Upload folder using huggingface_hub
Browse files- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/args.json +1 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/dataloader_08749.pt +3 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/log2.txt +267 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/model_08749.pt +3 -0
- lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/optimizer_08749.pt +3 -0
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/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_seed1341", "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": 1341, "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_1341", "n_embd": 256}
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/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_seed1341/log2.txt
ADDED
|
@@ -0,0 +1,267 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
max_steps: 8750
|
| 2 |
+
0 val loss 11.3022
|
| 3 |
+
0 val perplexity 81003.2969
|
| 4 |
+
0 train 11.302774 (lr=1.2000e-07) (hash(x)=145079536)
|
| 5 |
+
100 val loss 10.1813
|
| 6 |
+
100 val perplexity 26404.1973
|
| 7 |
+
100 train 10.158878 (lr=1.2120e-05) (hash(x)=155800595)
|
| 8 |
+
200 val loss 9.9129
|
| 9 |
+
200 val perplexity 20188.7715
|
| 10 |
+
200 train 9.868517 (lr=2.4120e-05) (hash(x)=145606733)
|
| 11 |
+
300 val loss 9.6447
|
| 12 |
+
300 val perplexity 15439.3730
|
| 13 |
+
300 train 9.670784 (lr=3.6120e-05) (hash(x)=150367139)
|
| 14 |
+
400 val loss 9.3377
|
| 15 |
+
400 val perplexity 11357.7764
|
| 16 |
+
400 train 9.477560 (lr=4.8120e-05) (hash(x)=155747374)
|
| 17 |
+
500 val loss 9.1418
|
| 18 |
+
500 val perplexity 9337.9492
|
| 19 |
+
500 train 9.147530 (lr=6.0000e-05) (hash(x)=140604760)
|
| 20 |
+
600 val loss 8.9356
|
| 21 |
+
600 val perplexity 7597.6377
|
| 22 |
+
600 train 8.935740 (lr=5.9980e-05) (hash(x)=148404734)
|
| 23 |
+
700 val loss 8.7538
|
| 24 |
+
700 val perplexity 6334.8677
|
| 25 |
+
700 train 8.754250 (lr=5.9922e-05) (hash(x)=148115934)
|
| 26 |
+
800 val loss 8.6325
|
| 27 |
+
800 val perplexity 5611.0654
|
| 28 |
+
800 train 8.589380 (lr=5.9824e-05) (hash(x)=137464699)
|
| 29 |
+
900 val loss 8.5862
|
| 30 |
+
900 val perplexity 5357.2109
|
| 31 |
+
900 train 8.529787 (lr=5.9687e-05) (hash(x)=143886042)
|
| 32 |
+
1000 val loss 8.5439
|
| 33 |
+
1000 val perplexity 5135.5269
|
| 34 |
+
1000 train 8.577871 (lr=5.9512e-05) (hash(x)=163799796)
|
| 35 |
+
1100 val loss 8.4062
|
| 36 |
+
1100 val perplexity 4474.9478
|
| 37 |
+
1100 train 8.374664 (lr=5.9298e-05) (hash(x)=144592844)
|
| 38 |
+
1200 val loss 8.4067
|
| 39 |
+
1200 val perplexity 4477.0439
|
| 40 |
+
1200 train 8.628639 (lr=5.9046e-05) (hash(x)=204706354)
|
| 41 |
+
1300 val loss 8.4059
|
| 42 |
+
1300 val perplexity 4473.3350
|
| 43 |
+
1300 train 8.400589 (lr=5.8757e-05) (hash(x)=150862210)
|
| 44 |
+
1400 val loss 8.3872
|
| 45 |
+
1400 val perplexity 4390.5337
|
| 46 |
+
1400 train 8.388836 (lr=5.8430e-05) (hash(x)=147766811)
|
| 47 |
+
1500 val loss 8.3914
|
| 48 |
+
1500 val perplexity 4408.8359
|
| 49 |
+
1500 train 8.313634 (lr=5.8066e-05) (hash(x)=135925327)
|
| 50 |
+
1600 val loss 8.3716
|
| 51 |
+
1600 val perplexity 4322.3784
|
| 52 |
+
1600 train 8.422943 (lr=5.7666e-05) (hash(x)=160440642)
|
| 53 |
+
1700 val loss 8.4504
|
| 54 |
+
1700 val perplexity 4676.9580
|
| 55 |
+
1700 train 8.510411 (lr=5.7230e-05) (hash(x)=151184106)
|
| 56 |
+
1800 val loss 8.3393
|
| 57 |
+
1800 val perplexity 4185.3354
|
| 58 |
+
1800 train 8.295940 (lr=5.6759e-05) (hash(x)=148100580)
|
| 59 |
+
1900 val loss 8.3477
|
| 60 |
+
1900 val perplexity 4220.3062
|
| 61 |
+
1900 train 8.342902 (lr=5.6253e-05) (hash(x)=149434659)
|
| 62 |
+
2000 val loss 8.3434
|
| 63 |
+
2000 val perplexity 4202.2373
|
| 64 |
+
2000 train 8.225173 (lr=5.5714e-05) (hash(x)=152285486)
|
| 65 |
+
2100 val loss 8.3391
|
| 66 |
+
2100 val perplexity 4184.3057
|
| 67 |
+
2100 train 8.232183 (lr=5.5142e-05) (hash(x)=144294295)
|
| 68 |
+
2200 val loss 8.3419
|
| 69 |
+
2200 val perplexity 4196.0781
|
| 70 |
+
2200 train 8.402203 (lr=5.4537e-05) (hash(x)=175030215)
|
| 71 |
+
2300 val loss 8.3638
|
| 72 |
+
2300 val perplexity 4289.0151
|
| 73 |
+
2300 train 8.304497 (lr=5.3902e-05) (hash(x)=150831428)
|
| 74 |
+
2400 val loss 8.3439
|
| 75 |
+
2400 val perplexity 4204.6426
|
| 76 |
+
2400 train 8.434280 (lr=5.3236e-05) (hash(x)=140808297)
|
| 77 |
+
2500 val loss 8.3428
|
| 78 |
+
2500 val perplexity 4199.7212
|
| 79 |
+
2500 train 8.343163 (lr=5.2541e-05) (hash(x)=153160275)
|
| 80 |
+
2600 val loss 8.3510
|
| 81 |
+
2600 val perplexity 4234.5576
|
| 82 |
+
2600 train 8.256983 (lr=5.1817e-05) (hash(x)=133990623)
|
| 83 |
+
2700 val loss 8.3480
|
| 84 |
+
2700 val perplexity 4221.7510
|
| 85 |
+
2700 train 8.404637 (lr=5.1067e-05) (hash(x)=142860944)
|
| 86 |
+
2800 val loss 8.3501
|
| 87 |
+
2800 val perplexity 4230.7393
|
| 88 |
+
2800 train 8.444883 (lr=5.0290e-05) (hash(x)=137959511)
|
| 89 |
+
2900 val loss 8.3519
|
| 90 |
+
2900 val perplexity 4238.2017
|
| 91 |
+
2900 train 8.355026 (lr=4.9487e-05) (hash(x)=147009873)
|
| 92 |
+
3000 val loss 8.3410
|
| 93 |
+
3000 val perplexity 4192.4863
|
| 94 |
+
3000 train 8.334948 (lr=4.8662e-05) (hash(x)=158264841)
|
| 95 |
+
3100 val loss 8.3390
|
| 96 |
+
3100 val perplexity 4183.8027
|
| 97 |
+
3100 train 8.372059 (lr=4.7813e-05) (hash(x)=139232251)
|
| 98 |
+
3200 val loss 8.3396
|
| 99 |
+
3200 val perplexity 4186.4932
|
| 100 |
+
3200 train 8.349333 (lr=4.6943e-05) (hash(x)=153436104)
|
| 101 |
+
3300 val loss 8.3409
|
| 102 |
+
3300 val perplexity 4191.8623
|
| 103 |
+
3300 train 8.275049 (lr=4.6052e-05) (hash(x)=149681831)
|
| 104 |
+
3400 val loss 8.3368
|
| 105 |
+
3400 val perplexity 4174.8350
|
| 106 |
+
3400 train 8.394770 (lr=4.5143e-05) (hash(x)=168425516)
|
| 107 |
+
3500 val loss 8.3488
|
| 108 |
+
3500 val perplexity 4225.0698
|
| 109 |
+
3500 train 8.363139 (lr=4.4216e-05) (hash(x)=163104338)
|
| 110 |
+
3600 val loss 8.3458
|
| 111 |
+
3600 val perplexity 4212.2642
|
| 112 |
+
3600 train 8.373493 (lr=4.3273e-05) (hash(x)=165109772)
|
| 113 |
+
3700 val loss 8.3445
|
| 114 |
+
3700 val perplexity 4206.8804
|
| 115 |
+
3700 train 8.283165 (lr=4.2315e-05) (hash(x)=153420306)
|
| 116 |
+
3800 val loss 8.3340
|
| 117 |
+
3800 val perplexity 4162.9238
|
| 118 |
+
3800 train 8.341437 (lr=4.1343e-05) (hash(x)=160168863)
|
| 119 |
+
3900 val loss 8.3616
|
| 120 |
+
3900 val perplexity 4279.3281
|
| 121 |
+
3900 train 8.420931 (lr=4.0360e-05) (hash(x)=153906073)
|
| 122 |
+
4000 val loss 8.3633
|
| 123 |
+
4000 val perplexity 4286.7129
|
| 124 |
+
4000 train 8.344972 (lr=3.9365e-05) (hash(x)=151055067)
|
| 125 |
+
4100 val loss 8.3393
|
| 126 |
+
4100 val perplexity 4184.9761
|
| 127 |
+
4100 train 8.326389 (lr=3.8362e-05) (hash(x)=149629830)
|
| 128 |
+
4200 val loss 8.3403
|
| 129 |
+
4200 val perplexity 4189.2769
|
| 130 |
+
4200 train 8.299853 (lr=3.7351e-05) (hash(x)=143101381)
|
| 131 |
+
4300 val loss 8.3360
|
| 132 |
+
4300 val perplexity 4171.4883
|
| 133 |
+
4300 train 8.311817 (lr=3.6333e-05) (hash(x)=149712044)
|
| 134 |
+
4400 val loss 8.3316
|
| 135 |
+
4400 val perplexity 4152.9551
|
| 136 |
+
4400 train 8.305692 (lr=3.5311e-05) (hash(x)=153446449)
|
| 137 |
+
4500 val loss 8.3397
|
| 138 |
+
4500 val perplexity 4186.8765
|
| 139 |
+
4500 train 8.340549 (lr=3.4285e-05) (hash(x)=146086947)
|
| 140 |
+
4600 val loss 8.3304
|
| 141 |
+
4600 val perplexity 4147.9678
|
| 142 |
+
4600 train 8.301800 (lr=3.3257e-05) (hash(x)=153800173)
|
| 143 |
+
4700 val loss 8.3333
|
| 144 |
+
4700 val perplexity 4159.9946
|
| 145 |
+
4700 train 8.280998 (lr=3.2229e-05) (hash(x)=155962726)
|
| 146 |
+
4800 val loss 8.3364
|
| 147 |
+
4800 val perplexity 4173.0518
|
| 148 |
+
4800 train 8.531429 (lr=3.1202e-05) (hash(x)=142045616)
|
| 149 |
+
4900 val loss 8.3406
|
| 150 |
+
4900 val perplexity 4190.6636
|
| 151 |
+
4900 train 8.315807 (lr=3.0178e-05) (hash(x)=143418248)
|
| 152 |
+
5000 val loss 8.3419
|
| 153 |
+
5000 val perplexity 4195.9302
|
| 154 |
+
5000 train 8.341086 (lr=2.9157e-05) (hash(x)=145789790)
|
| 155 |
+
5100 val loss 8.3421
|
| 156 |
+
5100 val perplexity 4196.7666
|
| 157 |
+
5100 train 8.367014 (lr=2.8143e-05) (hash(x)=137795633)
|
| 158 |
+
5200 val loss 8.3686
|
| 159 |
+
5200 val perplexity 4309.4624
|
| 160 |
+
5200 train 8.494040 (lr=2.7135e-05) (hash(x)=148907132)
|
| 161 |
+
5300 val loss 8.3396
|
| 162 |
+
5300 val perplexity 4186.5088
|
| 163 |
+
5300 train 8.326537 (lr=2.6136e-05) (hash(x)=152343580)
|
| 164 |
+
5400 val loss 8.3411
|
| 165 |
+
5400 val perplexity 4192.8940
|
| 166 |
+
5400 train 8.303223 (lr=2.5147e-05) (hash(x)=148578264)
|
| 167 |
+
5500 val loss 8.3422
|
| 168 |
+
5500 val perplexity 4197.2549
|
| 169 |
+
5500 train 8.441550 (lr=2.4169e-05) (hash(x)=145635833)
|
| 170 |
+
5600 val loss 8.3577
|
| 171 |
+
5600 val perplexity 4262.8521
|
| 172 |
+
5600 train 8.384826 (lr=2.3204e-05) (hash(x)=156337844)
|
| 173 |
+
5700 val loss 8.3529
|
| 174 |
+
5700 val perplexity 4242.5288
|
| 175 |
+
5700 train 8.379089 (lr=2.2253e-05) (hash(x)=147168506)
|
| 176 |
+
5800 val loss 8.3541
|
| 177 |
+
5800 val perplexity 4247.7271
|
| 178 |
+
5800 train 8.399971 (lr=2.1318e-05) (hash(x)=159566920)
|
| 179 |
+
5900 val loss 8.3549
|
| 180 |
+
5900 val perplexity 4250.8477
|
| 181 |
+
5900 train 8.418188 (lr=2.0400e-05) (hash(x)=158273929)
|
| 182 |
+
6000 val loss 8.3512
|
| 183 |
+
6000 val perplexity 4235.2886
|
| 184 |
+
6000 train 8.370876 (lr=1.9500e-05) (hash(x)=156649749)
|
| 185 |
+
6100 val loss 8.3442
|
| 186 |
+
6100 val perplexity 4205.6851
|
| 187 |
+
6100 train 8.319452 (lr=1.8620e-05) (hash(x)=146812388)
|
| 188 |
+
6200 val loss 8.3447
|
| 189 |
+
6200 val perplexity 4208.0122
|
| 190 |
+
6200 train 8.330069 (lr=1.7760e-05) (hash(x)=143522146)
|
| 191 |
+
6300 val loss 8.3403
|
| 192 |
+
6300 val perplexity 4189.2168
|
| 193 |
+
6300 train 8.282401 (lr=1.6923e-05) (hash(x)=150124474)
|
| 194 |
+
6400 val loss 8.3373
|
| 195 |
+
6400 val perplexity 4176.6748
|
| 196 |
+
6400 train 8.314153 (lr=1.6108e-05) (hash(x)=141242117)
|
| 197 |
+
6500 val loss 8.3378
|
| 198 |
+
6500 val perplexity 4178.8345
|
| 199 |
+
6500 train 8.311840 (lr=1.5319e-05) (hash(x)=143529762)
|
| 200 |
+
6600 val loss 8.3344
|
| 201 |
+
6600 val perplexity 4164.8696
|
| 202 |
+
6600 train 8.261427 (lr=1.4555e-05) (hash(x)=136948374)
|
| 203 |
+
6700 val loss 8.3367
|
| 204 |
+
6700 val perplexity 4174.3452
|
| 205 |
+
6700 train 8.335506 (lr=1.3817e-05) (hash(x)=146268592)
|
| 206 |
+
6800 val loss 8.3317
|
| 207 |
+
6800 val perplexity 4153.5806
|
| 208 |
+
6800 train 8.322599 (lr=1.3108e-05) (hash(x)=152676836)
|
| 209 |
+
6900 val loss 8.3327
|
| 210 |
+
6900 val perplexity 4157.7656
|
| 211 |
+
6900 train 8.323095 (lr=1.2427e-05) (hash(x)=134657776)
|
| 212 |
+
7000 val loss 8.3310
|
| 213 |
+
7000 val perplexity 4150.4053
|
| 214 |
+
7000 train 8.412806 (lr=1.1777e-05) (hash(x)=166721861)
|
| 215 |
+
7100 val loss 8.3285
|
| 216 |
+
7100 val perplexity 4140.0005
|
| 217 |
+
7100 train 8.277332 (lr=1.1157e-05) (hash(x)=135496702)
|
| 218 |
+
7200 val loss 8.3322
|
| 219 |
+
7200 val perplexity 4155.3638
|
| 220 |
+
7200 train 8.379076 (lr=1.0568e-05) (hash(x)=155567461)
|
| 221 |
+
7300 val loss 8.3251
|
| 222 |
+
7300 val perplexity 4126.2163
|
| 223 |
+
7300 train 8.254663 (lr=1.0012e-05) (hash(x)=142803829)
|
| 224 |
+
7400 val loss 8.3322
|
| 225 |
+
7400 val perplexity 4155.6250
|
| 226 |
+
7400 train 8.265259 (lr=9.4899e-06) (hash(x)=145294178)
|
| 227 |
+
7500 val loss 8.3261
|
| 228 |
+
7500 val perplexity 4130.3462
|
| 229 |
+
7500 train 8.223894 (lr=9.0014e-06) (hash(x)=150573713)
|
| 230 |
+
7600 val loss 8.3267
|
| 231 |
+
7600 val perplexity 4132.7104
|
| 232 |
+
7600 train 8.316321 (lr=8.5478e-06) (hash(x)=142771511)
|
| 233 |
+
7700 val loss 8.3275
|
| 234 |
+
7700 val perplexity 4136.1567
|
| 235 |
+
7700 train 8.304243 (lr=8.1297e-06) (hash(x)=143602175)
|
| 236 |
+
7800 val loss 8.3305
|
| 237 |
+
7800 val perplexity 4148.6875
|
| 238 |
+
7800 train 8.387541 (lr=7.7476e-06) (hash(x)=152379862)
|
| 239 |
+
7900 val loss 8.3261
|
| 240 |
+
7900 val perplexity 4130.3384
|
| 241 |
+
7900 train 8.336418 (lr=7.4021e-06) (hash(x)=146655921)
|
| 242 |
+
8000 val loss 8.3267
|
| 243 |
+
8000 val perplexity 4132.8447
|
| 244 |
+
8000 train 8.387903 (lr=7.0937e-06) (hash(x)=148262482)
|
| 245 |
+
8100 val loss 8.3324
|
| 246 |
+
8100 val perplexity 4156.3071
|
| 247 |
+
8100 train 8.308302 (lr=6.8229e-06) (hash(x)=147683655)
|
| 248 |
+
8200 val loss 8.3241
|
| 249 |
+
8200 val perplexity 4122.0786
|
| 250 |
+
8200 train 8.360710 (lr=6.5900e-06) (hash(x)=157312987)
|
| 251 |
+
8300 val loss 8.3282
|
| 252 |
+
8300 val perplexity 4138.7725
|
| 253 |
+
8300 train 8.397158 (lr=6.3954e-06) (hash(x)=141107543)
|
| 254 |
+
8400 val loss 8.3242
|
| 255 |
+
8400 val perplexity 4122.5781
|
| 256 |
+
8400 train 8.281261 (lr=6.2395e-06) (hash(x)=141323024)
|
| 257 |
+
8500 val loss 8.3244
|
| 258 |
+
8500 val perplexity 4123.1050
|
| 259 |
+
8500 train 8.405952 (lr=6.1223e-06) (hash(x)=150696521)
|
| 260 |
+
8600 val loss 8.3265
|
| 261 |
+
8600 val perplexity 4131.8198
|
| 262 |
+
8600 train 8.440543 (lr=6.0440e-06) (hash(x)=162288191)
|
| 263 |
+
8700 val loss 8.3254
|
| 264 |
+
8700 val perplexity 4127.2988
|
| 265 |
+
8700 train 8.242277 (lr=6.0049e-06) (hash(x)=152860941)
|
| 266 |
+
8749 val loss 8.3260
|
| 267 |
+
8749 val perplexity 4130.0156
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/model_08749.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f68d7f3c79549fcf734bb47e387a8ee0cc3e5a16cbb39fd4ea83f994e1862720
|
| 3 |
+
size 95220594
|
lr6e-5_total_batch_size61440_one_mask_per_head_2_latent_vectors_seed1341/optimizer_08749.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bd1fda4a5fb4baf239617453d02d4937c3a855fb876ebb66b552a5e4fd2d8039
|
| 3 |
+
size 184163894
|