Upload folder using huggingface_hub
Browse files- 12_head_baseline_lr_8e-4_head_dim_22_fixed/args.json +1 -0
- 12_head_baseline_lr_8e-4_head_dim_22_fixed/dataloader_04374.pt +3 -0
- 12_head_baseline_lr_8e-4_head_dim_22_fixed/log2.txt +529 -0
- 12_head_baseline_lr_8e-4_head_dim_22_fixed/model_04374.pt +3 -0
- 12_head_baseline_lr_8e-4_head_dim_22_fixed/optimizer_04374.pt +3 -0
12_head_baseline_lr_8e-4_head_dim_22_fixed/args.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"hellaswag": true, "attention_kind": "selective", "log_dir": "logs/repro_selective_pattern_rankings/12_head_baseline_lr_8e-4_head_dim_22_fixed", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 12, "n_embd": 264, "head_dim": 22, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 4375, "warmup_steps": 250, "group": "repro_selective_pattern_rankings", "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": 64, "total_batch_size": 131072, "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": 0.0008, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "12_head_baseline_lr_8e-4_fixed"}
|
12_head_baseline_lr_8e-4_head_dim_22_fixed/dataloader_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ec6682852ed60a9d42cb4047300f65a04f87328c0ad7a4516be84d11b28f216
|
| 3 |
+
size 964
|
12_head_baseline_lr_8e-4_head_dim_22_fixed/log2.txt
ADDED
|
@@ -0,0 +1,529 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
max_steps: 4375
|
| 2 |
+
0 val loss 11.4043
|
| 3 |
+
0 val perplexity 89707.0391
|
| 4 |
+
0 train 11.388923 (lr=1.1189e-06) (hash(x)=93356070)
|
| 5 |
+
10 train 10.462421 (lr=1.2308e-05) (hash(x)=91253010)
|
| 6 |
+
20 train 9.955700 (lr=2.3497e-05) (hash(x)=74934453)
|
| 7 |
+
30 train 9.740845 (lr=3.4685e-05) (hash(x)=79278034)
|
| 8 |
+
40 train 9.385580 (lr=4.5874e-05) (hash(x)=80107892)
|
| 9 |
+
50 train 8.995117 (lr=5.7063e-05) (hash(x)=70745428)
|
| 10 |
+
60 train 8.666317 (lr=6.8252e-05) (hash(x)=80077589)
|
| 11 |
+
70 train 8.131595 (lr=7.9441e-05) (hash(x)=76213766)
|
| 12 |
+
80 train 7.883564 (lr=9.0629e-05) (hash(x)=83218328)
|
| 13 |
+
90 train 7.670492 (lr=1.0182e-04) (hash(x)=74855845)
|
| 14 |
+
100 val loss 7.5677
|
| 15 |
+
100 val perplexity 1934.7378
|
| 16 |
+
100 train 7.439291 (lr=1.1301e-04) (hash(x)=82814902)
|
| 17 |
+
110 train 7.377951 (lr=1.2420e-04) (hash(x)=84907741)
|
| 18 |
+
120 train 7.278985 (lr=1.3538e-04) (hash(x)=82613223)
|
| 19 |
+
130 train 7.124608 (lr=1.4657e-04) (hash(x)=83540876)
|
| 20 |
+
140 train 7.105237 (lr=1.5776e-04) (hash(x)=75095216)
|
| 21 |
+
150 train 7.163971 (lr=1.6895e-04) (hash(x)=97190944)
|
| 22 |
+
160 train 7.069863 (lr=1.8014e-04) (hash(x)=82117809)
|
| 23 |
+
170 train 6.954305 (lr=1.9133e-04) (hash(x)=70514724)
|
| 24 |
+
180 train 6.854761 (lr=2.0252e-04) (hash(x)=81029624)
|
| 25 |
+
190 train 6.748855 (lr=2.1371e-04) (hash(x)=78996841)
|
| 26 |
+
200 val loss 6.6903
|
| 27 |
+
200 val perplexity 804.6004
|
| 28 |
+
200 train 6.695497 (lr=2.2490e-04) (hash(x)=79845097)
|
| 29 |
+
210 train 6.598349 (lr=2.3608e-04) (hash(x)=78997683)
|
| 30 |
+
220 train 6.464209 (lr=2.4727e-04) (hash(x)=74895865)
|
| 31 |
+
230 train 6.621018 (lr=2.5846e-04) (hash(x)=80933276)
|
| 32 |
+
240 train 6.331894 (lr=2.6965e-04) (hash(x)=77664606)
|
| 33 |
+
250 train 6.348096 (lr=2.8084e-04) (hash(x)=75903930)
|
| 34 |
+
260 train 6.254451 (lr=2.9203e-04) (hash(x)=82996853)
|
| 35 |
+
270 train 6.183151 (lr=3.0322e-04) (hash(x)=73269568)
|
| 36 |
+
280 train 6.121641 (lr=3.1441e-04) (hash(x)=82906427)
|
| 37 |
+
290 train 6.040687 (lr=3.2559e-04) (hash(x)=71797895)
|
| 38 |
+
300 val loss 6.1844
|
| 39 |
+
300 val perplexity 485.1360
|
| 40 |
+
300 train 5.911830 (lr=3.3678e-04) (hash(x)=85232249)
|
| 41 |
+
310 train 5.850702 (lr=3.4797e-04) (hash(x)=75252489)
|
| 42 |
+
320 train 5.953071 (lr=3.5916e-04) (hash(x)=74394644)
|
| 43 |
+
330 train 6.002592 (lr=3.7035e-04) (hash(x)=82772910)
|
| 44 |
+
340 train 6.045619 (lr=3.8154e-04) (hash(x)=81627464)
|
| 45 |
+
350 train 6.028246 (lr=3.9273e-04) (hash(x)=84678053)
|
| 46 |
+
360 train 5.967566 (lr=4.0392e-04) (hash(x)=81884128)
|
| 47 |
+
370 train 5.959550 (lr=4.1510e-04) (hash(x)=76379242)
|
| 48 |
+
380 train 5.737064 (lr=4.2629e-04) (hash(x)=98182875)
|
| 49 |
+
390 train 5.732515 (lr=4.3748e-04) (hash(x)=79710436)
|
| 50 |
+
400 val loss 5.7809
|
| 51 |
+
400 val perplexity 324.0513
|
| 52 |
+
400 train 5.733783 (lr=4.4867e-04) (hash(x)=79841071)
|
| 53 |
+
410 train 5.654058 (lr=4.5986e-04) (hash(x)=75844151)
|
| 54 |
+
420 train 5.652066 (lr=4.7105e-04) (hash(x)=73125036)
|
| 55 |
+
430 train 5.688927 (lr=4.8224e-04) (hash(x)=84214858)
|
| 56 |
+
440 train 5.576507 (lr=4.9343e-04) (hash(x)=80456994)
|
| 57 |
+
450 train 5.543094 (lr=5.0462e-04) (hash(x)=76735962)
|
| 58 |
+
460 train 5.481975 (lr=5.1580e-04) (hash(x)=81845446)
|
| 59 |
+
470 train 5.316375 (lr=5.2699e-04) (hash(x)=76094689)
|
| 60 |
+
480 train 5.416342 (lr=5.3818e-04) (hash(x)=83806686)
|
| 61 |
+
490 train 5.242144 (lr=5.4937e-04) (hash(x)=84690227)
|
| 62 |
+
500 val loss 5.5484
|
| 63 |
+
500 val perplexity 256.8339
|
| 64 |
+
500 train 5.310766 (lr=5.6056e-04) (hash(x)=71851938)
|
| 65 |
+
510 train 5.387978 (lr=5.7175e-04) (hash(x)=77159346)
|
| 66 |
+
520 train 5.454401 (lr=5.8294e-04) (hash(x)=80755753)
|
| 67 |
+
530 train 5.543831 (lr=5.9413e-04) (hash(x)=75379679)
|
| 68 |
+
540 train 5.463854 (lr=6.0531e-04) (hash(x)=82458619)
|
| 69 |
+
550 train 5.437819 (lr=6.1650e-04) (hash(x)=86366396)
|
| 70 |
+
560 train 5.414557 (lr=6.2769e-04) (hash(x)=85095044)
|
| 71 |
+
570 train 5.398359 (lr=6.3888e-04) (hash(x)=78385159)
|
| 72 |
+
580 train 5.349250 (lr=6.5007e-04) (hash(x)=79342394)
|
| 73 |
+
590 train 5.349400 (lr=6.6126e-04) (hash(x)=70782192)
|
| 74 |
+
600 val loss 5.3490
|
| 75 |
+
600 val perplexity 210.3880
|
| 76 |
+
600 train 5.324270 (lr=6.7245e-04) (hash(x)=76130353)
|
| 77 |
+
610 train 5.223341 (lr=6.8364e-04) (hash(x)=74778440)
|
| 78 |
+
620 train 5.211050 (lr=6.9483e-04) (hash(x)=79129709)
|
| 79 |
+
630 train 5.114326 (lr=7.0601e-04) (hash(x)=76469962)
|
| 80 |
+
640 train 5.111544 (lr=7.1720e-04) (hash(x)=78288049)
|
| 81 |
+
650 train 5.070465 (lr=7.2839e-04) (hash(x)=76641388)
|
| 82 |
+
660 train 4.989712 (lr=7.3958e-04) (hash(x)=75937906)
|
| 83 |
+
670 train 4.988499 (lr=7.5077e-04) (hash(x)=74807157)
|
| 84 |
+
680 train 4.895653 (lr=7.6196e-04) (hash(x)=77490144)
|
| 85 |
+
690 train 4.934090 (lr=7.7315e-04) (hash(x)=72900124)
|
| 86 |
+
700 val loss 5.1880
|
| 87 |
+
700 val perplexity 179.1159
|
| 88 |
+
700 train 4.891424 (lr=7.8434e-04) (hash(x)=73424218)
|
| 89 |
+
710 train 5.172663 (lr=7.9552e-04) (hash(x)=86052345)
|
| 90 |
+
720 train 5.129795 (lr=8.0000e-04) (hash(x)=77462613)
|
| 91 |
+
730 train 5.138889 (lr=7.9997e-04) (hash(x)=79027471)
|
| 92 |
+
740 train 5.111305 (lr=7.9992e-04) (hash(x)=78149992)
|
| 93 |
+
750 train 5.089380 (lr=7.9984e-04) (hash(x)=89147499)
|
| 94 |
+
760 train 5.078344 (lr=7.9973e-04) (hash(x)=74931538)
|
| 95 |
+
770 train 5.028081 (lr=7.9960e-04) (hash(x)=89648204)
|
| 96 |
+
780 train 5.072010 (lr=7.9944e-04) (hash(x)=83708293)
|
| 97 |
+
790 train 5.080956 (lr=7.9925e-04) (hash(x)=84336512)
|
| 98 |
+
800 val loss 5.0401
|
| 99 |
+
800 val perplexity 154.4910
|
| 100 |
+
800 train 4.961507 (lr=7.9904e-04) (hash(x)=75025285)
|
| 101 |
+
810 train 4.890038 (lr=7.9880e-04) (hash(x)=79436536)
|
| 102 |
+
820 train 4.874693 (lr=7.9854e-04) (hash(x)=90467390)
|
| 103 |
+
830 train 4.930626 (lr=7.9825e-04) (hash(x)=77292020)
|
| 104 |
+
840 train 4.839594 (lr=7.9793e-04) (hash(x)=75568927)
|
| 105 |
+
850 train 4.856942 (lr=7.9759e-04) (hash(x)=79671781)
|
| 106 |
+
860 train 4.948665 (lr=7.9722e-04) (hash(x)=84280943)
|
| 107 |
+
870 train 4.911723 (lr=7.9682e-04) (hash(x)=80901294)
|
| 108 |
+
880 train 5.053764 (lr=7.9640e-04) (hash(x)=81437584)
|
| 109 |
+
890 train 4.903955 (lr=7.9595e-04) (hash(x)=73893778)
|
| 110 |
+
900 val loss 4.9112
|
| 111 |
+
900 val perplexity 135.8005
|
| 112 |
+
900 train 4.828310 (lr=7.9547e-04) (hash(x)=74987794)
|
| 113 |
+
910 train 4.853667 (lr=7.9497e-04) (hash(x)=93721374)
|
| 114 |
+
920 train 4.839195 (lr=7.9444e-04) (hash(x)=79149678)
|
| 115 |
+
930 train 4.785468 (lr=7.9389e-04) (hash(x)=83179387)
|
| 116 |
+
940 train 4.787247 (lr=7.9331e-04) (hash(x)=89009978)
|
| 117 |
+
950 train 4.741975 (lr=7.9270e-04) (hash(x)=76483735)
|
| 118 |
+
960 train 4.679790 (lr=7.9207e-04) (hash(x)=78638850)
|
| 119 |
+
970 train 4.705347 (lr=7.9141e-04) (hash(x)=90415206)
|
| 120 |
+
980 train 4.637120 (lr=7.9073e-04) (hash(x)=90819195)
|
| 121 |
+
990 train 4.774509 (lr=7.9002e-04) (hash(x)=85938852)
|
| 122 |
+
1000 val loss 4.7558
|
| 123 |
+
1000 val perplexity 116.2588
|
| 124 |
+
1000 train 4.811062 (lr=7.8928e-04) (hash(x)=82436789)
|
| 125 |
+
1010 train 4.803025 (lr=7.8852e-04) (hash(x)=63022149)
|
| 126 |
+
1020 train 4.752833 (lr=7.8773e-04) (hash(x)=83501199)
|
| 127 |
+
1030 train 4.726783 (lr=7.8692e-04) (hash(x)=77093326)
|
| 128 |
+
1040 train 4.589956 (lr=7.8608e-04) (hash(x)=73437559)
|
| 129 |
+
1050 train 4.687597 (lr=7.8522e-04) (hash(x)=84550388)
|
| 130 |
+
1060 train 4.686138 (lr=7.8433e-04) (hash(x)=86184566)
|
| 131 |
+
1070 train 4.673692 (lr=7.8342e-04) (hash(x)=85644922)
|
| 132 |
+
1080 train 4.606560 (lr=7.8248e-04) (hash(x)=86179801)
|
| 133 |
+
1090 train 4.577961 (lr=7.8151e-04) (hash(x)=82868303)
|
| 134 |
+
1100 val loss 4.6420
|
| 135 |
+
1100 val perplexity 103.7554
|
| 136 |
+
1100 train 4.677928 (lr=7.8052e-04) (hash(x)=96780388)
|
| 137 |
+
1110 train 4.608566 (lr=7.7951e-04) (hash(x)=87223122)
|
| 138 |
+
1120 train 4.575022 (lr=7.7847e-04) (hash(x)=77292786)
|
| 139 |
+
1130 train 4.594349 (lr=7.7740e-04) (hash(x)=88761375)
|
| 140 |
+
1140 train 4.508932 (lr=7.7631e-04) (hash(x)=80956468)
|
| 141 |
+
1150 train 4.687494 (lr=7.7520e-04) (hash(x)=84725462)
|
| 142 |
+
1160 train 4.656981 (lr=7.7406e-04) (hash(x)=76098113)
|
| 143 |
+
1170 train 4.631224 (lr=7.7289e-04) (hash(x)=84387685)
|
| 144 |
+
1180 train 4.569960 (lr=7.7170e-04) (hash(x)=76120321)
|
| 145 |
+
1190 train 4.530944 (lr=7.7049e-04) (hash(x)=85974065)
|
| 146 |
+
1200 val loss 4.5507
|
| 147 |
+
1200 val perplexity 94.6945
|
| 148 |
+
1200 train 4.429749 (lr=7.6925e-04) (hash(x)=79340644)
|
| 149 |
+
1210 train 4.521499 (lr=7.6799e-04) (hash(x)=88345025)
|
| 150 |
+
1220 train 4.438153 (lr=7.6670e-04) (hash(x)=86406230)
|
| 151 |
+
1230 train 4.421066 (lr=7.6539e-04) (hash(x)=80608975)
|
| 152 |
+
1240 train 4.475980 (lr=7.6406e-04) (hash(x)=81635225)
|
| 153 |
+
1250 train 4.315457 (lr=7.6270e-04) (hash(x)=82126376)
|
| 154 |
+
1260 train 4.481150 (lr=7.6132e-04) (hash(x)=83737972)
|
| 155 |
+
1270 train 4.388452 (lr=7.5992e-04) (hash(x)=81132345)
|
| 156 |
+
1280 train 4.482712 (lr=7.5849e-04) (hash(x)=85850404)
|
| 157 |
+
1290 train 4.494557 (lr=7.5704e-04) (hash(x)=83490640)
|
| 158 |
+
1300 val loss 4.4970
|
| 159 |
+
1300 val perplexity 89.7513
|
| 160 |
+
1300 train 4.463191 (lr=7.5556e-04) (hash(x)=77545187)
|
| 161 |
+
1310 train 4.511926 (lr=7.5406e-04) (hash(x)=86412685)
|
| 162 |
+
1320 train 4.402415 (lr=7.5254e-04) (hash(x)=86429640)
|
| 163 |
+
1330 train 4.545220 (lr=7.5099e-04) (hash(x)=79645281)
|
| 164 |
+
1340 train 4.501055 (lr=7.4943e-04) (hash(x)=82668541)
|
| 165 |
+
1350 train 4.353905 (lr=7.4784e-04) (hash(x)=73627248)
|
| 166 |
+
1360 train 4.444376 (lr=7.4622e-04) (hash(x)=74845133)
|
| 167 |
+
1370 train 4.358597 (lr=7.4459e-04) (hash(x)=77850497)
|
| 168 |
+
1380 train 4.382380 (lr=7.4293e-04) (hash(x)=76143954)
|
| 169 |
+
1390 train 4.391551 (lr=7.4125e-04) (hash(x)=80675544)
|
| 170 |
+
1400 val loss 4.4580
|
| 171 |
+
1400 val perplexity 86.3124
|
| 172 |
+
1400 train 4.365345 (lr=7.3954e-04) (hash(x)=76010938)
|
| 173 |
+
1410 train 4.420878 (lr=7.3782e-04) (hash(x)=81260300)
|
| 174 |
+
1420 train 4.521589 (lr=7.3607e-04) (hash(x)=83803629)
|
| 175 |
+
1430 train 4.358571 (lr=7.3430e-04) (hash(x)=83840110)
|
| 176 |
+
1440 train 4.554413 (lr=7.3251e-04) (hash(x)=97711831)
|
| 177 |
+
1450 train 4.424284 (lr=7.3070e-04) (hash(x)=81110139)
|
| 178 |
+
1460 train 4.337059 (lr=7.2887e-04) (hash(x)=90942839)
|
| 179 |
+
1470 train 4.376791 (lr=7.2701e-04) (hash(x)=80721564)
|
| 180 |
+
1480 train 4.474306 (lr=7.2514e-04) (hash(x)=77852059)
|
| 181 |
+
1490 train 4.370626 (lr=7.2324e-04) (hash(x)=80011365)
|
| 182 |
+
1500 val loss 4.4096
|
| 183 |
+
1500 val perplexity 82.2353
|
| 184 |
+
1500 train 4.392446 (lr=7.2132e-04) (hash(x)=84496142)
|
| 185 |
+
1510 train 4.334968 (lr=7.1938e-04) (hash(x)=70186729)
|
| 186 |
+
1520 train 4.360988 (lr=7.1742e-04) (hash(x)=74854227)
|
| 187 |
+
1530 train 4.262049 (lr=7.1544e-04) (hash(x)=77468161)
|
| 188 |
+
1540 train 4.280844 (lr=7.1344e-04) (hash(x)=87853059)
|
| 189 |
+
1550 train 4.203699 (lr=7.1142e-04) (hash(x)=71225436)
|
| 190 |
+
1560 train 4.309296 (lr=7.0939e-04) (hash(x)=84072783)
|
| 191 |
+
1570 train 4.365957 (lr=7.0733e-04) (hash(x)=72723098)
|
| 192 |
+
1580 train 4.384117 (lr=7.0525e-04) (hash(x)=90409866)
|
| 193 |
+
1590 train 4.446133 (lr=7.0315e-04) (hash(x)=87481378)
|
| 194 |
+
1600 val loss 4.3710
|
| 195 |
+
1600 val perplexity 79.1224
|
| 196 |
+
1600 train 4.429656 (lr=7.0103e-04) (hash(x)=77643862)
|
| 197 |
+
1610 train 4.321773 (lr=6.9889e-04) (hash(x)=88862575)
|
| 198 |
+
1620 train 4.329190 (lr=6.9673e-04) (hash(x)=84612581)
|
| 199 |
+
1630 train 4.367182 (lr=6.9456e-04) (hash(x)=87075989)
|
| 200 |
+
1640 train 4.252344 (lr=6.9236e-04) (hash(x)=88277361)
|
| 201 |
+
1650 train 4.364820 (lr=6.9015e-04) (hash(x)=78750236)
|
| 202 |
+
1660 train 4.291776 (lr=6.8792e-04) (hash(x)=82604581)
|
| 203 |
+
1670 train 4.119716 (lr=6.8567e-04) (hash(x)=68482265)
|
| 204 |
+
1680 train 4.250303 (lr=6.8340e-04) (hash(x)=75088835)
|
| 205 |
+
1690 train 4.195225 (lr=6.8112e-04) (hash(x)=66060989)
|
| 206 |
+
1700 val loss 4.3646
|
| 207 |
+
1700 val perplexity 78.6181
|
| 208 |
+
1700 train 4.359484 (lr=6.7881e-04) (hash(x)=79986754)
|
| 209 |
+
1710 train 4.350958 (lr=6.7649e-04) (hash(x)=83657930)
|
| 210 |
+
1720 train 4.292342 (lr=6.7415e-04) (hash(x)=81754135)
|
| 211 |
+
1730 train 4.368599 (lr=6.7180e-04) (hash(x)=78904427)
|
| 212 |
+
1740 train 4.277336 (lr=6.6943e-04) (hash(x)=85920177)
|
| 213 |
+
1750 train 4.320709 (lr=6.6704e-04) (hash(x)=86573211)
|
| 214 |
+
1760 train 4.279339 (lr=6.6463e-04) (hash(x)=81737128)
|
| 215 |
+
1770 train 4.268407 (lr=6.6221e-04) (hash(x)=77400968)
|
| 216 |
+
1780 train 4.284335 (lr=6.5977e-04) (hash(x)=73545497)
|
| 217 |
+
1790 train 4.525265 (lr=6.5731e-04) (hash(x)=71641943)
|
| 218 |
+
1800 val loss 4.3151
|
| 219 |
+
1800 val perplexity 74.8181
|
| 220 |
+
1800 train 4.376225 (lr=6.5484e-04) (hash(x)=87819781)
|
| 221 |
+
1810 train 4.233324 (lr=6.5235e-04) (hash(x)=86870770)
|
| 222 |
+
1820 train 4.244097 (lr=6.4985e-04) (hash(x)=82522211)
|
| 223 |
+
1830 train 4.325108 (lr=6.4733e-04) (hash(x)=61947437)
|
| 224 |
+
1840 train 4.209496 (lr=6.4480e-04) (hash(x)=79865406)
|
| 225 |
+
1850 train 4.174800 (lr=6.4225e-04) (hash(x)=79828721)
|
| 226 |
+
1860 train 4.318232 (lr=6.3968e-04) (hash(x)=80869571)
|
| 227 |
+
1870 train 4.301679 (lr=6.3711e-04) (hash(x)=73780971)
|
| 228 |
+
1880 train 4.265572 (lr=6.3451e-04) (hash(x)=79249549)
|
| 229 |
+
1890 train 4.319492 (lr=6.3191e-04) (hash(x)=81041904)
|
| 230 |
+
1900 val loss 4.2906
|
| 231 |
+
1900 val perplexity 73.0124
|
| 232 |
+
1900 train 4.389087 (lr=6.2928e-04) (hash(x)=82456430)
|
| 233 |
+
1910 train 4.153023 (lr=6.2665e-04) (hash(x)=82222135)
|
| 234 |
+
1920 train 4.193025 (lr=6.2400e-04) (hash(x)=70033249)
|
| 235 |
+
1930 train 4.173045 (lr=6.2134e-04) (hash(x)=72887360)
|
| 236 |
+
1940 train 4.225553 (lr=6.1866e-04) (hash(x)=83251100)
|
| 237 |
+
1950 train 4.266335 (lr=6.1597e-04) (hash(x)=79660266)
|
| 238 |
+
1960 train 4.140210 (lr=6.1327e-04) (hash(x)=78919068)
|
| 239 |
+
1970 train 4.237479 (lr=6.1055e-04) (hash(x)=94690431)
|
| 240 |
+
1980 train 4.085918 (lr=6.0783e-04) (hash(x)=88153756)
|
| 241 |
+
1990 train 4.129201 (lr=6.0509e-04) (hash(x)=77195688)
|
| 242 |
+
2000 val loss 4.2654
|
| 243 |
+
2000 val perplexity 71.1960
|
| 244 |
+
2000 train 4.233664 (lr=6.0233e-04) (hash(x)=81308591)
|
| 245 |
+
2010 train 4.278122 (lr=5.9957e-04) (hash(x)=80205479)
|
| 246 |
+
2020 train 4.193841 (lr=5.9679e-04) (hash(x)=81008704)
|
| 247 |
+
2030 train 4.258300 (lr=5.9401e-04) (hash(x)=82914358)
|
| 248 |
+
2040 train 4.270494 (lr=5.9121e-04) (hash(x)=88064399)
|
| 249 |
+
2050 train 4.273919 (lr=5.8840e-04) (hash(x)=82889390)
|
| 250 |
+
2060 train 4.344082 (lr=5.8558e-04) (hash(x)=80913554)
|
| 251 |
+
2070 train 4.184629 (lr=5.8275e-04) (hash(x)=78121791)
|
| 252 |
+
2080 train 4.319590 (lr=5.7990e-04) (hash(x)=75356657)
|
| 253 |
+
2090 train 4.145003 (lr=5.7705e-04) (hash(x)=74048412)
|
| 254 |
+
2100 val loss 4.2389
|
| 255 |
+
2100 val perplexity 69.3323
|
| 256 |
+
2100 train 4.062921 (lr=5.7419e-04) (hash(x)=68928225)
|
| 257 |
+
2110 train 4.116064 (lr=5.7132e-04) (hash(x)=93734745)
|
| 258 |
+
2120 train 4.113101 (lr=5.6844e-04) (hash(x)=72878164)
|
| 259 |
+
2130 train 4.240356 (lr=5.6554e-04) (hash(x)=71508226)
|
| 260 |
+
2140 train 4.196486 (lr=5.6264e-04) (hash(x)=84168671)
|
| 261 |
+
2150 train 4.136520 (lr=5.5973e-04) (hash(x)=74673639)
|
| 262 |
+
2160 train 4.284075 (lr=5.5681e-04) (hash(x)=75470031)
|
| 263 |
+
2170 train 4.245594 (lr=5.5389e-04) (hash(x)=74307890)
|
| 264 |
+
2180 train 4.359695 (lr=5.5095e-04) (hash(x)=77214245)
|
| 265 |
+
2190 train 4.139488 (lr=5.4801e-04) (hash(x)=88628359)
|
| 266 |
+
2200 val loss 4.2191
|
| 267 |
+
2200 val perplexity 67.9702
|
| 268 |
+
2200 train 4.151878 (lr=5.4506e-04) (hash(x)=74779126)
|
| 269 |
+
2210 train 4.192382 (lr=5.4210e-04) (hash(x)=79817976)
|
| 270 |
+
2220 train 4.142745 (lr=5.3913e-04) (hash(x)=75258996)
|
| 271 |
+
2230 train 4.120797 (lr=5.3616e-04) (hash(x)=88164047)
|
| 272 |
+
2240 train 4.098520 (lr=5.3317e-04) (hash(x)=78580686)
|
| 273 |
+
2250 train 4.041303 (lr=5.3019e-04) (hash(x)=76510617)
|
| 274 |
+
2260 train 4.118978 (lr=5.2719e-04) (hash(x)=76043966)
|
| 275 |
+
2270 train 4.094014 (lr=5.2419e-04) (hash(x)=73062098)
|
| 276 |
+
2280 train 4.321898 (lr=5.2118e-04) (hash(x)=81885909)
|
| 277 |
+
2290 train 4.255442 (lr=5.1817e-04) (hash(x)=81481741)
|
| 278 |
+
2300 val loss 4.1963
|
| 279 |
+
2300 val perplexity 66.4379
|
| 280 |
+
2300 train 4.117752 (lr=5.1515e-04) (hash(x)=82104275)
|
| 281 |
+
2310 train 4.166573 (lr=5.1213e-04) (hash(x)=82137309)
|
| 282 |
+
2320 train 4.142792 (lr=5.0910e-04) (hash(x)=86741167)
|
| 283 |
+
2330 train 4.157535 (lr=5.0606e-04) (hash(x)=84796263)
|
| 284 |
+
2340 train 4.136901 (lr=5.0302e-04) (hash(x)=169966529)
|
| 285 |
+
2350 train 3.997656 (lr=4.9998e-04) (hash(x)=80006061)
|
| 286 |
+
2360 train 4.062540 (lr=4.9693e-04) (hash(x)=71090999)
|
| 287 |
+
2370 train 4.190221 (lr=4.9387e-04) (hash(x)=86641850)
|
| 288 |
+
2380 train 4.092087 (lr=4.9082e-04) (hash(x)=84573253)
|
| 289 |
+
2390 train 4.126458 (lr=4.8776e-04) (hash(x)=74483764)
|
| 290 |
+
2400 val loss 4.1890
|
| 291 |
+
2400 val perplexity 65.9553
|
| 292 |
+
2400 train 4.099054 (lr=4.8469e-04) (hash(x)=78327659)
|
| 293 |
+
2410 train 4.161429 (lr=4.8162e-04) (hash(x)=77273627)
|
| 294 |
+
2420 train 4.022769 (lr=4.7855e-04) (hash(x)=76938049)
|
| 295 |
+
2430 train 4.036468 (lr=4.7548e-04) (hash(x)=98449442)
|
| 296 |
+
2440 train 3.888872 (lr=4.7240e-04) (hash(x)=78157797)
|
| 297 |
+
2450 train 3.931483 (lr=4.6932e-04) (hash(x)=80637582)
|
| 298 |
+
2460 train 3.918687 (lr=4.6624e-04) (hash(x)=80023854)
|
| 299 |
+
2470 train 3.983843 (lr=4.6316e-04) (hash(x)=86678884)
|
| 300 |
+
2480 train 4.202255 (lr=4.6008e-04) (hash(x)=84086469)
|
| 301 |
+
2490 train 4.227712 (lr=4.5699e-04) (hash(x)=83131141)
|
| 302 |
+
2500 val loss 4.1625
|
| 303 |
+
2500 val perplexity 64.2324
|
| 304 |
+
2500 train 4.140171 (lr=4.5390e-04) (hash(x)=82583497)
|
| 305 |
+
2510 train 4.206031 (lr=4.5081e-04) (hash(x)=79164326)
|
| 306 |
+
2520 train 4.106184 (lr=4.4772e-04) (hash(x)=76456503)
|
| 307 |
+
2530 train 4.167851 (lr=4.4464e-04) (hash(x)=82001228)
|
| 308 |
+
2540 train 4.205464 (lr=4.4155e-04) (hash(x)=75615595)
|
| 309 |
+
2550 train 4.150397 (lr=4.3845e-04) (hash(x)=78397869)
|
| 310 |
+
2560 train 4.151053 (lr=4.3536e-04) (hash(x)=75547032)
|
| 311 |
+
2570 train 4.063576 (lr=4.3228e-04) (hash(x)=89201025)
|
| 312 |
+
2580 train 4.049784 (lr=4.2919e-04) (hash(x)=89856704)
|
| 313 |
+
2590 train 4.056904 (lr=4.2610e-04) (hash(x)=82175682)
|
| 314 |
+
2600 val loss 4.1449
|
| 315 |
+
2600 val perplexity 63.1142
|
| 316 |
+
2600 train 4.062811 (lr=4.2301e-04) (hash(x)=78312826)
|
| 317 |
+
2610 train 4.054043 (lr=4.1992e-04) (hash(x)=77066588)
|
| 318 |
+
2620 train 3.898928 (lr=4.1684e-04) (hash(x)=78666061)
|
| 319 |
+
2630 train 3.893532 (lr=4.1376e-04) (hash(x)=93762143)
|
| 320 |
+
2640 train 3.911001 (lr=4.1068e-04) (hash(x)=83191587)
|
| 321 |
+
2650 train 3.840140 (lr=4.0760e-04) (hash(x)=87169585)
|
| 322 |
+
2660 train 4.049696 (lr=4.0452e-04) (hash(x)=86426388)
|
| 323 |
+
2670 train 4.191960 (lr=4.0145e-04) (hash(x)=76692638)
|
| 324 |
+
2680 train 4.100883 (lr=3.9838e-04) (hash(x)=77446063)
|
| 325 |
+
2690 train 4.105713 (lr=3.9531e-04) (hash(x)=79809050)
|
| 326 |
+
2700 val loss 4.1373
|
| 327 |
+
2700 val perplexity 62.6346
|
| 328 |
+
2700 train 4.034157 (lr=3.9224e-04) (hash(x)=83116823)
|
| 329 |
+
2710 train 4.146717 (lr=3.8918e-04) (hash(x)=75622148)
|
| 330 |
+
2720 train 4.071832 (lr=3.8613e-04) (hash(x)=80690512)
|
| 331 |
+
2730 train 4.141748 (lr=3.8307e-04) (hash(x)=78009984)
|
| 332 |
+
2740 train 4.097802 (lr=3.8002e-04) (hash(x)=74205488)
|
| 333 |
+
2750 train 4.099316 (lr=3.7698e-04) (hash(x)=91013332)
|
| 334 |
+
2760 train 4.024355 (lr=3.7394e-04) (hash(x)=76954961)
|
| 335 |
+
2770 train 4.034565 (lr=3.7090e-04) (hash(x)=79390317)
|
| 336 |
+
2780 train 3.900937 (lr=3.6787e-04) (hash(x)=70168783)
|
| 337 |
+
2790 train 4.056646 (lr=3.6485e-04) (hash(x)=76028417)
|
| 338 |
+
2800 val loss 4.1185
|
| 339 |
+
2800 val perplexity 61.4661
|
| 340 |
+
2800 train 3.889193 (lr=3.6183e-04) (hash(x)=77656050)
|
| 341 |
+
2810 train 3.936734 (lr=3.5882e-04) (hash(x)=94248216)
|
| 342 |
+
2820 train 4.001037 (lr=3.5581e-04) (hash(x)=78305078)
|
| 343 |
+
2830 train 3.912551 (lr=3.5281e-04) (hash(x)=79948848)
|
| 344 |
+
2840 train 3.784479 (lr=3.4981e-04) (hash(x)=85341024)
|
| 345 |
+
2850 train 4.086265 (lr=3.4683e-04) (hash(x)=78735170)
|
| 346 |
+
2860 train 4.337343 (lr=3.4384e-04) (hash(x)=71616419)
|
| 347 |
+
2870 train 4.109067 (lr=3.4087e-04) (hash(x)=78656517)
|
| 348 |
+
2880 train 4.122245 (lr=3.3790e-04) (hash(x)=80073987)
|
| 349 |
+
2890 train 4.055164 (lr=3.3494e-04) (hash(x)=76894809)
|
| 350 |
+
2900 val loss 4.0947
|
| 351 |
+
2900 val perplexity 60.0190
|
| 352 |
+
2900 train 4.031393 (lr=3.3199e-04) (hash(x)=80499838)
|
| 353 |
+
2910 train 4.079880 (lr=3.2905e-04) (hash(x)=72673354)
|
| 354 |
+
2920 train 4.062461 (lr=3.2611e-04) (hash(x)=84265768)
|
| 355 |
+
2930 train 4.095368 (lr=3.2319e-04) (hash(x)=79612060)
|
| 356 |
+
2940 train 3.943176 (lr=3.2027e-04) (hash(x)=74970087)
|
| 357 |
+
2950 train 4.101684 (lr=3.1736e-04) (hash(x)=84166818)
|
| 358 |
+
2960 train 4.089890 (lr=3.1446e-04) (hash(x)=89410221)
|
| 359 |
+
2970 train 4.031025 (lr=3.1156e-04) (hash(x)=75672566)
|
| 360 |
+
2980 train 3.884170 (lr=3.0868e-04) (hash(x)=81760314)
|
| 361 |
+
2990 train 3.860748 (lr=3.0581e-04) (hash(x)=80605200)
|
| 362 |
+
3000 val loss 4.0945
|
| 363 |
+
3000 val perplexity 60.0080
|
| 364 |
+
3000 train 3.860792 (lr=3.0295e-04) (hash(x)=83804735)
|
| 365 |
+
3010 train 3.787617 (lr=3.0010e-04) (hash(x)=77015303)
|
| 366 |
+
3020 train 3.858312 (lr=2.9725e-04) (hash(x)=81464523)
|
| 367 |
+
3030 train 4.048865 (lr=2.9442e-04) (hash(x)=80168230)
|
| 368 |
+
3040 train 3.996626 (lr=2.9160e-04) (hash(x)=84817006)
|
| 369 |
+
3050 train 4.100596 (lr=2.8879e-04) (hash(x)=71601811)
|
| 370 |
+
3060 train 4.015127 (lr=2.8599e-04) (hash(x)=85499733)
|
| 371 |
+
3070 train 4.075488 (lr=2.8321e-04) (hash(x)=78670408)
|
| 372 |
+
3080 train 3.963888 (lr=2.8043e-04) (hash(x)=77120468)
|
| 373 |
+
3090 train 4.078160 (lr=2.7767e-04) (hash(x)=77927426)
|
| 374 |
+
3100 val loss 4.0761
|
| 375 |
+
3100 val perplexity 58.9171
|
| 376 |
+
3100 train 4.089870 (lr=2.7491e-04) (hash(x)=83998606)
|
| 377 |
+
3110 train 3.985148 (lr=2.7217e-04) (hash(x)=77990218)
|
| 378 |
+
3120 train 3.901287 (lr=2.6945e-04) (hash(x)=81623970)
|
| 379 |
+
3130 train 3.875212 (lr=2.6673e-04) (hash(x)=75014781)
|
| 380 |
+
3140 train 3.929268 (lr=2.6403e-04) (hash(x)=72591250)
|
| 381 |
+
3150 train 3.946052 (lr=2.6134e-04) (hash(x)=81421847)
|
| 382 |
+
3160 train 3.992490 (lr=2.5866e-04) (hash(x)=75112631)
|
| 383 |
+
3170 train 4.094281 (lr=2.5600e-04) (hash(x)=87518033)
|
| 384 |
+
3180 train 3.943526 (lr=2.5335e-04) (hash(x)=78648348)
|
| 385 |
+
3190 train 4.060004 (lr=2.5072e-04) (hash(x)=78270029)
|
| 386 |
+
3200 val loss 4.0585
|
| 387 |
+
3200 val perplexity 57.8863
|
| 388 |
+
3200 train 4.136689 (lr=2.4809e-04) (hash(x)=83380714)
|
| 389 |
+
3210 train 4.304782 (lr=2.4549e-04) (hash(x)=67804991)
|
| 390 |
+
3220 train 4.043062 (lr=2.4289e-04) (hash(x)=76968804)
|
| 391 |
+
3230 train 3.977912 (lr=2.4032e-04) (hash(x)=75798670)
|
| 392 |
+
3240 train 4.049535 (lr=2.3775e-04) (hash(x)=83192811)
|
| 393 |
+
3250 train 3.988622 (lr=2.3520e-04) (hash(x)=81659789)
|
| 394 |
+
3260 train 4.029355 (lr=2.3267e-04) (hash(x)=80622502)
|
| 395 |
+
3270 train 4.003949 (lr=2.3015e-04) (hash(x)=85436511)
|
| 396 |
+
3280 train 3.941789 (lr=2.2765e-04) (hash(x)=80589180)
|
| 397 |
+
3290 train 4.080757 (lr=2.2516e-04) (hash(x)=85747193)
|
| 398 |
+
3300 val loss 4.0538
|
| 399 |
+
3300 val perplexity 57.6185
|
| 400 |
+
3300 train 3.948888 (lr=2.2269e-04) (hash(x)=74635692)
|
| 401 |
+
3310 train 3.939957 (lr=2.2023e-04) (hash(x)=82045455)
|
| 402 |
+
3320 train 3.806700 (lr=2.1779e-04) (hash(x)=73249173)
|
| 403 |
+
3330 train 3.771123 (lr=2.1537e-04) (hash(x)=71492338)
|
| 404 |
+
3340 train 4.098222 (lr=2.1296e-04) (hash(x)=76374071)
|
| 405 |
+
3350 train 4.070264 (lr=2.1057e-04) (hash(x)=80474064)
|
| 406 |
+
3360 train 3.984107 (lr=2.0820e-04) (hash(x)=84390892)
|
| 407 |
+
3370 train 4.032338 (lr=2.0585e-04) (hash(x)=83399949)
|
| 408 |
+
3380 train 4.079006 (lr=2.0351e-04) (hash(x)=77648059)
|
| 409 |
+
3390 train 4.049698 (lr=2.0119e-04) (hash(x)=80152701)
|
| 410 |
+
3400 val loss 4.0370
|
| 411 |
+
3400 val perplexity 56.6555
|
| 412 |
+
3400 train 4.099397 (lr=1.9888e-04) (hash(x)=81952545)
|
| 413 |
+
3410 train 4.060771 (lr=1.9660e-04) (hash(x)=80908993)
|
| 414 |
+
3420 train 4.100306 (lr=1.9433e-04) (hash(x)=79610037)
|
| 415 |
+
3430 train 3.980257 (lr=1.9208e-04) (hash(x)=87624382)
|
| 416 |
+
3440 train 4.027635 (lr=1.8985e-04) (hash(x)=82336381)
|
| 417 |
+
3450 train 3.998418 (lr=1.8764e-04) (hash(x)=87687835)
|
| 418 |
+
3460 train 4.004923 (lr=1.8544e-04) (hash(x)=76376135)
|
| 419 |
+
3470 train 3.919445 (lr=1.8327e-04) (hash(x)=79331391)
|
| 420 |
+
3480 train 3.831715 (lr=1.8111e-04) (hash(x)=107002681)
|
| 421 |
+
3490 train 3.905044 (lr=1.7897e-04) (hash(x)=77465514)
|
| 422 |
+
3500 val loss 4.0327
|
| 423 |
+
3500 val perplexity 56.4119
|
| 424 |
+
3500 train 3.871272 (lr=1.7685e-04) (hash(x)=88237229)
|
| 425 |
+
3510 train 3.935917 (lr=1.7475e-04) (hash(x)=81011739)
|
| 426 |
+
3520 train 3.979258 (lr=1.7267e-04) (hash(x)=64643427)
|
| 427 |
+
3530 train 4.255639 (lr=1.7061e-04) (hash(x)=78029539)
|
| 428 |
+
3540 train 4.095768 (lr=1.6858e-04) (hash(x)=83188968)
|
| 429 |
+
3550 train 4.113892 (lr=1.6656e-04) (hash(x)=86104185)
|
| 430 |
+
3560 train 4.040516 (lr=1.6456e-04) (hash(x)=81993629)
|
| 431 |
+
3570 train 4.010998 (lr=1.6258e-04) (hash(x)=72719368)
|
| 432 |
+
3580 train 4.022573 (lr=1.6062e-04) (hash(x)=90091487)
|
| 433 |
+
3590 train 3.910327 (lr=1.5868e-04) (hash(x)=77393152)
|
| 434 |
+
3600 val loss 4.0165
|
| 435 |
+
3600 val perplexity 55.5089
|
| 436 |
+
3600 train 4.018743 (lr=1.5676e-04) (hash(x)=73103504)
|
| 437 |
+
3610 train 4.026032 (lr=1.5486e-04) (hash(x)=84429400)
|
| 438 |
+
3620 train 3.998312 (lr=1.5299e-04) (hash(x)=77021795)
|
| 439 |
+
3630 train 3.952541 (lr=1.5113e-04) (hash(x)=92830605)
|
| 440 |
+
3640 train 4.142409 (lr=1.4930e-04) (hash(x)=78313175)
|
| 441 |
+
3650 train 3.944538 (lr=1.4749e-04) (hash(x)=85395549)
|
| 442 |
+
3660 train 3.742173 (lr=1.4570e-04) (hash(x)=78114459)
|
| 443 |
+
3670 train 3.869202 (lr=1.4393e-04) (hash(x)=74968316)
|
| 444 |
+
3680 train 3.796798 (lr=1.4218e-04) (hash(x)=73358737)
|
| 445 |
+
3690 train 3.904264 (lr=1.4046e-04) (hash(x)=76399442)
|
| 446 |
+
3700 val loss 4.0180
|
| 447 |
+
3700 val perplexity 55.5926
|
| 448 |
+
3700 train 3.904297 (lr=1.3875e-04) (hash(x)=74039273)
|
| 449 |
+
3710 train 4.075643 (lr=1.3707e-04) (hash(x)=79299680)
|
| 450 |
+
3720 train 3.994953 (lr=1.3541e-04) (hash(x)=94701498)
|
| 451 |
+
3730 train 4.075052 (lr=1.3378e-04) (hash(x)=75352071)
|
| 452 |
+
3740 train 3.953482 (lr=1.3216e-04) (hash(x)=77909487)
|
| 453 |
+
3750 train 4.054442 (lr=1.3057e-04) (hash(x)=73986730)
|
| 454 |
+
3760 train 4.009195 (lr=1.2901e-04) (hash(x)=79325763)
|
| 455 |
+
3770 train 4.056023 (lr=1.2746e-04) (hash(x)=72457818)
|
| 456 |
+
3780 train 4.198357 (lr=1.2594e-04) (hash(x)=71775590)
|
| 457 |
+
3790 train 4.088856 (lr=1.2444e-04) (hash(x)=82638943)
|
| 458 |
+
3800 val loss 4.0030
|
| 459 |
+
3800 val perplexity 54.7611
|
| 460 |
+
3800 train 4.051215 (lr=1.2296e-04) (hash(x)=79965893)
|
| 461 |
+
3810 train 3.990251 (lr=1.2151e-04) (hash(x)=72598235)
|
| 462 |
+
3820 train 3.971640 (lr=1.2008e-04) (hash(x)=83113889)
|
| 463 |
+
3830 train 4.033205 (lr=1.1868e-04) (hash(x)=74434590)
|
| 464 |
+
3840 train 3.948935 (lr=1.1730e-04) (hash(x)=82860348)
|
| 465 |
+
3850 train 3.885401 (lr=1.1594e-04) (hash(x)=78067565)
|
| 466 |
+
3860 train 3.953950 (lr=1.1461e-04) (hash(x)=82592498)
|
| 467 |
+
3870 train 3.890995 (lr=1.1330e-04) (hash(x)=81820733)
|
| 468 |
+
3880 train 3.904396 (lr=1.1201e-04) (hash(x)=87709040)
|
| 469 |
+
3890 train 3.946584 (lr=1.1075e-04) (hash(x)=70379093)
|
| 470 |
+
3900 val loss 3.9978
|
| 471 |
+
3900 val perplexity 54.4802
|
| 472 |
+
3900 train 4.078245 (lr=1.0951e-04) (hash(x)=76597431)
|
| 473 |
+
3910 train 3.959372 (lr=1.0830e-04) (hash(x)=90490716)
|
| 474 |
+
3920 train 4.055665 (lr=1.0711e-04) (hash(x)=81970659)
|
| 475 |
+
3930 train 4.047575 (lr=1.0594e-04) (hash(x)=81496334)
|
| 476 |
+
3940 train 4.018819 (lr=1.0480e-04) (hash(x)=75717605)
|
| 477 |
+
3950 train 3.970653 (lr=1.0369e-04) (hash(x)=82749357)
|
| 478 |
+
3960 train 4.004234 (lr=1.0260e-04) (hash(x)=89355157)
|
| 479 |
+
3970 train 3.954911 (lr=1.0153e-04) (hash(x)=78980403)
|
| 480 |
+
3980 train 3.928860 (lr=1.0049e-04) (hash(x)=76627217)
|
| 481 |
+
3990 train 3.928572 (lr=9.9479e-05) (hash(x)=72412879)
|
| 482 |
+
4000 val loss 3.9930
|
| 483 |
+
4000 val perplexity 54.2161
|
| 484 |
+
4000 train 3.938426 (lr=9.8489e-05) (hash(x)=83018142)
|
| 485 |
+
4010 train 3.956633 (lr=9.7524e-05) (hash(x)=81272436)
|
| 486 |
+
4020 train 3.887358 (lr=9.6585e-05) (hash(x)=85497482)
|
| 487 |
+
4030 train 3.965064 (lr=9.5670e-05) (hash(x)=85241734)
|
| 488 |
+
4040 train 3.925635 (lr=9.4781e-05) (hash(x)=77925307)
|
| 489 |
+
4050 train 3.921662 (lr=9.3917e-05) (hash(x)=84826179)
|
| 490 |
+
4060 train 3.984097 (lr=9.3079e-05) (hash(x)=83606764)
|
| 491 |
+
4070 train 3.895870 (lr=9.2267e-05) (hash(x)=80567590)
|
| 492 |
+
4080 train 3.945748 (lr=9.1480e-05) (hash(x)=76860998)
|
| 493 |
+
4090 train 3.983865 (lr=9.0718e-05) (hash(x)=74902328)
|
| 494 |
+
4100 val loss 3.9834
|
| 495 |
+
4100 val perplexity 53.7005
|
| 496 |
+
4100 train 4.029130 (lr=8.9983e-05) (hash(x)=82832041)
|
| 497 |
+
4110 train 3.910758 (lr=8.9273e-05) (hash(x)=79143262)
|
| 498 |
+
4120 train 3.930658 (lr=8.8589e-05) (hash(x)=77038149)
|
| 499 |
+
4130 train 4.117441 (lr=8.7931e-05) (hash(x)=86339074)
|
| 500 |
+
4140 train 4.181167 (lr=8.7299e-05) (hash(x)=76686216)
|
| 501 |
+
4150 train 3.959009 (lr=8.6693e-05) (hash(x)=70522682)
|
| 502 |
+
4160 train 3.927616 (lr=8.6113e-05) (hash(x)=90958555)
|
| 503 |
+
4170 train 3.950793 (lr=8.5559e-05) (hash(x)=91463532)
|
| 504 |
+
4180 train 3.939378 (lr=8.5031e-05) (hash(x)=81959329)
|
| 505 |
+
4190 train 3.890112 (lr=8.4529e-05) (hash(x)=83146752)
|
| 506 |
+
4200 val loss 3.9859
|
| 507 |
+
4200 val perplexity 53.8312
|
| 508 |
+
4200 train 3.878052 (lr=8.4054e-05) (hash(x)=78361715)
|
| 509 |
+
4210 train 4.027769 (lr=8.3605e-05) (hash(x)=87364889)
|
| 510 |
+
4220 train 3.875301 (lr=8.3182e-05) (hash(x)=70465156)
|
| 511 |
+
4230 train 4.028909 (lr=8.2785e-05) (hash(x)=84524081)
|
| 512 |
+
4240 train 3.988286 (lr=8.2414e-05) (hash(x)=77824868)
|
| 513 |
+
4250 train 4.000423 (lr=8.2070e-05) (hash(x)=81710711)
|
| 514 |
+
4260 train 3.982910 (lr=8.1752e-05) (hash(x)=76362728)
|
| 515 |
+
4270 train 3.972584 (lr=8.1461e-05) (hash(x)=83115208)
|
| 516 |
+
4280 train 3.884435 (lr=8.1196e-05) (hash(x)=87218314)
|
| 517 |
+
4290 train 4.044530 (lr=8.0958e-05) (hash(x)=74582673)
|
| 518 |
+
4300 val loss 3.9812
|
| 519 |
+
4300 val perplexity 53.5801
|
| 520 |
+
4300 train 3.918951 (lr=8.0746e-05) (hash(x)=77379615)
|
| 521 |
+
4310 train 3.898298 (lr=8.0560e-05) (hash(x)=78669579)
|
| 522 |
+
4320 train 4.018243 (lr=8.0401e-05) (hash(x)=83066608)
|
| 523 |
+
4330 train 3.976596 (lr=8.0269e-05) (hash(x)=83037340)
|
| 524 |
+
4340 train 4.018907 (lr=8.0162e-05) (hash(x)=82849771)
|
| 525 |
+
4350 train 3.950427 (lr=8.0083e-05) (hash(x)=76693985)
|
| 526 |
+
4360 train 4.038631 (lr=8.0030e-05) (hash(x)=77745394)
|
| 527 |
+
4370 train 4.000293 (lr=8.0003e-05) (hash(x)=79954388)
|
| 528 |
+
4374 val loss 3.9749
|
| 529 |
+
4374 val perplexity 53.2455
|
12_head_baseline_lr_8e-4_head_dim_22_fixed/model_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3f42195ae10961caf47df8ac4366ecbda39ae2b4b0dd9a35e6a93e860e8c0d0e
|
| 3 |
+
size 96858242
|
12_head_baseline_lr_8e-4_head_dim_22_fixed/optimizer_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:54f76bd93b577e794013bc8b0018ab4884260a80ce5605302f32cca0907b58cf
|
| 3 |
+
size 187435910
|