.1208 +43100 train 5.586102 (lr=4.2589e-06) (hash(x)=48730321) +47000 val loss 5.6473 +47000 val perplexity 283.5148 +47000 train 5.184170 (lr=2.1607e-06) (hash(x)=43196571) +43200 val loss 5.6465 +43200 val perplexity 283.2866 +43200 train 5.909928 (lr=4.2232e-06) (hash(x)=56536090) +47100 val loss 5.6469 +47100 val perplexity 283.3986 +47100 train 5.506231 (lr=2.1502e-06) (hash(x)=51224987) +43300 val loss 5.6451 +43300 val perplexity 282.8889 +43300 train 5.803907 (lr=4.1880e-06) (hash(x)=54154116) +47200 val loss 5.6481 +47200 val perplexity 283.7391 +47200 train 5.334624 (lr=2.1400e-06) (hash(x)=47943697) +43400 val loss 5.6488 +43400 val perplexity 283.9452 +43400 train 5.537530 (lr=4.1533e-06) (hash(x)=50058055) +47300 val loss 5.6471 +47300 val perplexity 283.4691 +47300 train 5.570846 (lr=2.1302e-06) (hash(x)=47351003) +43500 val loss 5.6468 +43500 val perplexity 283.3955 +43500 train 5.546353 (lr=4.1191e-06) (hash(x)=48743802) +47400 val loss 5.6471 +47400 val perplexity 283.4613 +47400 train 5.665713 (lr=2.1208e-06) (hash(x)=55562243) +43600 val loss 5.6474 +43600 val perplexity 283.5404 +43600 train 5.278276 (lr=4.0854e-06) (hash(x)=42792886) +47500 val loss 5.6475 +47500 val perplexity 283.5900 +47500 train 5.499609 (lr=2.1117e-06) (hash(x)=53544850) +43700 val loss 5.6444 +43700 val perplexity 282.7097 +43700 train 6.064008 (lr=4.0522e-06) (hash(x)=56446070) +47600 val loss 5.6465 +47600 val perplexity 283.2854 +47600 train 5.851956 (lr=2.1030e-06) (hash(x)=43634907) +43800 val loss 5.6454 +43800 val perplexity 282.9943 +43800 train 5.476852 (lr=4.0195e-06) (hash(x)=45584354) +47700 val loss 5.6451 +47700 val perplexity 282.8987 +47700 train 5.649112 (lr=2.0946e-06) (hash(x)=47909383) +43900 val loss 5.6466 +43900 val perplexity 283.3298 +43900 train 5.498499 (lr=3.9873e-06) (hash(x)=49339253) +47800 val loss 5.6434 +47800 val perplexity 282.4189 +44000 val loss 5.6447 +44000 val perplexity 282.7755 +47800 train 5.531595 (lr=2.0865e-06) (hash(x)=45871079) +44000 train 5.471801 (lr=3.9556e-06) (hash(x)=46183203) +44100 val loss 5.6407 +44100 val perplexity 281.6720 +47900 val loss 5.6430 +47900 val perplexity 282.3196 +44100 train 5.648279 (lr=3.9243e-06) (hash(x)=47849630) +47900 train 5.672690 (lr=2.0789e-06) (hash(x)=47333324) +44200 val loss 5.6419 +44200 val perplexity 282.0052 +44200 train 5.976722 (lr=3.8936e-06) (hash(x)=49834275) +48000 val loss 5.6422 +48000 val perplexity 282.0783 +48000 train 5.759875 (lr=2.0715e-06) (hash(x)=52758020) +44300 val loss 5.6382 +44300 val perplexity 280.9696 +44300 train 6.134610 (lr=3.8634e-06) (hash(x)=62535257) +48100 val loss 5.6394 +48100 val perplexity 281.2993 +48100 train 5.653091 (lr=2.0646e-06) (hash(x)=49806349) +44400 val loss 5.6365 +44400 val perplexity 280.4667 +44400 train 5.690608 (lr=3.8337e-06) (hash(x)=49253957) +48200 val loss 5.6404 +48200 val perplexity 281.5673 +48200 train 5.694753 (lr=2.0580e-06) (hash(x)=53220839) +44500 val loss 5.6362 +44500 val perplexity 280.4003 +44500 train 5.609005 (lr=3.8045e-06) (hash(x)=55368339) +48300 val loss 5.6420 +48300 val perplexity 282.0179 +48300 train 5.696584 (lr=2.0517e-06) (hash(x)=56052541) +44600 val loss 5.6382 +44600 val perplexity 280.9584 +44600 train 5.762136 (lr=3.7758e-06) (hash(x)=47098476) +48400 val loss 5.6391 +48400 val perplexity 281.2224 +48400 train 5.537630 (lr=2.0458e-06) (hash(x)=44482356) +44700 val loss 5.6378 +44700 val perplexity 280.8406 +44700 train 5.538322 (lr=3.7476e-06) (hash(x)=48280562) +48500 val loss 5.6394 +48500 val perplexity 281.2876 +48500 train 5.326652 (lr=2.0403e-06) (hash(x)=45714818) +44800 val loss 5.6371 +44800 val perplexity 280.6385 +44800 train 5.949906 (lr=3.7199e-06) (hash(x)=55591638) +48600 val loss 5.6399 +48600 val perplexity 281.4334 +48600 train 5.420673 (lr=2.0351e-06) (hash(x)=49476556) +44900 val loss 5.6371 +44900 val perplexity 280.6382 +44900 train 5.778909 (lr=3.6927e-06) (hash(x)=53757748) +48700 val loss 5.6406 +48700 val perplexity 281.6422 +48700 train 5.370193 (lr=2.0302e-06) (hash(x)=42508579) +45000 val loss 5.6349 +45000 val perplexity 280.0444 +48800 val loss 5.6401 +48800 val perplexity 281.4903 +48800 train 5.861522 (lr=2.0258e-06) (hash(x)=52737449) +45000 train 5.803207 (lr=3.6660e-06) (hash(x)=51685087) +45100 val loss 5.6354 +45100 val perplexity 280.1590 +45100 train 5.648643 (lr=3.6398e-06) (hash(x)=50093774) +48900 val loss 5.6402 +48900 val perplexity 281.5067 +48900 train 5.502568 (lr=2.0217e-06) (hash(x)=47057569) +45200 val loss 5.6352 +45200 val perplexity 280.1136 +45200 train 5.383186 (lr=3.6142e-06) (hash(x)=43460450) +49000 val loss 5.6407 +49000 val perplexity 281.6610 +49000 train 5.629939 (lr=2.0179e-06) (hash(x)=49908975) +45300 val loss 5.6382 +45300 val perplexity 280.9558 +45300 train 5.865899 (lr=3.5891e-06) (hash(x)=49935488) +49100 val loss 5.6403 +49100 val perplexity 281.5545 +49100 train 5.463228 (lr=2.0145e-06) (hash(x)=48427414) +49200 val loss 5.6398 +49200 val perplexity 281.4078 +49200 train 5.453797 (lr=2.0115e-06) (hash(x)=50246074) +45400 val loss 5.6367 +45400 val perplexity 280.5479 +45400 train 5.633832 (lr=3.5644e-06) (hash(x)=49447929) +49300 val loss 5.6388 +49300 val perplexity 281.1304 +49300 train 5.673634 (lr=2.0088e-06) (hash(x)=47715359) +45500 val loss 5.6347 +45500 val perplexity 279.9761 +45500 train 5.603483 (lr=3.5403e-06) (hash(x)=50713904) +49400 val loss 5.6407 +49400 val perplexity 281.6634 +49400 train 5.668647 (lr=2.0064e-06) (hash(x)=50175867) +45600 val loss 5.6575 +45600 val perplexity 286.4227 +45600 train 5.589744 (lr=3.5167e-06) (hash(x)=47674606) +49500 val loss 5.6433 +49500 val perplexity 282.3814 +49500 train 5.493413 (lr=2.0045e-06) (hash(x)=49336040) +45700 val loss 5.6359 +45700 val perplexity 280.3044 +45700 train 5.544065 (lr=3.4936e-06) (hash(x)=51539617) +49600 val loss 5.6427 +49600 val perplexity 282.2176 +49600 train 5.702200 (lr=2.0029e-06) (hash(x)=52039357) +45800 val loss 5.6347 +45800 val perplexity 279.9731 +45800 train 5.587180 (lr=3.4711e-06) (hash(x)=44448785) +49700 val loss 5.6435 +49700 val perplexity 282.4498 +49700 train 5.481979 (lr=2.0016e-06) (hash(x)=47568707) +45900 val loss 5.6354 +45900 val perplexity 280.1649 +45900 train 5.460416 (lr=3.4490e-06) (hash(x)=51499105) +49800 val loss 5.6427 +49800 val perplexity 282.2162 +49800 train 5.572678 (lr=2.0007e-06) (hash(x)=48451274) +46000 val loss 5.6313 +46000 val perplexity 279.0161 +46000 train 5.556747 (lr=3.4275e-06) (hash(x)=48359464) +49900 val loss 5.6446 +49900 val perplexity 282.7643 +49900 train 5.333293 (lr=2.0002e-06) (hash(x)=44523603) +46100 val loss 5.6334 +46100 val perplexity 279.6065 +46100 train 5.709271 (lr=3.4065e-06) (hash(x)=51885986) +49999 val loss 5.6378 +49999 val perplexity 280.8474 +46200 val loss 5.6293 +46200 val perplexity 278.4709 +46200 train 6.181342 (lr=3.3860e-06) (hash(x)=65186615) +46300 val loss 5.6282 +46300 val perplexity 278.1739 +46300 train 5.606453 (lr=3.3661e-06) (hash(x)=49626999) +46400 val loss 5.6281 +46400 val perplexity 278.1389 +46400 train 5.550458 (lr=3.3466e-06) (hash(x)=43325701) +46500 val loss 5.6286 +46500 val perplexity 278.2797 +46500 train 5.936493 (lr=3.3277e-06) (hash(x)=54028595) +46600 val loss 5.6273 +46600 val perplexity 277.9013 +46600 train 5.477195 (lr=3.3093e-06) (hash(x)=44519175) +46700 val loss 5.6282 +46700 val perplexity 278.1648 +46700 train 5.664310 (lr=3.2915e-06) (hash(x)=48357998) +46800 val loss 5.6273 +46800 val perplexity 277.9197 +46800 train 5.810308 (lr=3.2741e-06) (hash(x)=55911353) +46900 val loss 5.6287 +46900 val perplexity 278.2995 +46900 train 5.547019 (lr=3.2573e-06) (hash(x)=47897187) +47000 val loss 5.6277 +47000 val perplexity 278.0203 +47000 train 5.161955 (lr=3.2410e-06) (hash(x)=43196571) +47100 val loss 5.6272 +47100 val perplexity 277.8806 +47100 train 5.466049 (lr=3.2253e-06) (hash(x)=51224987) +47200 val loss 5.6285 +47200 val perplexity 278.2572 +47200 train 5.323849 (lr=3.2101e-06) (hash(x)=47943697) +47300 val loss 5.6279 +47300 val perplexity 278.0871 +47300 train 5.559111 (lr=3.1954e-06) (hash(x)=47351003) +47400 val loss 5.6270 +47400 val perplexity 277.8231 +47400 train 5.655081 (lr=3.1812e-06) (hash(x)=55562243) +47500 val loss 5.6286 +47500 val perplexity 278.2666 +47500 train 5.455912 (lr=3.1675e-06) (hash(x)=53544850) +47600 val loss 5.6274 +47600 val perplexity 277.9441 +47600 train 5.831614 (lr=3.1544e-06) (hash(x)=43634907) +47700 val loss 5.6252 +47700 val perplexity 277.3192 +47700 train 5.618991 (lr=3.1419e-06) (hash(x)=47909383) +47800 val loss 5.6246 +47800 val perplexity 277.1566 +47800 train 5.519455 (lr=3.1298e-06) (hash(x)=45871079) +47900 val loss 5.6240 +47900 val perplexity 276.9862 +47900 train 5.634283 (lr=3.1183e-06) (hash(x)=47333324) +48000 val loss 5.6250 +48000 val perplexity 277.2649 +48000 train 5.725096 (lr=3.1073e-06) (hash(x)=52758020) +48100 val loss 5.6217 +48100 val perplexity 276.3683 +48100 train 5.630651 (lr=3.0969e-06) (hash(x)=49806349) +48200 val loss 5.6209 +48200 val perplexity 276.1285 +48200 train 5.680059 (lr=3.0869e-06) (hash(x)=53220839) +48300 val loss 5.6216 +48300 val perplexity 276.3374 +48300 train 5.660135 (lr=3.0776e-06) (hash(x)=56052541) +48400 val loss 5.6200 +48400 val perplexity 275.8965 +48400 train 5.522215 (lr=3.0687e-06) (hash(x)=44482356) +48500 val loss 5.6219 +48500 val perplexity 276.4207 +48500 train 5.311810 (lr=3.0604e-06) (hash(x)=45714818) +48600 val loss 5.6208 +48600 val perplexity 276.1157 +48600 train 5.379840 (lr=3.0526e-06) (hash(x)=49476556) +48700 val loss 5.6205 +48700 val perplexity 276.0317 +48700 train 5.337888 (lr=3.0454e-06) (hash(x)=42508579) +48800 val loss 5.6216 +48800 val perplexity 276.3209 +48800 train 5.845224 (lr=3.0387e-06) (hash(x)=52737449) diff --git a/attention_kindselective_n_heads4_seed1340/model_02500.pt b/attention_kindselective_n_heads4_seed1340/model_02500.pt index 3cf751cc9e33d65094f0dd08a89d86bdc4aaf31e..f85f4fe1c4100d75d62f99005f0c94d0e2d0bb5c 100644 --- a/attention_kindselective_n_heads4_seed1340/model_02500.pt +++ b/attention_kindselective_n_heads4_seed1340/model_02500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:e856700b2ba8c1c78fec371f511980a9165c3a304e65b1e2ecf19f9a9ba2c4f5 +oid sha256:d40b44be7123ddcbe8c8270c95b84e3da6036f8ad4a89f29bd78174771f5ad1d size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_05000.pt b/attention_kindselective_n_heads4_seed1340/model_05000.pt index 08b7cf75b1cce0c930bbeed9bfff44cf6bcdb302..1bfc4975eab66497f19052eecded0a4f3d524995 100644 --- a/attention_kindselective_n_heads4_seed1340/model_05000.pt +++ b/attention_kindselective_n_heads4_seed1340/model_05000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:6ce0f91188a61cf2a20045bdc8d3dc3070278a30ac99975978bf3d232dd03f70 +oid sha256:64114b9e470f27de31f0af97063b74e5cb302fb1c4b95c1bab4f07ef2f27c50e size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_07500.pt b/attention_kindselective_n_heads4_seed1340/model_07500.pt index 9c3c64ec881d0b8e2e69a7be6470c8d9ce8e59dd..f23ad6c9feaaf1d8d68938c9beab620d553a30d5 100644 --- a/attention_kindselective_n_heads4_seed1340/model_07500.pt +++ b/attention_kindselective_n_heads4_seed1340/model_07500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:21edaf2540d7212d65d0bd210419f7d6a0f9281cde1d06f864c3ec1d63fff124 +oid sha256:e359763ad74be085ba664e365dd25cebe3d59184a9895012e9a3e79d28815537 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_10000.pt b/attention_kindselective_n_heads4_seed1340/model_10000.pt index 9cc3538de065058608ccc9c3a3f01d7df72073fe..08f3b74b0bd14ea1885c5b1de00efdac336e89f0 100644 --- a/attention_kindselective_n_heads4_seed1340/model_10000.pt +++ b/attention_kindselective_n_heads4_seed1340/model_10000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9e28d809d0626f302c7278753b8c13072c86cc80bea7caeb91264050622dd5e7 +oid sha256:acfc53c9a7cd4ecd4786fe1dbfd12b3665d56e5218f464f0a7377d3c19cc0571 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_12500.pt b/attention_kindselective_n_heads4_seed1340/model_12500.pt index ab3086e97cac460a59bea10823e6b1ce18ee4bcc..e1dd39ca431f09b92b1aba8f36fa27a499331191 100644 --- a/attention_kindselective_n_heads4_seed1340/model_12500.pt +++ b/attention_kindselective_n_heads4_seed1340/model_12500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:69a6cd3908266914a289ba86b8a7d2af8ce51bf4034f6b670c6730502d6f65d4 +oid sha256:869376a3b0bb3fde527f196c220c30e7f871fd914115e19e80c0b37bf4c442fd size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_15000.pt b/attention_kindselective_n_heads4_seed1340/model_15000.pt new file mode 100644 index 0000000000000000000000000000000000000000..9d83dc3cad3378303db84f869db4d46f4dbdc480 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_15000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:915e146b8111649e6a6db269b7778691b7cbe651ee1e60dec23d29627b76fd80 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_17500.pt b/attention_kindselective_n_heads4_seed1340/model_17500.pt new file mode 100644 index 0000000000000000000000000000000000000000..b47f6b464256b416b0fc0a8d4489d4b463e5d7df --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_17500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cce67e01256c5cce1cde6f7cc2d691f251068070c6622f2e4497418d105cfdf7 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_20000.pt b/attention_kindselective_n_heads4_seed1340/model_20000.pt new file mode 100644 index 0000000000000000000000000000000000000000..d9bd2bd285080cdf7d2db997107756bd1763be60 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_20000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e58a9162de1095f0dc094f589088a4a286f14df34fde95304844bd261b137114 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_22500.pt b/attention_kindselective_n_heads4_seed1340/model_22500.pt new file mode 100644 index 0000000000000000000000000000000000000000..2669e9b05754dd5192841a7d8dad155b94b09c1e --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_22500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6cb8425cc4b92ac619610aa095caad0181fb1e64a22bdf07e0f8a478d4fbc637 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_25000.pt b/attention_kindselective_n_heads4_seed1340/model_25000.pt new file mode 100644 index 0000000000000000000000000000000000000000..e55c329dd201ef1e2aae80fc0ddfa809791a8c1a --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_25000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7a54319fac2ccb5018c7a6c9d11a35d9c18c453909cbc138ac59131dc99fb576 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_27500.pt b/attention_kindselective_n_heads4_seed1340/model_27500.pt new file mode 100644 index 0000000000000000000000000000000000000000..d7fd00cb5fab3fb0fbb32cf1785b0fb74f19ee30 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_27500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:178b2ef1de395feb11e65b597ad37d5e8925e9d5a1e95e955a6707aea00e122f +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_30000.pt b/attention_kindselective_n_heads4_seed1340/model_30000.pt new file mode 100644 index 0000000000000000000000000000000000000000..2108d2fae698f54447418eb9ff11bb0f912372bc --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_30000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7f23f9d5154f322670c4460e9fe3c45b108edee73a84595627941c922acd3bf4 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_32500.pt b/attention_kindselective_n_heads4_seed1340/model_32500.pt new file mode 100644 index 0000000000000000000000000000000000000000..53c918c43651da11c3d821e836c7b605e46fbda5 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_32500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c72a4c4302438f826476f9fd4f4472b9054a5b02e723684d14f8c5a41b51e5d +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_35000.pt b/attention_kindselective_n_heads4_seed1340/model_35000.pt new file mode 100644 index 0000000000000000000000000000000000000000..6d29c28a99d05aa83b2c02690cc9629fe3f043d2 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_35000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f29591ab1e09c43505e5fdc647dd4e9d21abfe6a0ab7f50ab47f9fe4207045ac +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_37500.pt b/attention_kindselective_n_heads4_seed1340/model_37500.pt new file mode 100644 index 0000000000000000000000000000000000000000..6d4bee712c22202045c169fffd6c953df4307528 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_37500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:06a37d0898b5b62e808aaefe9b89f988cf90d4fb9c10d5fd38410275bfecbc21 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_40000.pt b/attention_kindselective_n_heads4_seed1340/model_40000.pt new file mode 100644 index 0000000000000000000000000000000000000000..f5e05dfe66605829f7798312261e45d475bd8ccf --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/model_40000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:03aa4bc431bade8c18c5c59922094723549f04f741f4366931664738dd354a17 +size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_42500.pt b/attention_kindselective_n_heads4_seed1340/model_42500.pt index 1582a7515effc8f9960d883824ffcdd246af20af..40b269563e2a2e99461c27d728bf020b4baf29be 100644 --- a/attention_kindselective_n_heads4_seed1340/model_42500.pt +++ b/attention_kindselective_n_heads4_seed1340/model_42500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:7e937e5bde413b5f1d5cd2e7a03428c77cca6ac8ef7547cc1d8aa0565c33ea0d +oid sha256:5aa2c435ba305fa32a8e5282696e70255f511cfaf0c5b519f809a0a2368b0f9c size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_45000.pt b/attention_kindselective_n_heads4_seed1340/model_45000.pt index a0d8357566f22c046b930592717b873acc52c9e1..93f3e76e391d502e5aa1b7ddfdde52f106636541 100644 --- a/attention_kindselective_n_heads4_seed1340/model_45000.pt +++ b/attention_kindselective_n_heads4_seed1340/model_45000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:03ef3fa0f39e905fced4c8e9f6d4c4c36d3dfdec8a8589b65a5f4c6363469015 +oid sha256:1d0d999a95806ce0df25259e77ae207370f9d5fbc520455ebc8c6df91d2a1049 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_47500.pt b/attention_kindselective_n_heads4_seed1340/model_47500.pt index 2e14d9834227886ac51ae22e552f5c47bab89320..62e976af736aa93215dbb7d19403be39b024ee4f 100644 --- a/attention_kindselective_n_heads4_seed1340/model_47500.pt +++ b/attention_kindselective_n_heads4_seed1340/model_47500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:58579aca525c6640a50ed6de3f5bc997ccf568b1c74ac74080a8413181c72725 +oid sha256:e8c612bd311cf5bf36b7cd0eaf2d390b96cbd61422cc6c461817195a65724ea8 size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/model_49999.pt b/attention_kindselective_n_heads4_seed1340/model_49999.pt index c07cd50aaf883d1de929947dc45625af116e944a..6cfc646753fb064801eaaff874d89ad79ad15ffe 100644 --- a/attention_kindselective_n_heads4_seed1340/model_49999.pt +++ b/attention_kindselective_n_heads4_seed1340/model_49999.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ddd2671bc53d3d47c2a3597b04a352fe6e56e1aa82a19bed3ee6f4e205890842 +oid sha256:329cc7c07d336b08babec9f25a79e6cb3f45e98b7201a91059aa2e8f0a57113b size 92843394 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_02500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_02500.pt index 0e7db4740d5d16d05d6f5137a6be8c3d69712087..30ec58836fead13177507af1bf1851aff835e27f 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_02500.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_02500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:71d5d80bc8621c01c7517a9c8966efcf316b98161adb2e069c5201828111ea9b +oid sha256:9e96f244321b4ceb32e937deaaff4052c2e16226a5c21490d87da1ae07a2e463 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_05000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_05000.pt index 7516c3b76e5b9a33f0fc37e554cdc959063b4603..a03f245174dbe441390101b2825941e9f748a5b0 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_05000.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_05000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4cf0986c29b0869c7e9016f8a369414e67e321e3e73344a91dc6a3451b88ff99 +oid sha256:d2eb29f56578c3cff4293ded7486dacf3c591b28bf309ee22630b302d36f8f24 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_07500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_07500.pt index a022464dd21f009682545b2f046b51770946b70b..77b4019d08607d9ab601438165097fd1f94cdf6a 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_07500.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_07500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:f45a3b4647bd40c9e04b365a1e5c1703f4207fa3698dc385fc2cfd5bc0ea4bd5 +oid sha256:f46662c30e0c586b62614381c40c4d40361fe89d16f64503cefe582b93844dd2 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_10000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_10000.pt index 73aa4df26fde3473a9a2c2fb719064d5b69cb4d9..3a6c44fb6b7e746a43aa286f60a62e08eb26ae98 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_10000.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_10000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:06bc9c8845c72cf18f14e45b43f7c4c19b17aa50a69cff163f981ea53f88744f +oid sha256:07246a91a4608bf5ab9f139643924fd088b34009b22bdfeb77ec5a8dc3d6e94b size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_12500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_12500.pt index 98d3cf301de6ea3bca07b14ebbaa3ec6f3b7ed40..985919cde1fbd73bdb98f08e0d70af096926b178 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_12500.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_12500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:61803c8687b0d4b6de778e759324ef063faecd26bcfb6e193e45505fe759b8d9 +oid sha256:cc0fdd1cddbb185439cc502fcc422c882464781999d991dc77c352104ede18c0 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_15000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_15000.pt new file mode 100644 index 0000000000000000000000000000000000000000..cb87223bd90b9731df945770ecb30bd3ba22730f --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_15000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d994c5254e5c42075db07a54cd8b3196eeed538d59eb3f014786b64ec3da70c +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_17500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_17500.pt new file mode 100644 index 0000000000000000000000000000000000000000..241f19e7aa32b830d46f8d67d1a58f4264ebf8fa --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_17500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d2670345760eecd01305e061ae5a4ee5dfeaa24cf41a44adf5919d03ea63db3 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_20000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_20000.pt new file mode 100644 index 0000000000000000000000000000000000000000..cac6e587db17261ceafe7342cb71aef14882f8e6 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_20000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:72b225f844a81f011bef5ad885e1f90ef3eb824869267a3703f1a6434b680fc2 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_22500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_22500.pt new file mode 100644 index 0000000000000000000000000000000000000000..e5ba6e2398a546f842d7f390f9c6c720de236b8f --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_22500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a6f83f09e13afe18b673705f672eb4c982a0acdcd440f50922885b335c734953 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_25000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_25000.pt new file mode 100644 index 0000000000000000000000000000000000000000..04adbc624f844566ff859cbd2d3a69eb5be2a728 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_25000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6eeb58ca059298e41a9d34cf2a5fcfc8f342f052ea97c57e3106bc36533f5bfe +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_27500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_27500.pt new file mode 100644 index 0000000000000000000000000000000000000000..6e825e20c02032001d65d19ca79e9bc27e542032 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_27500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18ce8d19d6111ae0bd7475511f2adab74b2164eb9aacff4ab75e722b6f73a311 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_30000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_30000.pt new file mode 100644 index 0000000000000000000000000000000000000000..a82115549b963f7c84377d7565a12636b6061ada --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_30000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:42c27ce04b2b34ff778efa00ae992b9cfaa0b12cbaf53df09c38f5745821a6a0 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_32500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_32500.pt new file mode 100644 index 0000000000000000000000000000000000000000..7fb750f0b5f065106fa43780efa61ff15f4b2405 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_32500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38c3b9a8f3ab7c52f68a25fcbe1dd87ab4bb869b6d9d5e7d9cf7698080629f9e +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_35000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_35000.pt new file mode 100644 index 0000000000000000000000000000000000000000..63ceda8476b12da0c8b113013733f847edf62162 --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_35000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f3bc2abb7ee0d8214d323a2d1e726e5b56e1bbf13ce1fc078bd5ad0738eb44c9 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_37500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_37500.pt new file mode 100644 index 0000000000000000000000000000000000000000..e69eafd0c0739e330c6a0041e13d50a45522c66c --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_37500.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:115d76cf5edc7fdc26fe7491f3cf57823c30944da97f62912cfd04b9b8cdb858 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_40000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_40000.pt new file mode 100644 index 0000000000000000000000000000000000000000..1bb6c39f06eff56cc1a2cb0fea5af2ac59c9459f --- /dev/null +++ b/attention_kindselective_n_heads4_seed1340/optimizer_40000.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ba7f05d99562a7da60349b27638be67608c000adbd7c48fd923b87f1e0542277 +size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_42500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_42500.pt index 5a138fc85937070865258e9296146f25e71b1d37..def7fe70982b42d735aed488091080aac139a21a 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_42500.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_42500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:dffbcca299af84d6d052050b7c9c7230f9052fbb489bbc7104607f111ccd656e +oid sha256:a626934e4e9e095eecfd4bd018cf53ce8e61b807d9c1512a49275dea6dc52a73 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_45000.pt b/attention_kindselective_n_heads4_seed1340/optimizer_45000.pt index fa386cdd35d28cd7aaf247a5b01485da19a69553..ce4da3ad55b157a2d4e196ed16e8a654776989c3 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_45000.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_45000.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:8b4dba3c28b0a3140da1cda29187583bfb0ffbc5b49bd20693dae0ef1a68772c +oid sha256:8c174347256b7c8c475375a623d2d4d9a2b62df9a0d0c8a8937e9506b32daaf3 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_47500.pt b/attention_kindselective_n_heads4_seed1340/optimizer_47500.pt index 063fd09efd4330da8d20d84733bef58fd940f702..75f4efbd937c2175c65e45fea5a1a0c748f0a917 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_47500.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_47500.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ee4537d10c929400a183bc037015dc153913766e3bc65eeecde61d2e9ef87311 +oid sha256:475db5a1b73d01446ce1d2c901a66c15a4df5f461910601f0e4206c1e76dbd13 size 179406214 diff --git a/attention_kindselective_n_heads4_seed1340/optimizer_49999.pt b/attention_kindselective_n_heads4_seed1340/optimizer_49999.pt index fa6044c86e34a9b4bb1ade369cb5f294654ac39f..24844973a61754fecbce15faf99ffc10e130407d 100644 --- a/attention_kindselective_n_heads4_seed1340/optimizer_49999.pt +++ b/attention_kindselective_n_heads4_seed1340/optimizer_49999.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:f5280d94beeaac57235e9ca63bb2521ca98af393b93180fe9a674caafe9bb168 +oid sha256:031552b880a7fda4f1cc2102a8b65478d94dee2f33b6be0a22bde6deef7ce83b size 179406214