Upload folder using huggingface_hub
Browse files- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/args.json +1 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/dataloader_08749.pt +3 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/log2.txt +267 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/model_08749.pt +3 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/optimizer_08749.pt +3 -0
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/args.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"hellaswag": true, "attention_kind": "selective", "log_dir": "wider_is_better_10/lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338", "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": 8750, "warmup_steps": 500, "group": "wider_is_better_10", "use_wandb": true, "kill_self_after_run": false, "random_seed": 1338, "memory_penalty_epsilon": 0.1, "selection_head_linear_combo": "n_latent_masks", "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": 120, "total_batch_size": 61440, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_masks": null, "n_latent_masks": 2, "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": 6e-05, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "6e-5_61440_1338", "n_embd": 256}
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/dataloader_08749.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:953385078aa3787b69fc6857dfd48b0a2cd2f4d27c6f8892e01211aca53d07f5
|
| 3 |
+
size 964
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/log2.txt
ADDED
|
@@ -0,0 +1,267 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
max_steps: 8750
|
| 2 |
+
0 val loss 11.2524
|
| 3 |
+
0 val perplexity 77061.3750
|
| 4 |
+
0 train 11.253548 (lr=1.2000e-07) (hash(x)=150327452)
|
| 5 |
+
100 val loss 10.0800
|
| 6 |
+
100 val perplexity 23860.5059
|
| 7 |
+
100 train 10.110741 (lr=1.2120e-05) (hash(x)=166780046)
|
| 8 |
+
200 val loss 9.8451
|
| 9 |
+
200 val perplexity 18865.0098
|
| 10 |
+
200 train 9.856663 (lr=2.4120e-05) (hash(x)=155040610)
|
| 11 |
+
300 val loss 9.6589
|
| 12 |
+
300 val perplexity 15660.1348
|
| 13 |
+
300 train 9.699568 (lr=3.6120e-05) (hash(x)=155504036)
|
| 14 |
+
400 val loss 9.3308
|
| 15 |
+
400 val perplexity 11280.6152
|
| 16 |
+
400 train 9.324850 (lr=4.8120e-05) (hash(x)=143823248)
|
| 17 |
+
500 val loss 9.0827
|
| 18 |
+
500 val perplexity 8801.6543
|
| 19 |
+
500 train 9.094957 (lr=6.0000e-05) (hash(x)=143734685)
|
| 20 |
+
600 val loss 8.9223
|
| 21 |
+
600 val perplexity 7497.6660
|
| 22 |
+
600 train 8.886049 (lr=5.9980e-05) (hash(x)=150678249)
|
| 23 |
+
700 val loss 8.6511
|
| 24 |
+
700 val perplexity 5716.1870
|
| 25 |
+
700 train 8.739485 (lr=5.9922e-05) (hash(x)=175802021)
|
| 26 |
+
800 val loss 8.4693
|
| 27 |
+
800 val perplexity 4766.2109
|
| 28 |
+
800 train 8.462464 (lr=5.9824e-05) (hash(x)=158681215)
|
| 29 |
+
900 val loss 8.6006
|
| 30 |
+
900 val perplexity 5435.0195
|
| 31 |
+
900 train 8.579044 (lr=5.9687e-05) (hash(x)=146108145)
|
| 32 |
+
1000 val loss 8.3132
|
| 33 |
+
1000 val perplexity 4077.5107
|
| 34 |
+
1000 train 8.297157 (lr=5.9512e-05) (hash(x)=154996086)
|
| 35 |
+
1100 val loss 8.2820
|
| 36 |
+
1100 val perplexity 3952.1191
|
| 37 |
+
1100 train 8.307036 (lr=5.9298e-05) (hash(x)=153885445)
|
| 38 |
+
1200 val loss 8.2964
|
| 39 |
+
1200 val perplexity 4009.6040
|
| 40 |
+
1200 train 8.308805 (lr=5.9046e-05) (hash(x)=142353087)
|
| 41 |
+
1300 val loss 8.2845
|
| 42 |
+
1300 val perplexity 3961.9912
|
| 43 |
+
1300 train 8.296026 (lr=5.8757e-05) (hash(x)=150750353)
|
| 44 |
+
1400 val loss 8.2955
|
| 45 |
+
1400 val perplexity 4005.9958
|
| 46 |
+
1400 train 8.237262 (lr=5.8430e-05) (hash(x)=152767913)
|
| 47 |
+
1500 val loss 8.2893
|
| 48 |
+
1500 val perplexity 3981.2324
|
| 49 |
+
1500 train 8.343046 (lr=5.8066e-05) (hash(x)=151562048)
|
| 50 |
+
1600 val loss 8.2925
|
| 51 |
+
1600 val perplexity 3993.6597
|
| 52 |
+
1600 train 8.273527 (lr=5.7666e-05) (hash(x)=166486165)
|
| 53 |
+
1700 val loss 8.3093
|
| 54 |
+
1700 val perplexity 4061.3157
|
| 55 |
+
1700 train 8.232641 (lr=5.7230e-05) (hash(x)=130835396)
|
| 56 |
+
1800 val loss 8.2637
|
| 57 |
+
1800 val perplexity 3880.2791
|
| 58 |
+
1800 train 8.302708 (lr=5.6759e-05) (hash(x)=158851816)
|
| 59 |
+
1900 val loss 8.2549
|
| 60 |
+
1900 val perplexity 3846.5444
|
| 61 |
+
1900 train 8.269764 (lr=5.6253e-05) (hash(x)=153313879)
|
| 62 |
+
2000 val loss 8.2642
|
| 63 |
+
2000 val perplexity 3882.2075
|
| 64 |
+
2000 train 8.243088 (lr=5.5714e-05) (hash(x)=158245023)
|
| 65 |
+
2100 val loss 8.2627
|
| 66 |
+
2100 val perplexity 3876.6284
|
| 67 |
+
2100 train 8.180153 (lr=5.5142e-05) (hash(x)=157204896)
|
| 68 |
+
2200 val loss 8.2419
|
| 69 |
+
2200 val perplexity 3796.5608
|
| 70 |
+
2200 train 8.225559 (lr=5.4537e-05) (hash(x)=137541932)
|
| 71 |
+
2300 val loss 8.2622
|
| 72 |
+
2300 val perplexity 3874.6990
|
| 73 |
+
2300 train 8.268251 (lr=5.3902e-05) (hash(x)=150149692)
|
| 74 |
+
2400 val loss 8.3845
|
| 75 |
+
2400 val perplexity 4378.5288
|
| 76 |
+
2400 train 8.434961 (lr=5.3236e-05) (hash(x)=151730720)
|
| 77 |
+
2500 val loss 8.2590
|
| 78 |
+
2500 val perplexity 3862.1814
|
| 79 |
+
2500 train 8.027147 (lr=5.2541e-05) (hash(x)=143406752)
|
| 80 |
+
2600 val loss 8.2304
|
| 81 |
+
2600 val perplexity 3753.4353
|
| 82 |
+
2600 train 8.191346 (lr=5.1817e-05) (hash(x)=157272496)
|
| 83 |
+
2700 val loss 8.2296
|
| 84 |
+
2700 val perplexity 3750.4656
|
| 85 |
+
2700 train 8.293442 (lr=5.1067e-05) (hash(x)=155342327)
|
| 86 |
+
2800 val loss 8.2184
|
| 87 |
+
2800 val perplexity 3708.7090
|
| 88 |
+
2800 train 8.141396 (lr=5.0290e-05) (hash(x)=140626679)
|
| 89 |
+
2900 val loss 8.2138
|
| 90 |
+
2900 val perplexity 3691.7178
|
| 91 |
+
2900 train 8.162292 (lr=4.9487e-05) (hash(x)=144953350)
|
| 92 |
+
3000 val loss 8.2398
|
| 93 |
+
3000 val perplexity 3788.6865
|
| 94 |
+
3000 train 8.280184 (lr=4.8662e-05) (hash(x)=172449837)
|
| 95 |
+
3100 val loss 8.2512
|
| 96 |
+
3100 val perplexity 3832.2061
|
| 97 |
+
3100 train 8.230724 (lr=4.7813e-05) (hash(x)=141710086)
|
| 98 |
+
3200 val loss 8.2153
|
| 99 |
+
3200 val perplexity 3697.0061
|
| 100 |
+
3200 train 8.172878 (lr=4.6943e-05) (hash(x)=151299772)
|
| 101 |
+
3300 val loss 8.2158
|
| 102 |
+
3300 val perplexity 3698.9880
|
| 103 |
+
3300 train 8.188112 (lr=4.6052e-05) (hash(x)=146473110)
|
| 104 |
+
3400 val loss 8.2062
|
| 105 |
+
3400 val perplexity 3663.7258
|
| 106 |
+
3400 train 8.266872 (lr=4.5143e-05) (hash(x)=153954157)
|
| 107 |
+
3500 val loss 8.2016
|
| 108 |
+
3500 val perplexity 3646.8433
|
| 109 |
+
3500 train 8.141716 (lr=4.4216e-05) (hash(x)=153717336)
|
| 110 |
+
3600 val loss 8.2424
|
| 111 |
+
3600 val perplexity 3798.6143
|
| 112 |
+
3600 train 8.138775 (lr=4.3273e-05) (hash(x)=144965161)
|
| 113 |
+
3700 val loss 8.1980
|
| 114 |
+
3700 val perplexity 3633.8105
|
| 115 |
+
3700 train 8.066285 (lr=4.2315e-05) (hash(x)=125969741)
|
| 116 |
+
3800 val loss 8.1957
|
| 117 |
+
3800 val perplexity 3625.3232
|
| 118 |
+
3800 train 8.171384 (lr=4.1343e-05) (hash(x)=155070487)
|
| 119 |
+
3900 val loss 8.1979
|
| 120 |
+
3900 val perplexity 3633.4087
|
| 121 |
+
3900 train 8.142708 (lr=4.0360e-05) (hash(x)=149444644)
|
| 122 |
+
4000 val loss 8.1966
|
| 123 |
+
4000 val perplexity 3628.7200
|
| 124 |
+
4000 train 8.213164 (lr=3.9365e-05) (hash(x)=151663033)
|
| 125 |
+
4100 val loss 8.1984
|
| 126 |
+
4100 val perplexity 3635.1208
|
| 127 |
+
4100 train 8.284185 (lr=3.8362e-05) (hash(x)=143688282)
|
| 128 |
+
4200 val loss 8.2049
|
| 129 |
+
4200 val perplexity 3658.7188
|
| 130 |
+
4200 train 8.179093 (lr=3.7351e-05) (hash(x)=163361651)
|
| 131 |
+
4300 val loss 8.1999
|
| 132 |
+
4300 val perplexity 3640.7378
|
| 133 |
+
4300 train 8.385876 (lr=3.6333e-05) (hash(x)=153619361)
|
| 134 |
+
4400 val loss 8.2212
|
| 135 |
+
4400 val perplexity 3718.9766
|
| 136 |
+
4400 train 8.387505 (lr=3.5311e-05) (hash(x)=168527064)
|
| 137 |
+
4500 val loss 8.2074
|
| 138 |
+
4500 val perplexity 3668.1589
|
| 139 |
+
4500 train 8.171785 (lr=3.4285e-05) (hash(x)=125588037)
|
| 140 |
+
4600 val loss 8.1562
|
| 141 |
+
4600 val perplexity 3484.7673
|
| 142 |
+
4600 train 8.087983 (lr=3.3257e-05) (hash(x)=143710941)
|
| 143 |
+
4700 val loss 8.1115
|
| 144 |
+
4700 val perplexity 3332.6169
|
| 145 |
+
4700 train 7.910356 (lr=3.2229e-05) (hash(x)=150952742)
|
| 146 |
+
4800 val loss 8.0541
|
| 147 |
+
4800 val perplexity 3146.7600
|
| 148 |
+
4800 train 7.988973 (lr=3.1202e-05) (hash(x)=145323659)
|
| 149 |
+
4900 val loss 8.0506
|
| 150 |
+
4900 val perplexity 3135.6790
|
| 151 |
+
4900 train 8.083582 (lr=3.0178e-05) (hash(x)=153151397)
|
| 152 |
+
5000 val loss 8.0310
|
| 153 |
+
5000 val perplexity 3074.7332
|
| 154 |
+
5000 train 7.980440 (lr=2.9157e-05) (hash(x)=143182059)
|
| 155 |
+
5100 val loss 8.0517
|
| 156 |
+
5100 val perplexity 3139.2427
|
| 157 |
+
5100 train 8.127379 (lr=2.8143e-05) (hash(x)=170083586)
|
| 158 |
+
5200 val loss 8.0293
|
| 159 |
+
5200 val perplexity 3069.6060
|
| 160 |
+
5200 train 8.001753 (lr=2.7135e-05) (hash(x)=149363919)
|
| 161 |
+
5300 val loss 8.0128
|
| 162 |
+
5300 val perplexity 3019.3169
|
| 163 |
+
5300 train 8.077737 (lr=2.6136e-05) (hash(x)=152033784)
|
| 164 |
+
5400 val loss 8.0620
|
| 165 |
+
5400 val perplexity 3171.7368
|
| 166 |
+
5400 train 8.036554 (lr=2.5147e-05) (hash(x)=154614289)
|
| 167 |
+
5500 val loss 8.0108
|
| 168 |
+
5500 val perplexity 3013.4399
|
| 169 |
+
5500 train 8.097231 (lr=2.4169e-05) (hash(x)=157745174)
|
| 170 |
+
5600 val loss 8.0236
|
| 171 |
+
5600 val perplexity 3052.2285
|
| 172 |
+
5600 train 7.957674 (lr=2.3204e-05) (hash(x)=147693222)
|
| 173 |
+
5700 val loss 8.0121
|
| 174 |
+
5700 val perplexity 3017.2300
|
| 175 |
+
5700 train 7.969118 (lr=2.2253e-05) (hash(x)=149784627)
|
| 176 |
+
5800 val loss 8.0002
|
| 177 |
+
5800 val perplexity 2981.4414
|
| 178 |
+
5800 train 7.991749 (lr=2.1318e-05) (hash(x)=158620729)
|
| 179 |
+
5900 val loss 7.9887
|
| 180 |
+
5900 val perplexity 2947.3838
|
| 181 |
+
5900 train 7.935953 (lr=2.0400e-05) (hash(x)=159763910)
|
| 182 |
+
6000 val loss 7.9742
|
| 183 |
+
6000 val perplexity 2905.1582
|
| 184 |
+
6000 train 7.939895 (lr=1.9500e-05) (hash(x)=147640561)
|
| 185 |
+
6100 val loss 7.9775
|
| 186 |
+
6100 val perplexity 2914.5393
|
| 187 |
+
6100 train 8.000101 (lr=1.8620e-05) (hash(x)=156613394)
|
| 188 |
+
6200 val loss 7.9554
|
| 189 |
+
6200 val perplexity 2850.8755
|
| 190 |
+
6200 train 8.084249 (lr=1.7760e-05) (hash(x)=186221290)
|
| 191 |
+
6300 val loss 7.9477
|
| 192 |
+
6300 val perplexity 2828.9556
|
| 193 |
+
6300 train 7.887838 (lr=1.6923e-05) (hash(x)=152081419)
|
| 194 |
+
6400 val loss 7.9432
|
| 195 |
+
6400 val perplexity 2816.4341
|
| 196 |
+
6400 train 7.907034 (lr=1.6108e-05) (hash(x)=154808349)
|
| 197 |
+
6500 val loss 7.9465
|
| 198 |
+
6500 val perplexity 2825.6970
|
| 199 |
+
6500 train 7.908285 (lr=1.5319e-05) (hash(x)=159437208)
|
| 200 |
+
6600 val loss 7.9413
|
| 201 |
+
6600 val perplexity 2811.1423
|
| 202 |
+
6600 train 7.853995 (lr=1.4555e-05) (hash(x)=157933074)
|
| 203 |
+
6700 val loss 7.9418
|
| 204 |
+
6700 val perplexity 2812.2805
|
| 205 |
+
6700 train 7.949341 (lr=1.3817e-05) (hash(x)=161560240)
|
| 206 |
+
6800 val loss 7.9350
|
| 207 |
+
6800 val perplexity 2793.3040
|
| 208 |
+
6800 train 7.891088 (lr=1.3108e-05) (hash(x)=155424292)
|
| 209 |
+
6900 val loss 7.9357
|
| 210 |
+
6900 val perplexity 2795.4253
|
| 211 |
+
6900 train 7.958529 (lr=1.2427e-05) (hash(x)=148561470)
|
| 212 |
+
7000 val loss 7.9329
|
| 213 |
+
7000 val perplexity 2787.5876
|
| 214 |
+
7000 train 7.877773 (lr=1.1777e-05) (hash(x)=141527450)
|
| 215 |
+
7100 val loss 7.9321
|
| 216 |
+
7100 val perplexity 2785.2307
|
| 217 |
+
7100 train 7.904116 (lr=1.1157e-05) (hash(x)=151066339)
|
| 218 |
+
7200 val loss 7.9393
|
| 219 |
+
7200 val perplexity 2805.4858
|
| 220 |
+
7200 train 7.909233 (lr=1.0568e-05) (hash(x)=155231264)
|
| 221 |
+
7300 val loss 7.9481
|
| 222 |
+
7300 val perplexity 2830.2332
|
| 223 |
+
7300 train 7.962085 (lr=1.0012e-05) (hash(x)=150281149)
|
| 224 |
+
7400 val loss 7.9688
|
| 225 |
+
7400 val perplexity 2889.2656
|
| 226 |
+
7400 train 7.895494 (lr=9.4899e-06) (hash(x)=148421717)
|
| 227 |
+
7500 val loss 7.9474
|
| 228 |
+
7500 val perplexity 2828.2634
|
| 229 |
+
7500 train 7.826111 (lr=9.0014e-06) (hash(x)=146921118)
|
| 230 |
+
7600 val loss 7.9379
|
| 231 |
+
7600 val perplexity 2801.5449
|
| 232 |
+
7600 train 8.004966 (lr=8.5478e-06) (hash(x)=150660048)
|
| 233 |
+
7700 val loss 7.9298
|
| 234 |
+
7700 val perplexity 2778.9531
|
| 235 |
+
7700 train 7.889890 (lr=8.1297e-06) (hash(x)=148059852)
|
| 236 |
+
7800 val loss 7.9186
|
| 237 |
+
7800 val perplexity 2747.9478
|
| 238 |
+
7800 train 7.761279 (lr=7.7476e-06) (hash(x)=148331002)
|
| 239 |
+
7900 val loss 7.9113
|
| 240 |
+
7900 val perplexity 2728.0078
|
| 241 |
+
7900 train 7.897078 (lr=7.4021e-06) (hash(x)=164923883)
|
| 242 |
+
8000 val loss 7.9089
|
| 243 |
+
8000 val perplexity 2721.4011
|
| 244 |
+
8000 train 7.841110 (lr=7.0937e-06) (hash(x)=143545384)
|
| 245 |
+
8100 val loss 7.9112
|
| 246 |
+
8100 val perplexity 2727.7437
|
| 247 |
+
8100 train 7.907278 (lr=6.8229e-06) (hash(x)=160686959)
|
| 248 |
+
8200 val loss 7.9073
|
| 249 |
+
8200 val perplexity 2717.0005
|
| 250 |
+
8200 train 7.982006 (lr=6.5900e-06) (hash(x)=156501889)
|
| 251 |
+
8300 val loss 7.9087
|
| 252 |
+
8300 val perplexity 2720.9185
|
| 253 |
+
8300 train 7.858608 (lr=6.3954e-06) (hash(x)=142716875)
|
| 254 |
+
8400 val loss 7.9064
|
| 255 |
+
8400 val perplexity 2714.7188
|
| 256 |
+
8400 train 7.816538 (lr=6.2395e-06) (hash(x)=154436684)
|
| 257 |
+
8500 val loss 7.9057
|
| 258 |
+
8500 val perplexity 2712.7000
|
| 259 |
+
8500 train 7.910065 (lr=6.1223e-06) (hash(x)=147965839)
|
| 260 |
+
8600 val loss 7.9059
|
| 261 |
+
8600 val perplexity 2713.3364
|
| 262 |
+
8600 train 7.847117 (lr=6.0440e-06) (hash(x)=145228097)
|
| 263 |
+
8700 val loss 7.9046
|
| 264 |
+
8700 val perplexity 2709.6787
|
| 265 |
+
8700 train 8.005460 (lr=6.0049e-06) (hash(x)=152910357)
|
| 266 |
+
8749 val loss 7.9064
|
| 267 |
+
8749 val perplexity 2714.6890
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/model_08749.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:72ac3fd834e67725f07b094d453c8dd81bab6d2c285179aeb657cf088dc5ba38
|
| 3 |
+
size 95220594
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1338/optimizer_08749.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6073bbe3959570287e57667148c646cacc12ab4898e35772d6978b9c47a6086e
|
| 3 |
+
size 184163894
|