Upload folder using huggingface_hub

attention_kindselective_n_heads4_seed1339/args.json

@@ -1 +1 @@
-{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_6/attention_kindselective_n_heads4_seed1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 4, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 10000, "warmup_steps": 200, "group": "wider_is_better_6", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1339, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "none", "selection_head_linear_combo_scale": 1.0, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": true, "use_mini_model": false, "upload_to_hf": true, "seq_len": 256, "batch_size": 40, "total_batch_size": 10240, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": null, "mask_layernorm": false, "residual_attention_masks": false, "compute_base_shapes": false, "base_shapes_savefile": null, "mup": true, "disable_selection": false, "mup_enable_coord_check_logging": false, "max_lr": 0.00015, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "15e-5_10240_4_1339", "n_embd": 256}
+{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_7/attention_kindselective_n_heads4_seed1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 4, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 50000, "warmup_steps": 200, "group": "wider_is_better_7", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1339, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "none", "selection_head_linear_combo_scale": 1.0, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": true, "use_mini_model": false, "upload_to_hf": true, "seq_len": 256, "batch_size": 40, "total_batch_size": 10240, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": null, "mask_layernorm": false, "residual_attention_masks": false, "compute_base_shapes": false, "base_shapes_savefile": null, "mup": true, "disable_selection": false, "mup_enable_coord_check_logging": false, "max_lr": 0.0001, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "10e-5_10240_4_1339", "n_embd": 256}

attention_kindselective_n_heads4_seed1339/dataloader_10000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3858f6c832feea78a674d8c5c384061cc7d4f22cddbd0a2be6de33bc91e2c72
+size 964

attention_kindselective_n_heads4_seed1339/dataloader_40000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84f58ba3b4a1f9be7da4e697e06782f0e1ce4d3aca49f1997087fc83aa466dd9
+size 964

attention_kindselective_n_heads4_seed1339/dataloader_42500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf6d24c78d89100d146bce9f26be940db3d71092473d9b55db97d6b35531eac2
+size 964

attention_kindselective_n_heads4_seed1339/dataloader_45000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22bb90b43d81f3da5454f91a70e1ed29aeb2f470a727ce38390ff8a5c4924889
+size 964

attention_kindselective_n_heads4_seed1339/dataloader_47500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55507725e6988f190e4963078652fafa6b68e8d4f79221387612612babf3e1c1
+size 964

attention_kindselective_n_heads4_seed1339/dataloader_49999.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47776cddb8021172f048a950b83f25b692cb340214b800ce3837c15ceb58907c
+size 964

attention_kindselective_n_heads4_seed1339/log2.txt

@@ -1,603 +1,699 @@
-0 train 11.297240 (lr=5.0000e-07) (hash(x)=55241167)
-max_steps: 10000
-100 val loss 9.4113
-100 val perplexity 12226.2285
-100 train 9.392327 (lr=5.0500e-05) (hash(x)=46387919)
+max_steps: 50000
+38300 val loss 5.6784
+38300 val perplexity 292.4702
+38300 train 5.395936 (lr=1.0855e-05) (hash(x)=49878964)
 0 val loss 11.2766
 0 val perplexity 78951.9062
-200 val loss 7.8296
-200 val perplexity 2513.9221
-200 train 7.923450 (lr=1.0000e-04) (hash(x)=51276369)
-300 val loss 7.6256
-300 val perplexity 2049.9619
-300 train 7.519519 (lr=9.9977e-05) (hash(x)=48572728)
-0 train 11.297240 (lr=7.5000e-07) (hash(x)=55241167)
-400 val loss 7.5602
-400 val perplexity 1920.2125
-400 train 7.274498 (lr=9.9908e-05) (hash(x)=44943400)
-100 val loss 9.0840
-100 val perplexity 8813.1025
-100 train 9.057368 (lr=7.5750e-05) (hash(x)=46387919)
-500 val loss 7.5149
-500 val perplexity 1835.2683
-500 train 7.412261 (lr=9.9792e-05) (hash(x)=48434590)
-200 val loss 7.7009
-200 val perplexity 2210.4219
-200 train 7.782194 (lr=1.5000e-04) (hash(x)=51276369)
-600 val loss 7.4697
-600 val perplexity 1754.1002
-600 train 7.309482 (lr=9.9631e-05) (hash(x)=47442266)
-300 val loss 7.6257
-300 val perplexity 2050.2629
-300 train 7.521665 (lr=1.4997e-04) (hash(x)=48572728)
-700 val loss 7.4248
-700 val perplexity 1676.9968
-700 train 7.487512 (lr=9.9423e-05) (hash(x)=56495712)
-400 val loss 7.6210
-400 val perplexity 2040.5312
-400 train 7.349378 (lr=1.4986e-04) (hash(x)=44943400)
-800 val loss 7.4002
-800 val perplexity 1636.3661
-800 train 7.382157 (lr=9.9170e-05) (hash(x)=50093738)
-500 val loss 7.5935
-500 val perplexity 1985.2406
-500 train 7.501320 (lr=1.4969e-04) (hash(x)=48434590)
-900 val loss 7.3485
-900 val perplexity 1553.9221
-900 train 7.314708 (lr=9.8872e-05) (hash(x)=49436061)
-600 val loss 7.5432
-600 val perplexity 1887.9182
-600 train 7.397351 (lr=1.4945e-04) (hash(x)=47442266)
-1000 val loss 7.3260
-1000 val perplexity 1519.2688
-1000 train 7.252484 (lr=9.8528e-05) (hash(x)=48818282)
-700 val loss 7.5011
-700 val perplexity 1809.9755
-700 train 7.557298 (lr=1.4913e-04) (hash(x)=56495712)
-1100 val loss 7.2749
-1100 val perplexity 1443.6171
-1100 train 7.333434 (lr=9.8140e-05) (hash(x)=55387425)
-800 val loss 7.4475
-800 val perplexity 1715.6282
-800 train 7.430103 (lr=1.4876e-04) (hash(x)=50093738)
-1200 val loss 7.2674
-1200 val perplexity 1432.7633
-1200 train 7.136762 (lr=9.7708e-05) (hash(x)=48328378)
-900 val loss 7.3857
-900 val perplexity 1612.8206
-900 train 7.335268 (lr=1.4831e-04) (hash(x)=49436061)
-1300 val loss 7.2723
-1300 val perplexity 1439.9219
-1300 train 7.192955 (lr=9.7231e-05) (hash(x)=48956717)
-1000 val loss 7.3329
-1000 val perplexity 1529.8805
-1000 train 7.257126 (lr=1.4779e-04) (hash(x)=48818282)
-1400 val loss 7.2745
-1400 val perplexity 1443.0217
-1400 train 7.159895 (lr=9.6711e-05) (hash(x)=49782571)
-1100 val loss 7.2824
-1100 val perplexity 1454.4734
-1100 train 7.329366 (lr=1.4721e-04) (hash(x)=55387425)
-1500 val loss 7.2582
-1500 val perplexity 1419.7051
-1500 train 7.043027 (lr=9.6149e-05) (hash(x)=51081994)
-1200 val loss 7.2663
-1200 val perplexity 1431.2822
-1200 train 7.143242 (lr=1.4656e-04) (hash(x)=48328378)
-1600 val loss 7.2452
-1600 val perplexity 1401.4271
-1600 train 7.309634 (lr=9.5544e-05) (hash(x)=48791204)
-1300 val loss 7.2275
-1300 val perplexity 1376.7683
-1300 train 7.136340 (lr=1.4585e-04) (hash(x)=48956717)
-1700 val loss 7.2259
-1700 val perplexity 1374.5807
-1700 train 6.931890 (lr=9.4897e-05) (hash(x)=46425011)
-1400 val loss 7.2388
-1400 val perplexity 1392.4169
-1400 train 7.132669 (lr=1.4507e-04) (hash(x)=49782571)
-1800 val loss 7.2145
-1800 val perplexity 1358.9602
-1800 train 6.850976 (lr=9.4209e-05) (hash(x)=44919141)
-1500 val loss 7.2698
-1500 val perplexity 1436.2498
-1500 train 7.063611 (lr=1.4422e-04) (hash(x)=51081994)
-1900 val loss 7.1925
-1900 val perplexity 1329.3683
-1900 train 7.279314 (lr=9.3481e-05) (hash(x)=48299675)
-1600 val loss 7.2511
-1600 val perplexity 1409.6339
-1600 train 7.338916 (lr=1.4332e-04) (hash(x)=48791204)
-2000 val loss 7.1303
-2000 val perplexity 1249.2107
-2000 train 7.887177 (lr=9.2714e-05) (hash(x)=61804797)
-1700 val loss 7.2541
-1700 val perplexity 1413.9268
-1700 train 6.979441 (lr=1.4235e-04) (hash(x)=46425011)
-2100 val loss 7.0845
-2100 val perplexity 1193.3490
-2100 train 7.169927 (lr=9.1908e-05) (hash(x)=53638030)
-1800 val loss 7.2169
-1800 val perplexity 1362.2542
-1800 train 6.886115 (lr=1.4131e-04) (hash(x)=44919141)
-2200 val loss 7.0669
-2200 val perplexity 1172.4698
-2200 train 7.293911 (lr=9.1064e-05) (hash(x)=56364593)
-1900 val loss 7.1736
-1900 val perplexity 1304.4712
-1900 train 7.252931 (lr=1.4022e-04) (hash(x)=48299675)
-2300 val loss 7.0647
-2300 val perplexity 1169.9177
-2300 train 7.168211 (lr=9.0182e-05) (hash(x)=52945580)
-2000 val loss 7.1572
-2000 val perplexity 1283.3210
-2000 train 7.978416 (lr=1.3907e-04) (hash(x)=61804797)
-2400 val loss 7.0197
-2400 val perplexity 1118.4805
-2400 train 7.089632 (lr=8.9265e-05) (hash(x)=54595211)
-2100 val loss 7.0765
-2100 val perplexity 1183.8523
-2500 val loss 7.0181
-2500 val perplexity 1116.6877
-2100 train 7.178714 (lr=1.3786e-04) (hash(x)=53638030)
-2500 train 7.027974 (lr=8.8313e-05) (hash(x)=52323363)
-2200 val loss 7.0702
-2200 val perplexity 1176.4192
-2200 train 7.273838 (lr=1.3660e-04) (hash(x)=56364593)
-2600 val loss 6.9974
-2600 val perplexity 1093.8115
-2600 train 6.790699 (lr=8.7326e-05) (hash(x)=42652707)
-2700 val loss 6.9823
-2700 val perplexity 1077.4113
-2300 val loss 7.0534
-2300 val perplexity 1156.7692
-2700 train 6.681511 (lr=8.6306e-05) (hash(x)=45687855)
-2300 train 7.148480 (lr=1.3527e-04) (hash(x)=52945580)
-2800 val loss 6.9617
-2800 val perplexity 1055.4443
-2800 train 6.951207 (lr=8.5254e-05) (hash(x)=49962830)
-2400 val loss 7.0180
-2400 val perplexity 1116.6062
-2400 train 7.111838 (lr=1.3390e-04) (hash(x)=54595211)
-2900 val loss 6.9430
-2900 val perplexity 1035.8958
-2900 train 6.955583 (lr=8.4170e-05) (hash(x)=52958451)
-2500 val loss 6.9926
-2500 val perplexity 1088.5376
-2500 train 6.990477 (lr=1.3247e-04) (hash(x)=52323363)
-3000 val loss 6.9176
-3000 val perplexity 1009.8487
-3000 train 6.864679 (lr=8.3057e-05) (hash(x)=48896990)
-2600 val loss 7.0060
-2600 val perplexity 1103.2196
-2600 train 6.810057 (lr=1.3099e-04) (hash(x)=42652707)
-3100 val loss 6.9101
-3100 val perplexity 1002.2977
-3100 train 6.856146 (lr=8.1915e-05) (hash(x)=47064930)
-2700 val loss 6.9722
-2700 val perplexity 1066.5864
-2700 train 6.679486 (lr=1.2946e-04) (hash(x)=45687855)
-3200 val loss 6.8815
-3200 val perplexity 974.0536
-3200 train 6.596547 (lr=8.0745e-05) (hash(x)=45145921)
-2800 val loss 6.9433
-2800 val perplexity 1036.2302
-2800 train 6.932369 (lr=1.2788e-04) (hash(x)=49962830)
-3300 val loss 6.8839
-3300 val perplexity 976.3783
-3300 train 6.722473 (lr=7.9549e-05) (hash(x)=48085018)
-2900 val loss 6.9177
-2900 val perplexity 1009.9710
-2900 train 6.917989 (lr=1.2626e-04) (hash(x)=52958451)
-3400 val loss 6.8586
-3400 val perplexity 951.9899
-3400 train 6.852911 (lr=7.8328e-05) (hash(x)=49596053)
-3000 val loss 6.8656
-3000 val perplexity 958.6799
-3000 train 6.826707 (lr=1.2459e-04) (hash(x)=48896990)
-3500 val loss 6.8550
-3500 val perplexity 948.5732
-3500 train 6.454469 (lr=7.7082e-05) (hash(x)=41135634)
-3100 val loss 6.9046
-3100 val perplexity 996.8251
-3100 train 6.876366 (lr=1.2287e-04) (hash(x)=47064930)
-3600 val loss 6.8194
-3600 val perplexity 915.4774
-3600 train 6.813796 (lr=7.5814e-05) (hash(x)=54286330)
-3200 val loss 6.8627
-3200 val perplexity 955.9679
-3200 train 6.592710 (lr=1.2112e-04) (hash(x)=45145921)
-3700 val loss 6.8307
-3700 val perplexity 925.8421
-3700 train 6.721279 (lr=7.4525e-05) (hash(x)=57061993)
-3300 val loss 6.8442
-3300 val perplexity 938.4250
-3300 train 6.669895 (lr=1.1932e-04) (hash(x)=48085018)
-3800 val loss 6.8211
-3800 val perplexity 916.9943
-3800 train 6.377489 (lr=7.3215e-05) (hash(x)=46544884)
-3400 val loss 6.8391
-3400 val perplexity 933.6216
-3400 train 6.844838 (lr=1.1749e-04) (hash(x)=49596053)
-3900 val loss 6.7986
-3900 val perplexity 896.6097
-3900 train 6.638959 (lr=7.1887e-05) (hash(x)=48937435)
-3500 val loss 6.8285
-3500 val perplexity 923.8395
-3500 train 6.439719 (lr=1.1562e-04) (hash(x)=41135634)
-4000 val loss 6.7805
-4000 val perplexity 880.4681
-4000 train 6.811349 (lr=7.0541e-05) (hash(x)=54466186)
-3600 val loss 6.8305
-3600 val perplexity 925.6660
-3600 train 6.827404 (lr=1.1372e-04) (hash(x)=54286330)
-4100 val loss 6.7587
-4100 val perplexity 861.4922
-4100 train 6.585770 (lr=6.9180e-05) (hash(x)=51079995)
-3700 val loss 6.8390
-3700 val perplexity 933.5700
-3700 train 6.702039 (lr=1.1179e-04) (hash(x)=57061993)
-4200 val loss 6.7599
-4200 val perplexity 862.5894
-4200 train 6.885074 (lr=6.7804e-05) (hash(x)=56495581)
-3800 val loss 6.8435
-3800 val perplexity 937.7438
-3800 train 6.419742 (lr=1.0982e-04) (hash(x)=46544884)
-4300 val loss 6.7067
-4300 val perplexity 817.8332
-4300 train 6.550481 (lr=6.6414e-05) (hash(x)=45095478)
-3900 val loss 6.8543
-3900 val perplexity 947.9782
-3900 train 6.697006 (lr=1.0783e-04) (hash(x)=48937435)
-4400 val loss 6.6989
-4400 val perplexity 811.4722
-4400 train 6.710745 (lr=6.5013e-05) (hash(x)=49954028)
-4000 val loss 6.8472
-4000 val perplexity 941.2394
-4000 train 6.863346 (lr=1.0581e-04) (hash(x)=54466186)
-4500 val loss 6.6732
-4500 val perplexity 790.9513
-4500 train 6.604428 (lr=6.3602e-05) (hash(x)=48975821)
-4100 val loss 6.8166
-4100 val perplexity 912.8441
-4100 train 6.650379 (lr=1.0377e-04) (hash(x)=51079995)
-4600 val loss 6.6435
-4600 val perplexity 767.7492
-4600 train 6.736774 (lr=6.2182e-05) (hash(x)=49293028)
-4200 val loss 6.8169
-4200 val perplexity 913.1793
-4200 train 6.960119 (lr=1.0171e-04) (hash(x)=56495581)
-4700 val loss 6.6407
-4700 val perplexity 765.6613
-4700 train 6.549062 (lr=6.0754e-05) (hash(x)=48077753)
-4300 val loss 6.7746
-4300 val perplexity 875.3004
-4300 train 6.632710 (lr=9.9622e-05) (hash(x)=45095478)
-4800 val loss 6.6097
-4800 val perplexity 742.2697
-4800 train 6.485708 (lr=5.9321e-05) (hash(x)=45037879)
-4400 val loss 6.7593
-4400 val perplexity 862.0560
-4400 train 6.739607 (lr=9.7520e-05) (hash(x)=49954028)
-4900 val loss 6.5948
-4900 val perplexity 731.2781
-4900 train 6.503368 (lr=5.7883e-05) (hash(x)=50320154)
-4500 val loss 6.7230
-4500 val perplexity 831.2733
-4500 train 6.650680 (lr=9.5403e-05) (hash(x)=48975821)
-5000 val loss 6.5871
-5000 val perplexity 725.6609
-5000 train 6.830953 (lr=5.6442e-05) (hash(x)=55024523)
-4600 val loss 6.7008
-4600 val perplexity 813.0734
-4600 train 6.782051 (lr=9.3273e-05) (hash(x)=49293028)
-5100 val loss 6.5751
-5100 val perplexity 717.0258
-5100 train 6.487334 (lr=5.5000e-05) (hash(x)=47570607)
-4700 val loss 6.7082
-4700 val perplexity 819.0656
-4700 train 6.629680 (lr=9.1132e-05) (hash(x)=48077753)
-5200 val loss 6.5637
-5200 val perplexity 708.9213
-5200 train 6.681831 (lr=5.3558e-05) (hash(x)=54311349)
-4800 val loss 6.7095
-4800 val perplexity 820.1912
-4800 train 6.573892 (lr=8.8982e-05) (hash(x)=45037879)
-5300 val loss 6.5509
-5300 val perplexity 699.8708
-5300 train 6.444979 (lr=5.2117e-05) (hash(x)=47178524)
-4900 val loss 6.6770
-4900 val perplexity 793.9644
-4900 train 6.600446 (lr=8.6825e-05) (hash(x)=50320154)
-5400 val loss 6.5382
-5400 val perplexity 691.0194
-5400 train 6.540273 (lr=5.0679e-05) (hash(x)=49425088)
-5000 val loss 6.6830
-5000 val perplexity 798.7117
-5000 train 6.917428 (lr=8.4663e-05) (hash(x)=55024523)
-5500 val loss 6.5273
-5500 val perplexity 683.5435
-5500 train 6.369995 (lr=4.9246e-05) (hash(x)=46383189)
-5100 val loss 6.6579
-5100 val perplexity 778.9476
-5100 train 6.595668 (lr=8.2500e-05) (hash(x)=47570607)
-5600 val loss 6.5284
-5600 val perplexity 684.2851
-5600 train 6.712617 (lr=4.7818e-05) (hash(x)=54084990)
-5200 val loss 6.6476
-5200 val perplexity 770.9180
-5200 train 6.782622 (lr=8.0337e-05) (hash(x)=54311349)
-5700 val loss 6.5284
-5700 val perplexity 684.2838
-5700 train 6.222934 (lr=4.6398e-05) (hash(x)=47384182)
-5800 val loss 6.5110
-5800 val perplexity 672.5214
-5800 train 6.396578 (lr=4.4987e-05) (hash(x)=51683744)
-5300 val loss 6.6252
-5300 val perplexity 753.8685
-5300 train 6.532175 (lr=7.8175e-05) (hash(x)=47178524)
-5900 val loss 6.5014
-5900 val perplexity 666.0516
-5900 train 6.510395 (lr=4.3586e-05) (hash(x)=52798836)
-5400 val loss 6.6181
-5400 val perplexity 748.5306
-5400 train 6.616325 (lr=7.6018e-05) (hash(x)=49425088)
-6000 val loss 6.5040
-6000 val perplexity 667.8166
-6000 train 6.393989 (lr=4.2196e-05) (hash(x)=50945000)
-6100 val loss 6.5062
-6100 val perplexity 669.2753
-6100 train 6.196708 (lr=4.0820e-05) (hash(x)=48964427)
-5500 val loss 6.6147
-5500 val perplexity 745.9890
-5500 train 6.448332 (lr=7.3868e-05) (hash(x)=46383189)
-6200 val loss 6.5098
-6200 val perplexity 671.7039
-6200 train 6.268740 (lr=3.9459e-05) (hash(x)=47148610)
-5600 val loss 6.6037
-5600 val perplexity 737.8403
-5600 train 6.795812 (lr=7.1727e-05) (hash(x)=54084990)
-6300 val loss 6.5004
-6300 val perplexity 665.4230
-6300 train 6.412795 (lr=3.8113e-05) (hash(x)=50860553)
-6400 val loss 6.4993
-6400 val perplexity 664.6778
-6400 train 6.270996 (lr=3.6785e-05) (hash(x)=50553992)
-5700 val loss 6.5924
-5700 val perplexity 729.5043
-5700 train 6.275771 (lr=6.9597e-05) (hash(x)=47384182)
-6500 val loss 6.4847
-6500 val perplexity 655.0747
-6500 train 6.289717 (lr=3.5475e-05) (hash(x)=52521332)
-5800 val loss 6.5847
-5800 val perplexity 723.9128
-5800 train 6.476488 (lr=6.7480e-05) (hash(x)=51683744)
-6600 val loss 6.4606
-6600 val perplexity 639.4189
-6600 train 6.806929 (lr=3.4186e-05) (hash(x)=52609843)
-6700 val loss 6.4414
-6700 val perplexity 627.2814
-6700 train 6.480598 (lr=3.2918e-05) (hash(x)=53632957)
-5900 val loss 6.5876
-5900 val perplexity 726.0012
-5900 train 6.579369 (lr=6.5378e-05) (hash(x)=52798836)
-6800 val loss 6.4358
-6800 val perplexity 623.7556
-6800 train 6.441747 (lr=3.1672e-05) (hash(x)=46927608)
-6000 val loss 6.5899
-6000 val perplexity 727.6727
-6000 train 6.478214 (lr=6.3294e-05) (hash(x)=50945000)
-6900 val loss 6.4211
-6900 val perplexity 614.6884
-6900 train 6.448431 (lr=3.0451e-05) (hash(x)=54483057)
-7000 val loss 6.4171
-7000 val perplexity 612.2233
-6100 val loss 6.5737
-6100 val perplexity 716.0176
-7000 train 6.564780 (lr=2.9255e-05) (hash(x)=51493884)
-6100 train 6.276618 (lr=6.1230e-05) (hash(x)=48964427)
-7100 val loss 6.4078
-7100 val perplexity 606.5637
-7100 train 6.538711 (lr=2.8085e-05) (hash(x)=53303341)
-6200 val loss 6.5912
-6200 val perplexity 728.6831
-6200 train 6.360708 (lr=5.9188e-05) (hash(x)=47148610)
-7200 val loss 6.4123
-7200 val perplexity 609.2900
-7200 train 6.105514 (lr=2.6943e-05) (hash(x)=45272178)
-6300 val loss 6.5876
-6300 val perplexity 726.0035
-6300 train 6.502324 (lr=5.7169e-05) (hash(x)=50860553)
-7300 val loss 6.4032
-7300 val perplexity 603.7695
-7300 train 6.337297 (lr=2.5830e-05) (hash(x)=50389872)
-7400 val loss 6.3916
-7400 val perplexity 596.8148
-7400 train 6.031796 (lr=2.4746e-05) (hash(x)=43796301)
-6400 val loss 6.5650
-6400 val perplexity 709.8069
-6400 train 6.330476 (lr=5.5177e-05) (hash(x)=50553992)
-7500 val loss 6.3875
-7500 val perplexity 594.3956
-7500 train 6.322572 (lr=2.3694e-05) (hash(x)=47808686)
-6500 val loss 6.5650
-6500 val perplexity 709.7889
-6500 train 6.371252 (lr=5.3213e-05) (hash(x)=52521332)
-7600 val loss 6.3867
-7600 val perplexity 593.9199
-7600 train 6.118526 (lr=2.2674e-05) (hash(x)=41936898)
-7700 val loss 6.3863
-7700 val perplexity 593.6702
-7700 train 6.602107 (lr=2.1687e-05) (hash(x)=57550318)
-6600 val loss 6.5361
-6600 val perplexity 689.5613
-6600 train 6.896532 (lr=5.1279e-05) (hash(x)=52609843)
-7800 val loss 6.3803
-7800 val perplexity 590.1259
-7800 train 6.245681 (lr=2.0735e-05) (hash(x)=47485210)
-6700 val loss 6.5179
-6700 val perplexity 677.1226
-6700 train 6.555040 (lr=4.9377e-05) (hash(x)=53632957)
-7900 val loss 6.3767
-7900 val perplexity 587.9981
-7900 train 6.719403 (lr=1.9818e-05) (hash(x)=53228688)
-8000 val loss 6.3760
-8000 val perplexity 587.5701
-8000 train 6.432549 (lr=1.8936e-05) (hash(x)=52018673)
-6800 val loss 6.5208
-6800 val perplexity 679.0950
-6800 train 6.524570 (lr=4.7509e-05) (hash(x)=46927608)
-8100 val loss 6.3687
-8100 val perplexity 583.2793
-8100 train 6.201787 (lr=1.8092e-05) (hash(x)=47079349)
-6900 val loss 6.5172
-6900 val perplexity 676.6768
-6900 train 6.538620 (lr=4.5676e-05) (hash(x)=54483057)
-8200 val loss 6.3645
-8200 val perplexity 580.8560
-8200 train 6.434308 (lr=1.7286e-05) (hash(x)=57921563)
-8300 val loss 6.3590
-8300 val perplexity 577.6404
-8300 train 5.898799 (lr=1.6519e-05) (hash(x)=45038933)
-7000 val loss 6.5124
-7000 val perplexity 673.4177
-7000 train 6.681723 (lr=4.3882e-05) (hash(x)=51493884)
-8400 val loss 6.3775
-8400 val perplexity 588.4750
-8400 train 6.225904 (lr=1.5791e-05) (hash(x)=47763246)
-7100 val loss 6.4924
-7100 val perplexity 660.0879
-7100 train 6.627834 (lr=4.2128e-05) (hash(x)=53303341)
-8500 val loss 6.3612
-8500 val perplexity 578.9673
-8500 train 6.429732 (lr=1.5103e-05) (hash(x)=56176595)
-8600 val loss 6.3597
-8600 val perplexity 578.0970
-8600 train 6.434188 (lr=1.4456e-05) (hash(x)=55184249)
-7200 val loss 6.5003
-7200 val perplexity 665.3522
-7200 train 6.209903 (lr=4.0414e-05) (hash(x)=45272178)
-8700 val loss 6.3549
-8700 val perplexity 575.2811
-8700 train 6.294187 (lr=1.3851e-05) (hash(x)=46471646)
-8800 val loss 6.3584
-8800 val perplexity 577.3042
-8800 train 6.057069 (lr=1.3289e-05) (hash(x)=46233162)
-7300 val loss 6.4876
-7300 val perplexity 656.9751
-7300 train 6.401985 (lr=3.8745e-05) (hash(x)=50389872)
-8900 val loss 6.3599
-8900 val perplexity 578.1836
-8900 train 6.283579 (lr=1.2769e-05) (hash(x)=47233684)
-7400 val loss 6.4828
-7400 val perplexity 653.8246
-7400 train 6.129344 (lr=3.7120e-05) (hash(x)=43796301)
-9000 val loss 6.3502
-9000 val perplexity 572.6089
-9000 train 6.235326 (lr=1.2292e-05) (hash(x)=48374529)
-9100 val loss 6.3393
-9100 val perplexity 566.3860
-9100 train 6.325945 (lr=1.1860e-05) (hash(x)=48065371)
-7500 val loss 6.4820
-7500 val perplexity 653.2842
-7500 train 6.422829 (lr=3.5541e-05) (hash(x)=47808686)
-9200 val loss 6.3315
-9200 val perplexity 561.9741
-9200 train 6.326896 (lr=1.1472e-05) (hash(x)=47408078)
-7600 val loss 6.4791
-7600 val perplexity 651.3569
-7600 train 6.215595 (lr=3.4011e-05) (hash(x)=41936898)
-9300 val loss 6.3276
-9300 val perplexity 559.8051
-9300 train 6.309064 (lr=1.1128e-05) (hash(x)=50749781)
-9400 val loss 6.3237
-9400 val perplexity 557.6545
-9400 train 6.525542 (lr=1.0830e-05) (hash(x)=48560169)
-7700 val loss 6.4761
-7700 val perplexity 649.4601
-7700 train 6.682161 (lr=3.2531e-05) (hash(x)=57550318)
-9500 val loss 6.3229
-9500 val perplexity 557.1821
-9500 train 6.435276 (lr=1.0577e-05) (hash(x)=50936392)
-7800 val loss 6.4727
-7800 val perplexity 647.2555
-7800 train 6.342023 (lr=3.1102e-05) (hash(x)=47485210)
-9600 val loss 6.3189
-9600 val perplexity 554.9551
-9600 train 6.341009 (lr=1.0369e-05) (hash(x)=50651714)
-9700 val loss 6.3178
-9700 val perplexity 554.3513
-9700 train 6.391021 (lr=1.0208e-05) (hash(x)=47311384)
-7900 val loss 6.4673
-7900 val perplexity 643.7279
-7900 train 6.802589 (lr=2.9726e-05) (hash(x)=53228688)
-8000 val loss 6.4706
-8000 val perplexity 645.8663
-8000 train 6.519435 (lr=2.8405e-05) (hash(x)=52018673)
-9800 val loss 6.3160
-9800 val perplexity 553.3610
-9800 train 6.418117 (lr=1.0092e-05) (hash(x)=50921139)
-8100 val loss 6.4637
-8100 val perplexity 641.4133
-8100 train 6.298305 (lr=2.7138e-05) (hash(x)=47079349)
-9900 val loss 6.3219
-9900 val perplexity 556.6038
-9900 train 6.467185 (lr=1.0023e-05) (hash(x)=48142455)
-8200 val loss 6.4578
-8200 val perplexity 637.6469
-8200 train 6.539501 (lr=2.5929e-05) (hash(x)=57921563)
-9999 val loss 6.3137
-9999 val perplexity 552.0605
-8300 val loss 6.4564
-8300 val perplexity 636.7405
-8300 train 6.004556 (lr=2.4778e-05) (hash(x)=45038933)
-8400 val loss 6.4678
-8400 val perplexity 644.0859
-8400 train 6.313585 (lr=2.3686e-05) (hash(x)=47763246)
-8500 val loss 6.4619
-8500 val perplexity 640.2839
-8500 train 6.524305 (lr=2.2655e-05) (hash(x)=56176595)
-8600 val loss 6.4634
-8600 val perplexity 641.2393
-8600 train 6.531406 (lr=2.1685e-05) (hash(x)=55184249)
-8700 val loss 6.4605
-8700 val perplexity 639.3777
-8700 train 6.405812 (lr=2.0777e-05) (hash(x)=46471646)
-8800 val loss 6.4620
-8800 val perplexity 640.3541
-8800 train 6.169169 (lr=1.9933e-05) (hash(x)=46233162)
-8900 val loss 6.4575
-8900 val perplexity 637.4733
-8900 train 6.387800 (lr=1.9153e-05) (hash(x)=47233684)
-9000 val loss 6.4577
-9000 val perplexity 637.6104
-9000 train 6.348251 (lr=1.8439e-05) (hash(x)=48374529)
-9100 val loss 6.4384
-9100 val perplexity 625.4025
-9100 train 6.387834 (lr=1.7790e-05) (hash(x)=48065371)
-9200 val loss 6.4323
-9200 val perplexity 621.5872
-9200 train 6.400970 (lr=1.7208e-05) (hash(x)=47408078)
-9300 val loss 6.4267
-9300 val perplexity 618.1290
-9300 train 6.410742 (lr=1.6692e-05) (hash(x)=50749781)
-9400 val loss 6.4225
-9400 val perplexity 615.5146
-9400 train 6.621217 (lr=1.6245e-05) (hash(x)=48560169)
-9500 val loss 6.4212
-9500 val perplexity 614.7256
-9500 train 6.548310 (lr=1.5865e-05) (hash(x)=50936392)
-9600 val loss 6.4178
-9600 val perplexity 612.6835
-9600 train 6.430169 (lr=1.5554e-05) (hash(x)=50651714)
-9700 val loss 6.4189
-9700 val perplexity 613.3140
-9700 train 6.469123 (lr=1.5312e-05) (hash(x)=47311384)
-9800 val loss 6.4171
-9800 val perplexity 612.1956
-9800 train 6.518755 (lr=1.5139e-05) (hash(x)=50921139)
-9900 val loss 6.4295
-9900 val perplexity 619.8542
-9900 train 6.563539 (lr=1.5035e-05) (hash(x)=48142455)
-9999 val loss 6.4144
-9999 val perplexity 610.5499
+38400 val loss 5.6700
+38400 val perplexity 290.0423
+38400 train 5.604312 (lr=1.0760e-05) (hash(x)=52855313)
+38500 val loss 5.6699
+38500 val perplexity 289.9914
+38500 train 5.250736 (lr=1.0666e-05) (hash(x)=47213252)
+0 train 11.297240 (lr=5.0000e-07) (hash(x)=55241167)
+38600 val loss 5.6694
+38600 val perplexity 289.8528
+38600 train 5.830087 (lr=1.0572e-05) (hash(x)=50001498)
+100 val loss 9.4564
+100 val perplexity 12789.4766
+100 train 9.441436 (lr=5.0500e-05) (hash(x)=46387919)
+38700 val loss 5.6726
+38700 val perplexity 290.7901
+38700 train 5.401573 (lr=1.0479e-05) (hash(x)=49541736)
+200 val loss 7.8121
+200 val perplexity 2470.3337
+200 train 7.900826 (lr=1.0000e-04) (hash(x)=51276369)
+38800 val loss 5.6685
+38800 val perplexity 289.6135
+38800 train 5.601834 (lr=1.0386e-05) (hash(x)=49712820)
+300 val loss 7.6376
+300 val perplexity 2074.7930
+300 train 7.534567 (lr=9.9999e-05) (hash(x)=48572728)
+38900 val loss 5.6613
+38900 val perplexity 287.5113
+38900 train 5.612003 (lr=1.0294e-05) (hash(x)=50913051)
+400 val loss 7.5966
+400 val perplexity 1991.4926
+400 train 7.331833 (lr=9.9996e-05) (hash(x)=44943400)
+39000 val loss 5.6624
+39000 val perplexity 287.8408
+39000 train 5.621553 (lr=1.0203e-05) (hash(x)=46110703)
+500 val loss 7.5399
+500 val perplexity 1881.5571
+500 train 7.449239 (lr=9.9992e-05) (hash(x)=48434590)
+39100 val loss 5.6597
+39100 val perplexity 287.0684
+39100 train 5.493328 (lr=1.0113e-05) (hash(x)=43887045)
+600 val loss 7.4966
+600 val perplexity 1801.9823
+600 train 7.344361 (lr=9.9986e-05) (hash(x)=47442266)
+39200 val loss 5.6595
+39200 val perplexity 287.0032
+39200 train 5.897973 (lr=1.0023e-05) (hash(x)=53102477)
+700 val loss 7.4837
+700 val perplexity 1778.8489
+700 train 7.543903 (lr=9.9978e-05) (hash(x)=56495712)
+39300 val loss 5.6569
+39300 val perplexity 286.2463
+39300 train 5.512180 (lr=9.9341e-06) (hash(x)=44607212)
+800 val loss 7.4533
+800 val perplexity 1725.5752
+800 train 7.440802 (lr=9.9968e-05) (hash(x)=50093738)
+39400 val loss 5.6587
+39400 val perplexity 286.7726
+39400 train 5.844805 (lr=9.8458e-06) (hash(x)=55054109)
+900 val loss 7.4095
+900 val perplexity 1651.6809
+900 train 7.374321 (lr=9.9956e-05) (hash(x)=49436061)
+39500 val loss 5.6560
+39500 val perplexity 286.0097
+39500 train 5.519652 (lr=9.7581e-06) (hash(x)=47529853)
+1000 val loss 7.3944
+1000 val perplexity 1626.8689
+1000 train 7.331343 (lr=9.9943e-05) (hash(x)=48818282)
+39600 val loss 5.6537
+39600 val perplexity 285.3348
+39600 train 5.425322 (lr=9.6712e-06) (hash(x)=50783698)
+1100 val loss 7.3509
+1100 val perplexity 1557.6047
+1100 train 7.407344 (lr=9.9927e-05) (hash(x)=55387425)
+39700 val loss 5.6570
+39700 val perplexity 286.2832
+39700 train 5.444355 (lr=9.5849e-06) (hash(x)=47104160)
+1200 val loss 7.5253
+1200 val perplexity 1854.3712
+1200 train 7.421935 (lr=9.9910e-05) (hash(x)=48328378)
+39800 val loss 5.6608
+39800 val perplexity 287.3673
+39800 train 5.592721 (lr=9.4994e-06) (hash(x)=50941478)
+1300 val loss 7.3479
+1300 val perplexity 1553.0095
+1300 train 7.264929 (lr=9.9892e-05) (hash(x)=48956717)
+39900 val loss 5.6565
+39900 val perplexity 286.1535
+39900 train 5.388857 (lr=9.4146e-06) (hash(x)=47374830)
+1400 val loss 7.3611
+1400 val perplexity 1573.6216
+1400 train 7.274777 (lr=9.9871e-05) (hash(x)=49782571)
+40000 val loss 5.6581
+40000 val perplexity 286.6104
+40000 train 5.779648 (lr=9.3305e-06) (hash(x)=56749414)
+1500 val loss 7.3527
+1500 val perplexity 1560.3909
+1500 train 7.145610 (lr=9.9849e-05) (hash(x)=51081994)
+40100 val loss 5.6620
+40100 val perplexity 287.7258
+40100 train 5.607458 (lr=9.2472e-06) (hash(x)=49256359)
+1600 val loss 7.3263
+1600 val perplexity 1519.7659
+1600 train 7.390820 (lr=9.9825e-05) (hash(x)=48791204)
+40200 val loss 5.6555
+40200 val perplexity 285.8730
+40200 train 5.576644 (lr=9.1646e-06) (hash(x)=47932192)
+1700 val loss 7.2937
+1700 val perplexity 1471.0563
+1700 train 7.009460 (lr=9.9799e-05) (hash(x)=46425011)
+40300 val loss 5.6568
+40300 val perplexity 286.2384
+40300 train 5.476855 (lr=9.0827e-06) (hash(x)=42473499)
+1800 val loss 7.3241
+1800 val perplexity 1516.4382
+1800 train 6.985591 (lr=9.9771e-05) (hash(x)=44919141)
+40400 val loss 5.6550
+40400 val perplexity 285.7186
+40400 train 5.456766 (lr=9.0015e-06) (hash(x)=50469946)
+1900 val loss 7.2486
+1900 val perplexity 1406.1027
+1900 train 7.351365 (lr=9.9741e-05) (hash(x)=48299675)
+40500 val loss 5.6501
+40500 val perplexity 284.3304
+40500 train 5.743055 (lr=8.9211e-06) (hash(x)=48818656)
+2000 val loss 7.2173
+2000 val perplexity 1362.7544
+2000 train 8.040308 (lr=9.9710e-05) (hash(x)=61804797)
+40600 val loss 5.6457
+40600 val perplexity 283.0702
+40600 train 5.488149 (lr=8.8414e-06) (hash(x)=45215748)
+2100 val loss 7.2014
+2100 val perplexity 1341.2904
+2100 train 7.292386 (lr=9.9677e-05) (hash(x)=53638030)
+40700 val loss 5.6458
+40700 val perplexity 283.0904
+40700 train 5.773078 (lr=8.7624e-06) (hash(x)=48057577)
+2200 val loss 7.1685
+2200 val perplexity 1297.8553
+2200 train 7.388167 (lr=9.9642e-05) (hash(x)=56364593)
+40800 val loss 5.6444
+40800 val perplexity 282.7088
+40800 train 5.605831 (lr=8.6842e-06) (hash(x)=50103136)
+2300 val loss 7.1717
+2300 val perplexity 1302.0543
+2300 train 7.251419 (lr=9.9606e-05) (hash(x)=52945580)
+40900 val loss 5.6433
+40900 val perplexity 282.3985
+40900 train 5.499018 (lr=8.6068e-06) (hash(x)=45764202)
+2400 val loss 7.1734
+2400 val perplexity 1304.2727
+2400 train 7.251097 (lr=9.9567e-05) (hash(x)=54595211)
+41000 val loss 5.6416
+41000 val perplexity 281.9000
+41000 train 5.758846 (lr=8.5301e-06) (hash(x)=50724130)
+2500 val loss 7.1699
+2500 val perplexity 1299.7263
+2500 train 7.189177 (lr=9.9527e-05) (hash(x)=52323363)
+41100 val loss 5.6407
+41100 val perplexity 281.6723
+41100 train 5.572755 (lr=8.4541e-06) (hash(x)=48739125)
+2600 val loss 7.1373
+2600 val perplexity 1257.9912
+2600 train 6.959945 (lr=9.9485e-05) (hash(x)=42652707)
+41200 val loss 5.6379
+41200 val perplexity 280.8844
+41200 train 5.594127 (lr=8.3789e-06) (hash(x)=47473129)
+2700 val loss 7.1502
+2700 val perplexity 1274.4105
+2700 train 6.867459 (lr=9.9442e-05) (hash(x)=45687855)
+41300 val loss 5.6379
+41300 val perplexity 280.8649
+41300 train 5.563573 (lr=8.3045e-06) (hash(x)=51032823)
+2800 val loss 7.1220
+2800 val perplexity 1238.9615
+2800 train 7.126947 (lr=9.9396e-05) (hash(x)=49962830)
+41400 val loss 5.6520
+41400 val perplexity 284.8671
+41400 train 5.698008 (lr=8.2308e-06) (hash(x)=44411781)
+2900 val loss 7.1669
+2900 val perplexity 1295.7585
+2900 train 7.199690 (lr=9.9349e-05) (hash(x)=52958451)
+41500 val loss 5.6380
+41500 val perplexity 280.8879
+41500 train 5.683953 (lr=8.1579e-06) (hash(x)=48630357)
+3000 val loss 7.1216
+3000 val perplexity 1238.4033
+3000 train 7.069978 (lr=9.9300e-05) (hash(x)=48896990)
+41600 val loss 5.6353
+41600 val perplexity 280.1527
+41600 train 5.446326 (lr=8.0858e-06) (hash(x)=47836869)
+3100 val loss 7.1046
+3100 val perplexity 1217.5770
+3100 train 7.054947 (lr=9.9249e-05) (hash(x)=47064930)
+41700 val loss 5.6361
+41700 val perplexity 280.3648
+41700 train 5.515561 (lr=8.0144e-06) (hash(x)=49143658)
+3200 val loss 7.0866
+3200 val perplexity 1195.8274
+3200 train 6.860336 (lr=9.9197e-05) (hash(x)=45145921)
+41800 val loss 5.6389
+41800 val perplexity 281.1574
+41800 train 5.532392 (lr=7.9438e-06) (hash(x)=50426446)
+3300 val loss 7.0858
+3300 val perplexity 1194.9000
+3300 train 6.935767 (lr=9.9142e-05) (hash(x)=48085018)
+41900 val loss 5.6362
+41900 val perplexity 280.3896
+41900 train 5.521365 (lr=7.8740e-06) (hash(x)=48134303)
+3400 val loss 7.0704
+3400 val perplexity 1176.6060
+3400 train 7.080620 (lr=9.9086e-05) (hash(x)=49596053)
+42000 val loss 5.6366
+42000 val perplexity 280.5153
+42000 train 5.442455 (lr=7.8050e-06) (hash(x)=48151525)
+3500 val loss 7.0368
+3500 val perplexity 1137.7137
+3500 train 6.621675 (lr=9.9028e-05) (hash(x)=41135634)
+42100 val loss 5.6404
+42100 val perplexity 281.5832
+42100 train 5.503111 (lr=7.7368e-06) (hash(x)=49984437)
+3600 val loss 7.0387
+3600 val perplexity 1139.8794
+3600 train 7.045468 (lr=9.8969e-05) (hash(x)=54286330)
+42200 val loss 5.6393
+42200 val perplexity 281.2754
+42200 train 5.540201 (lr=7.6693e-06) (hash(x)=49103965)
+3700 val loss 7.0348
+3700 val perplexity 1135.4819
+3700 train 6.938444 (lr=9.8908e-05) (hash(x)=57061993)
+42300 val loss 5.6361
+42300 val perplexity 280.3770
+42300 train 5.343730 (lr=7.6027e-06) (hash(x)=52134795)
+3800 val loss 7.0378
+3800 val perplexity 1138.8237
+3800 train 6.616735 (lr=9.8845e-05) (hash(x)=46544884)
+42400 val loss 5.6382
+42400 val perplexity 280.9585
+42400 train 5.463866 (lr=7.5368e-06) (hash(x)=47289074)
+3900 val loss 7.0345
+3900 val perplexity 1135.1008
+3900 train 6.881027 (lr=9.8780e-05) (hash(x)=48937435)
+42500 val loss 5.6371
+42500 val perplexity 280.6571
+42500 train 6.041627 (lr=7.4717e-06) (hash(x)=54213272)
+4000 val loss 7.0398
+4000 val perplexity 1141.1144
+4000 train 6.977884 (lr=9.8713e-05) (hash(x)=54466186)
+42600 val loss 5.6364
+42600 val perplexity 280.4386
+42600 train 5.344518 (lr=7.4074e-06) (hash(x)=49347326)
+4100 val loss 7.0053
+4100 val perplexity 1102.4150
+4100 train 6.836554 (lr=9.8645e-05) (hash(x)=51079995)
+42700 val loss 5.6367
+42700 val perplexity 280.5423
+42700 train 5.406415 (lr=7.3440e-06) (hash(x)=45762383)
+4200 val loss 6.9973
+4200 val perplexity 1093.6879
+4200 train 7.157095 (lr=9.8575e-05) (hash(x)=56495581)
+42800 val loss 5.6380
+42800 val perplexity 280.8899
+42800 train 5.719070 (lr=7.2813e-06) (hash(x)=50377506)
+4300 val loss 6.9613
+4300 val perplexity 1055.0035
+4300 train 6.850510 (lr=9.8503e-05) (hash(x)=45095478)
+42900 val loss 5.6290
+42900 val perplexity 278.3965
+42900 train 5.998972 (lr=7.2194e-06) (hash(x)=57332431)
+4400 val loss 6.9297
+4400 val perplexity 1022.1531
+4400 train 6.888327 (lr=9.8430e-05) (hash(x)=49954028)
+43000 val loss 5.6294
+43000 val perplexity 278.5089
+43000 train 5.674274 (lr=7.1583e-06) (hash(x)=50599951)
+4500 val loss 6.9208
+4500 val perplexity 1013.0889
+4500 train 6.849075 (lr=9.8355e-05) (hash(x)=48975821)
+43100 val loss 5.6275
+43100 val perplexity 277.9758
+43100 train 5.593292 (lr=7.0981e-06) (hash(x)=51139362)
+4600 val loss 6.8843
+4600 val perplexity 976.8253
+4600 train 6.973327 (lr=9.8278e-05) (hash(x)=49293028)
+43200 val loss 5.6260
+43200 val perplexity 277.5468
+43200 train 5.478598 (lr=7.0386e-06) (hash(x)=45759314)
+4700 val loss 6.8725
+4700 val perplexity 965.3833
+4700 train 6.789083 (lr=9.8199e-05) (hash(x)=48077753)
+43300 val loss 5.6245
+43300 val perplexity 277.1455
+43300 train 5.677057 (lr=6.9800e-06) (hash(x)=48251106)
+4800 val loss 6.8344
+4800 val perplexity 929.3071
+4800 train 6.655324 (lr=9.8119e-05) (hash(x)=45037879)
+43400 val loss 5.6273
+43400 val perplexity 277.9023
+43400 train 5.983092 (lr=6.9222e-06) (hash(x)=49941825)
+4900 val loss 6.7952
+4900 val perplexity 893.5521
+4900 train 6.717526 (lr=9.8036e-05) (hash(x)=50320154)
+5000 val loss 6.8113
+5000 val perplexity 908.0694
+43500 val loss 5.6248
+43500 val perplexity 277.2100
+43500 train 6.025346 (lr=6.8652e-06) (hash(x)=47282597)
+5000 train 7.030609 (lr=9.7953e-05) (hash(x)=55024523)
+5100 val loss 6.7562
+5100 val perplexity 859.3580
+5100 train 6.689296 (lr=9.7867e-05) (hash(x)=47570607)
+43600 val loss 5.6236
+43600 val perplexity 276.8875
+43600 train 5.733875 (lr=6.8090e-06) (hash(x)=47311813)
+5200 val loss 6.7385
+5200 val perplexity 844.3349
+5200 train 6.866223 (lr=9.7780e-05) (hash(x)=54311349)
+43700 val loss 5.6253
+43700 val perplexity 277.3623
+43700 train 5.692357 (lr=6.7537e-06) (hash(x)=51739445)
+5300 val loss 6.7112
+5300 val perplexity 821.5640
+5300 train 6.617697 (lr=9.7691e-05) (hash(x)=47178524)
+43800 val loss 5.6239
+43800 val perplexity 276.9691
+43800 train 5.663078 (lr=6.6992e-06) (hash(x)=58597156)
+5400 val loss 6.6875
+5400 val perplexity 802.3312
+5400 train 6.683461 (lr=9.7600e-05) (hash(x)=49425088)
+43900 val loss 5.6227
+43900 val perplexity 276.6289
+43900 train 5.352713 (lr=6.6455e-06) (hash(x)=45351304)
+5500 val loss 6.6716
+5500 val perplexity 789.6293
+5500 train 6.500800 (lr=9.7508e-05) (hash(x)=46383189)
+44000 val loss 5.6236
+44000 val perplexity 276.8810
+44000 train 5.404914 (lr=6.5926e-06) (hash(x)=46603503)
+5600 val loss 6.6700
+5600 val perplexity 788.3792
+5600 train 6.856142 (lr=9.7414e-05) (hash(x)=54084990)
+44100 val loss 5.6214
+44100 val perplexity 276.2868
+44100 train 5.590413 (lr=6.5406e-06) (hash(x)=52189096)
+5700 val loss 6.6305
+5700 val perplexity 757.8840
+5700 train 6.303418 (lr=9.7318e-05) (hash(x)=47384182)
+44200 val loss 5.6225
+44200 val perplexity 276.5793
+44200 train 5.789032 (lr=6.4894e-06) (hash(x)=51527792)
+5800 val loss 6.6221
+5800 val perplexity 751.5416
+5800 train 6.522151 (lr=9.7221e-05) (hash(x)=51683744)
+44300 val loss 5.6208
+44300 val perplexity 276.1172
+44300 train 5.475013 (lr=6.4390e-06) (hash(x)=47982628)
+5900 val loss 6.6169
+5900 val perplexity 747.5950
+5900 train 6.624623 (lr=9.7122e-05) (hash(x)=52798836)
+44400 val loss 5.6230
+44400 val perplexity 276.7267
+44400 train 5.611701 (lr=6.3895e-06) (hash(x)=54902968)
+6000 val loss 6.6217
+6000 val perplexity 751.2009
+6000 train 6.511095 (lr=9.7021e-05) (hash(x)=50945000)
+44500 val loss 5.6205
+44500 val perplexity 276.0341
+44500 train 5.537436 (lr=6.3408e-06) (hash(x)=49008493)
+6100 val loss 6.6186
+6100 val perplexity 748.9037
+6100 train 6.304089 (lr=9.6919e-05) (hash(x)=48964427)
+44600 val loss 5.6222
+44600 val perplexity 276.4967
+44600 train 5.353870 (lr=6.2929e-06) (hash(x)=43643824)
+6200 val loss 6.6284
+6200 val perplexity 756.2642
+6200 train 6.390724 (lr=9.6815e-05) (hash(x)=47148610)
+44700 val loss 5.6218
+44700 val perplexity 276.3996
+44700 train 5.470081 (lr=6.2459e-06) (hash(x)=50465872)
+6300 val loss 6.5986
+6300 val perplexity 734.0524
+6300 train 6.511916 (lr=9.6709e-05) (hash(x)=50860553)
+44800 val loss 5.6218
+44800 val perplexity 276.3798
+44800 train 5.378158 (lr=6.1998e-06) (hash(x)=51962160)
+6400 val loss 6.5951
+6400 val perplexity 731.4692
+6400 train 6.364479 (lr=9.6602e-05) (hash(x)=50553992)
+44900 val loss 5.6235
+44900 val perplexity 276.8581
+44900 train 5.763238 (lr=6.1545e-06) (hash(x)=57370039)
+6500 val loss 6.5914
+6500 val perplexity 728.8076
+6500 train 6.389678 (lr=9.6493e-05) (hash(x)=52521332)
+45000 val loss 5.6243
+45000 val perplexity 277.0884
+45000 train 5.469142 (lr=6.1100e-06) (hash(x)=52648130)
+6600 val loss 6.5748
+6600 val perplexity 716.8139
+6600 train 6.929626 (lr=9.6382e-05) (hash(x)=52609843)
+45100 val loss 5.6554
+45100 val perplexity 285.8250
+45100 train 5.327322 (lr=6.0664e-06) (hash(x)=45155459)
+6700 val loss 6.5304
+6700 val perplexity 685.6540
+6700 train 6.558313 (lr=9.6270e-05) (hash(x)=53632957)
+45200 val loss 5.6170
+45200 val perplexity 275.0746
+45200 train 5.611628 (lr=6.0237e-06) (hash(x)=52304992)
+6800 val loss 6.5309
+6800 val perplexity 685.9938
+6800 train 6.539653 (lr=9.6156e-05) (hash(x)=46927608)
+45300 val loss 5.6149
+45300 val perplexity 274.4989
+45300 train 5.468201 (lr=5.9818e-06) (hash(x)=46456407)
+6900 val loss 6.5383
+6900 val perplexity 691.1232
+6900 train 6.552355 (lr=9.6040e-05) (hash(x)=54483057)
+45400 val loss 5.6139
+45400 val perplexity 274.2122
+45400 train 5.607281 (lr=5.9407e-06) (hash(x)=48325493)
+7000 val loss 6.5042
+7000 val perplexity 667.9302
+7000 train 6.676917 (lr=9.5923e-05) (hash(x)=51493884)
+45500 val loss 5.6116
+45500 val perplexity 273.5927
+45500 train 5.583129 (lr=5.9005e-06) (hash(x)=48869064)
+7100 val loss 6.4774
+7100 val perplexity 650.3052
+7100 train 6.625372 (lr=9.5804e-05) (hash(x)=53303341)
+7200 val loss 6.4821
+7200 val perplexity 653.3736
+7200 train 6.203289 (lr=9.5683e-05) (hash(x)=45272178)
+45600 val loss 5.6123
+45600 val perplexity 273.7640
+45600 train 5.690610 (lr=5.8612e-06) (hash(x)=46077805)
+7300 val loss 6.4693
+7300 val perplexity 645.0018
+7300 train 6.416029 (lr=9.5561e-05) (hash(x)=50389872)
+45700 val loss 5.6112
+45700 val perplexity 273.4644
+45700 train 5.849724 (lr=5.8227e-06) (hash(x)=56831165)
+7400 val loss 6.4705
+7400 val perplexity 645.8201
+7400 train 6.127152 (lr=9.5437e-05) (hash(x)=43796301)
+45800 val loss 5.6116
+45800 val perplexity 273.5795
+45800 train 5.459206 (lr=5.7851e-06) (hash(x)=50124921)
+7500 val loss 6.4612
+7500 val perplexity 639.8350
+7500 train 6.404037 (lr=9.5312e-05) (hash(x)=47808686)
+45900 val loss 5.6117
+45900 val perplexity 273.6022
+45900 train 5.222868 (lr=5.7484e-06) (hash(x)=43227400)
+7600 val loss 6.4687
+7600 val perplexity 644.6377
+7600 train 6.201755 (lr=9.5185e-05) (hash(x)=41936898)
+46000 val loss 5.6104
+46000 val perplexity 273.2413
+46000 train 5.602647 (lr=5.7125e-06) (hash(x)=58308887)
+7700 val loss 6.5138
+7700 val perplexity 674.3773
+7700 train 6.732550 (lr=9.5057e-05) (hash(x)=57550318)
+46100 val loss 5.6094
+46100 val perplexity 272.9798
+46100 train 5.677024 (lr=5.6775e-06) (hash(x)=48571632)
+7800 val loss 6.4561
+7800 val perplexity 636.5423
+7800 train 6.328799 (lr=9.4926e-05) (hash(x)=47485210)
+46200 val loss 5.6103
+46200 val perplexity 273.2153
+46200 train 5.874604 (lr=5.6434e-06) (hash(x)=46368862)
+7900 val loss 6.4544
+7900 val perplexity 635.5160
+7900 train 6.812044 (lr=9.4795e-05) (hash(x)=53228688)
+46300 val loss 5.6123
+46300 val perplexity 273.7760
+46300 train 5.376794 (lr=5.6101e-06) (hash(x)=52820317)
+8000 val loss 6.4555
+8000 val perplexity 636.1951
+8000 train 6.513896 (lr=9.4661e-05) (hash(x)=52018673)
+46400 val loss 5.6107
+46400 val perplexity 273.3253
+46400 train 5.410422 (lr=5.5777e-06) (hash(x)=51871487)
+8100 val loss 6.4554
+8100 val perplexity 636.1124
+8100 train 6.303450 (lr=9.4526e-05) (hash(x)=47079349)
+46500 val loss 5.6121
+46500 val perplexity 273.7071
+46500 train 5.565622 (lr=5.5462e-06) (hash(x)=47867973)
+8200 val loss 6.4537
+8200 val perplexity 635.0416
+8200 train 6.513351 (lr=9.4390e-05) (hash(x)=57921563)
+46600 val loss 5.6095
+46600 val perplexity 272.9956
+46600 train 5.271016 (lr=5.5156e-06) (hash(x)=50685713)
+8300 val loss 6.4452
+8300 val perplexity 629.6801
+8300 train 5.972620 (lr=9.4252e-05) (hash(x)=45038933)
+46700 val loss 5.6092
+46700 val perplexity 272.9158
+46700 train 5.732937 (lr=5.4858e-06) (hash(x)=50975526)
+8400 val loss 6.4808
+8400 val perplexity 652.5134
+8400 train 6.312977 (lr=9.4112e-05) (hash(x)=47763246)
+46800 val loss 5.6107
+46800 val perplexity 273.3414
+46800 train 5.750267 (lr=5.4569e-06) (hash(x)=51012129)
+8500 val loss 6.4612
+8500 val perplexity 639.8493
+8500 train 6.493088 (lr=9.3971e-05) (hash(x)=56176595)
+46900 val loss 5.6143
+46900 val perplexity 274.3083
+46900 train 5.333468 (lr=5.4289e-06) (hash(x)=51205353)
+8600 val loss 6.4510
+8600 val perplexity 633.3370
+8600 train 6.529971 (lr=9.3828e-05) (hash(x)=55184249)
+47000 val loss 5.6131
+47000 val perplexity 274.0013
+47000 train 5.759924 (lr=5.4017e-06) (hash(x)=48939661)
+8700 val loss 6.4552
+8700 val perplexity 636.0110
+8700 train 6.379639 (lr=9.3684e-05) (hash(x)=46471646)
+47100 val loss 5.6096
+47100 val perplexity 273.0318
+47100 train 5.531120 (lr=5.3755e-06) (hash(x)=51002557)
+8800 val loss 6.4945
+8800 val perplexity 661.4907
+8800 train 6.198891 (lr=9.3538e-05) (hash(x)=46233162)
+47200 val loss 5.6103
+47200 val perplexity 273.2244
+47200 train 5.444582 (lr=5.3501e-06) (hash(x)=43422209)
+8900 val loss 6.5133
+8900 val perplexity 674.0529
+8900 train 6.432613 (lr=9.3391e-05) (hash(x)=47233684)
+47300 val loss 5.6093
+47300 val perplexity 272.9432
+47300 train 5.365450 (lr=5.3256e-06) (hash(x)=45870552)
+9000 val loss 6.4790
+9000 val perplexity 651.3448
+9000 train 6.380252 (lr=9.3242e-05) (hash(x)=48374529)
+47400 val loss 5.6096
+47400 val perplexity 273.0215
+47400 train 5.453257 (lr=5.3020e-06) (hash(x)=47001810)
+9100 val loss 6.4612
+9100 val perplexity 639.8325
+9100 train 6.426319 (lr=9.3092e-05) (hash(x)=48065371)
+47500 val loss 5.6062
+47500 val perplexity 272.1136
+47500 train 5.676814 (lr=5.2792e-06) (hash(x)=49146619)
+47600 val loss 5.6047
+47600 val perplexity 271.6933
+47600 train 5.688634 (lr=5.2574e-06) (hash(x)=53742853)
+9200 val loss 6.4337
+9200 val perplexity 622.4777
+9200 train 6.365408 (lr=9.2940e-05) (hash(x)=47408078)
+47700 val loss 5.6023
+47700 val perplexity 271.0357
+47700 train 5.858260 (lr=5.2364e-06) (hash(x)=53019458)
+9300 val loss 6.4346
+9300 val perplexity 623.0532
+9300 train 6.420694 (lr=9.2786e-05) (hash(x)=50749781)
+47800 val loss 5.6015
+47800 val perplexity 270.8320
+47800 train 5.688690 (lr=5.2163e-06) (hash(x)=54497914)
+47900 val loss 5.6040
+47900 val perplexity 271.5063
+47900 train 5.530357 (lr=5.1972e-06) (hash(x)=47109012)
+9400 val loss 6.4372
+9400 val perplexity 624.6458
+9400 train 6.634754 (lr=9.2632e-05) (hash(x)=48560169)
+48000 val loss 5.6005
+48000 val perplexity 270.5672
+48000 train 5.590864 (lr=5.1788e-06) (hash(x)=48429816)
+9500 val loss 6.4296
+9500 val perplexity 619.9107
+9500 train 6.521097 (lr=9.2475e-05) (hash(x)=50936392)
+48100 val loss 5.6008
+48100 val perplexity 270.6499
+48100 train 5.544402 (lr=5.1614e-06) (hash(x)=46500465)
+48200 val loss 5.5982
+48200 val perplexity 269.9372
+48200 train 5.762817 (lr=5.1449e-06) (hash(x)=50888195)
+9600 val loss 6.4141
+9600 val perplexity 610.3770
+9600 train 6.430016 (lr=9.2317e-05) (hash(x)=50651714)
+48300 val loss 5.5988
+48300 val perplexity 270.1088
+48300 train 5.666016 (lr=5.1293e-06) (hash(x)=56406586)
+9700 val loss 6.4110
+9700 val perplexity 608.4774
+9700 train 6.487106 (lr=9.2158e-05) (hash(x)=47311384)
+48400 val loss 5.5989
+48400 val perplexity 270.1341
+48400 train 5.751607 (lr=5.1145e-06) (hash(x)=48713969)
+48500 val loss 5.5980
+48500 val perplexity 269.8947
+48500 train 5.026878 (lr=5.1007e-06) (hash(x)=37691696)
+9800 val loss 6.4147
+9800 val perplexity 610.7438
+9800 train 6.526601 (lr=9.1997e-05) (hash(x)=50921139)
+48600 val loss 5.5988
+48600 val perplexity 270.1123
+48600 train 5.892519 (lr=5.0877e-06) (hash(x)=48712321)
+48700 val loss 5.5987
+48700 val perplexity 270.0880
+9900 val loss 6.4657
+9900 val perplexity 642.6958
+48700 train 6.032611 (lr=5.0756e-06) (hash(x)=63513096)
+9900 train 6.607420 (lr=9.1835e-05) (hash(x)=48142455)
+48800 val loss 5.6003
+48800 val perplexity 270.4991
+48800 train 5.218384 (lr=5.0644e-06) (hash(x)=44917721)
+10000 val loss 6.4405
+10000 val perplexity 626.7271
+10000 train 6.596187 (lr=9.1671e-05) (hash(x)=50757167)
+48900 val loss 5.5981
+48900 val perplexity 269.9094
+48900 train 5.816957 (lr=5.0542e-06) (hash(x)=51426559)
+49000 val loss 5.5992
+49000 val perplexity 270.2160
+49000 train 5.508775 (lr=5.0448e-06) (hash(x)=47909690)
+10100 val loss 6.4674
+10100 val perplexity 643.8083
+10100 train 6.442811 (lr=9.1506e-05) (hash(x)=46128585)
+49100 val loss 5.5971
+49100 val perplexity 269.6422
+49100 train 5.500440 (lr=5.0363e-06) (hash(x)=48329226)
+10200 val loss 6.4787
+10200 val perplexity 651.1212
+10200 train 6.703011 (lr=9.1339e-05) (hash(x)=57016296)
+49200 val loss 5.5991
+49200 val perplexity 270.1703
+49200 train 5.600343 (lr=5.0286e-06) (hash(x)=49512472)
+49300 val loss 5.5989
+49300 val perplexity 270.1354
+49300 train 5.452957 (lr=5.0219e-06) (hash(x)=46541146)
+10300 val loss 6.5369
+10300 val perplexity 690.1228
+10300 train 6.600325 (lr=9.1171e-05) (hash(x)=49839464)
+49400 val loss 5.6009
+49400 val perplexity 270.6745
+49400 train 5.543798 (lr=5.0161e-06) (hash(x)=52192792)
+10400 val loss 6.5206
+10400 val perplexity 678.9719
+10400 train 6.516890 (lr=9.1001e-05) (hash(x)=50237852)
+49500 val loss 5.5995
+49500 val perplexity 270.2928
+49500 train 5.386661 (lr=5.0112e-06) (hash(x)=48954331)
+49600 val loss 5.5985
+49600 val perplexity 270.0147
+49600 train 5.711846 (lr=5.0072e-06) (hash(x)=50364098)
+10500 val loss 6.5185
+10500 val perplexity 677.5903
+10500 train 6.626068 (lr=9.0830e-05) (hash(x)=48003282)
+49700 val loss 5.5968
+49700 val perplexity 269.5677
+49700 train 5.554048 (lr=5.0040e-06) (hash(x)=52016774)
+49800 val loss 5.5979
+49800 val perplexity 269.8479
+49800 train 5.830290 (lr=5.0018e-06) (hash(x)=54182957)
+10600 val loss 6.5319
+10600 val perplexity 686.7331
+10600 train 6.466997 (lr=9.0658e-05) (hash(x)=53088427)
+49900 val loss 5.5940
+49900 val perplexity 268.8109
+49900 train 5.332826 (lr=5.0004e-06) (hash(x)=48188126)
+10700 val loss 6.5500
+10700 val perplexity 699.2339
+10700 train 6.378216 (lr=9.0484e-05) (hash(x)=47687508)
+49999 val loss 5.6180
+49999 val perplexity 275.3266
+10800 val loss 6.5394
+10800 val perplexity 691.8988
+10800 train 6.397013 (lr=9.0308e-05) (hash(x)=47658698)
+10900 val loss 6.4993
+10900 val perplexity 664.6454
+10900 train 6.503762 (lr=9.0132e-05) (hash(x)=52957012)
+11000 val loss 6.4866
+11000 val perplexity 656.2906
+11000 train 6.625940 (lr=8.9954e-05) (hash(x)=54632498)
+11100 val loss 6.5099
+11100 val perplexity 671.7436
+11100 train 6.986665 (lr=8.9774e-05) (hash(x)=54028593)
+11200 val loss 6.4837
+11200 val perplexity 654.3682
+11200 train 6.616082 (lr=8.9593e-05) (hash(x)=54084412)
+11300 val loss 6.4871
+11300 val perplexity 656.6074
+11300 train 6.309471 (lr=8.9411e-05) (hash(x)=48170961)
+11400 val loss 6.4986
+11400 val perplexity 664.2282
+11400 train 6.540001 (lr=8.9227e-05) (hash(x)=52124243)

attention_kindselective_n_heads4_seed1339/model_02500.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ad93158bdfd32071e25dc456532ed3fd36c61df4cc86f865449960dcdf6fdfcb
+oid sha256:2ec7853f39c020bf0817d8a9424f0786ad4b93fc0839e1f8734dc36c9b628b3c
 size 92843394

attention_kindselective_n_heads4_seed1339/model_05000.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6aa78e6516bbc5459cdc3154c9b778836f34c2c1dbac08076e7043de9bd5fe9c
+oid sha256:e7a3ea4733ac190d2c44f57c3fc804dc59f7571ad4f028f02fe4775d0b66df1b
 size 92843394

attention_kindselective_n_heads4_seed1339/model_07500.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20de18d92ff73af4207c8758d2c668cb193ef56f2e7857b8391859c03bfce23d
+oid sha256:5d9cba9958a987856e63371998d573139f4d7e12bf83d1dc6741e08c1656ae1e
 size 92843394

attention_kindselective_n_heads4_seed1339/model_10000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75c9a83ca6838f6643993a01ade7f926d7dccd9aaabbaa9d1cc052a66d45ad6a
+size 92843394

attention_kindselective_n_heads4_seed1339/model_40000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78a8fd9d6a750808245efe7c72bed034116984b1a3d236604a1e986d60e7ed58
+size 92843394

attention_kindselective_n_heads4_seed1339/model_42500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3363e83947e45c3c64503ecbdcc0b270aa74e74aa9ec5438a596bd3c8156913
+size 92843394

attention_kindselective_n_heads4_seed1339/model_45000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:652267ce4a3c3eb24b3fdd1182cbeaad71435a28d6533b8d65f60dfd783e5411
+size 92843394

attention_kindselective_n_heads4_seed1339/model_47500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d198491e4dd117689476940ce9f230be59d99c41399afc7f16477ab0abe13f73
+size 92843394

attention_kindselective_n_heads4_seed1339/model_49999.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb07662f40aa23921523b0c43b25a496ce763a42d93ca236f9ab6e69e0f9407d
+size 92843394

attention_kindselective_n_heads4_seed1339/optimizer_02500.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:edeb87f6d21d339a0559282c4a2db7e96080f40d421400d4df3816f05159bf21
+oid sha256:6d0daaea20dfea170f7f2a3011fad8c68fd0d34d7837cfc7e1b82e10c8afac2e
 size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_05000.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ef24c1b7a4309998b53d8f8c3106922fe90cd2c76ffb6673a28994960c96400
+oid sha256:d6b399697378d1367510c9f4104e498e29bd1617ea27a7878564f423ee87160d
 size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_07500.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e1171237d3a16e4a7ecf8a131c6ba1d82e5ffafb1c91d7d035a9b67b6ae1398e
+oid sha256:827fa057c0c4bae82594e7d86c915d39ef2208d8da842bda634d1d07f8a6f14c
 size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_10000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdb56e2388c4d739ff044d95f0f99d11044dc7064e262ea9e0da4677e4b7369f
+size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_40000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ecc8125e65f3bc259ac039c34b488e49892d0778ea59a14ffd86b25c567c0ac
+size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_42500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a22bc9fe743fba8fc4b4560471a341de420ff16d5fb5457b184c26c331fefcdb
+size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_45000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9b0a4a35f4948303d8fd2c63401b8e68948d5d7ffed4e82c8947d42f9eace1f
+size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_47500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca09c067aab98fdf7e31137500fc5459871aa21e3a7a17c7d424ef21a92d9905
+size 179406214

attention_kindselective_n_heads4_seed1339/optimizer_49999.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67ac5a0dfae5b7e50ccf39b379788bfee75cb4871c0743d86aacdb6f3b54b940
+size 179406214