Upload folder using huggingface_hub
Browse files- logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/args.json +1 -0
- logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/dataloader_04374.pt +3 -0
- logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/log2.txt +548 -0
- logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/model_04374.pt +3 -0
- logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/optimizer_04374.pt +3 -0
logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/args.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"hellaswag": true, "attention_kind": "selective", "log_dir": "logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 13, "n_embd": 768, "head_dim": 64, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 4375, "warmup_steps": 250, "group": "64_heads_bump_comparison", "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, "disable_selection_head_linear_combo_bias": false, "assert_latent_matches_no_head": false, "n_latent_masks": null, "init_latent_masks_to_identity": false, "init_latent_masks_to_inverse": false, "latent_mask_scale": null, "latent_mask_runtime_multiplier": null, "latent_mask_sigmoid": false, "S_layernorm": false, "one_head_per_latent_mask": false, "protection_kind": "none", "leaky_relu_alpha": null, "leaky_relu_bias": null, "use_compile": false, "use_mini_model": false, "upload_to_hf": true, "seq_len": 1024, "batch_size": 8, "total_batch_size": 131072, "protection_head_scaling_factor": 1.0, "protection_head_bias": 0.0, "n_sliced_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": 0.003, "decay_lr": true, "readout_zero_init": false, "query_zero_init": false, "l1_loss": false, "debugpy": false, "key": "13_heads_no_compile", "latent_mask_precision": "bfloat16"}
|
logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/dataloader_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6031fd3e2855a036f7a5531cc24555aabd1115f9dd6618b8b2ca6f55279ef0b2
|
| 3 |
+
size 964
|
logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/log2.txt
ADDED
|
@@ -0,0 +1,548 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
max_steps: 4375
|
| 2 |
+
0 val loss 10.9799
|
| 3 |
+
0 val perplexity 58680.1328
|
| 4 |
+
0 hella 0.2500
|
| 5 |
+
0 train 10.975788 (lr=4.1958e-06) (hash(x)=44992657)
|
| 6 |
+
10 train 9.665457 (lr=4.6154e-05) (hash(x)=33468021)
|
| 7 |
+
20 train 9.108110 (lr=8.8112e-05) (hash(x)=40941803)
|
| 8 |
+
30 train 8.458189 (lr=1.3007e-04) (hash(x)=36715902)
|
| 9 |
+
40 train 7.814971 (lr=1.7203e-04) (hash(x)=32710993)
|
| 10 |
+
50 train 7.664196 (lr=2.1399e-04) (hash(x)=43839896)
|
| 11 |
+
60 train 7.676256 (lr=2.5594e-04) (hash(x)=40675468)
|
| 12 |
+
70 train 7.303211 (lr=2.9790e-04) (hash(x)=34592631)
|
| 13 |
+
80 train 7.239693 (lr=3.3986e-04) (hash(x)=44444845)
|
| 14 |
+
90 train 7.034295 (lr=3.8182e-04) (hash(x)=41965258)
|
| 15 |
+
100 val loss 6.9202
|
| 16 |
+
100 val perplexity 1012.5585
|
| 17 |
+
100 train 6.780492 (lr=4.2378e-04) (hash(x)=41284750)
|
| 18 |
+
110 train 6.714340 (lr=4.6573e-04) (hash(x)=41118734)
|
| 19 |
+
120 train 6.562747 (lr=5.0769e-04) (hash(x)=37537547)
|
| 20 |
+
130 train 6.411744 (lr=5.4965e-04) (hash(x)=43625179)
|
| 21 |
+
140 train 6.368657 (lr=5.9161e-04) (hash(x)=41940760)
|
| 22 |
+
150 train 6.460468 (lr=6.3357e-04) (hash(x)=39210431)
|
| 23 |
+
160 train 6.386385 (lr=6.7552e-04) (hash(x)=41128294)
|
| 24 |
+
170 train 6.276641 (lr=7.1748e-04) (hash(x)=41590227)
|
| 25 |
+
180 train 6.183324 (lr=7.5944e-04) (hash(x)=38084871)
|
| 26 |
+
190 train 6.111420 (lr=8.0140e-04) (hash(x)=34534333)
|
| 27 |
+
200 val loss 6.0384
|
| 28 |
+
200 val perplexity 419.2068
|
| 29 |
+
200 train 6.064824 (lr=8.4336e-04) (hash(x)=41614299)
|
| 30 |
+
210 train 5.945435 (lr=8.8531e-04) (hash(x)=35551597)
|
| 31 |
+
220 train 5.863481 (lr=9.2727e-04) (hash(x)=44615432)
|
| 32 |
+
230 train 5.966683 (lr=9.6923e-04) (hash(x)=37078768)
|
| 33 |
+
240 train 5.667850 (lr=1.0112e-03) (hash(x)=40644078)
|
| 34 |
+
250 hella 0.2366
|
| 35 |
+
250 train 5.740993 (lr=1.0531e-03) (hash(x)=44636066)
|
| 36 |
+
260 train 5.628315 (lr=1.0951e-03) (hash(x)=43247952)
|
| 37 |
+
270 train 5.560978 (lr=1.1371e-03) (hash(x)=37989370)
|
| 38 |
+
280 train 5.497177 (lr=1.1790e-03) (hash(x)=36229730)
|
| 39 |
+
290 train 5.448112 (lr=1.2210e-03) (hash(x)=35032996)
|
| 40 |
+
300 val loss 5.5537
|
| 41 |
+
300 val perplexity 258.2032
|
| 42 |
+
300 train 5.295135 (lr=1.2629e-03) (hash(x)=33868636)
|
| 43 |
+
310 train 5.212380 (lr=1.3049e-03) (hash(x)=39782705)
|
| 44 |
+
320 train 5.310715 (lr=1.3469e-03) (hash(x)=43525765)
|
| 45 |
+
330 train 5.424822 (lr=1.3888e-03) (hash(x)=38074926)
|
| 46 |
+
340 train 5.505033 (lr=1.4308e-03) (hash(x)=50125702)
|
| 47 |
+
350 train 5.448676 (lr=1.4727e-03) (hash(x)=35062222)
|
| 48 |
+
360 train 5.389750 (lr=1.5147e-03) (hash(x)=39671001)
|
| 49 |
+
370 train 5.410851 (lr=1.5566e-03) (hash(x)=35748364)
|
| 50 |
+
380 train 5.236575 (lr=1.5986e-03) (hash(x)=39962666)
|
| 51 |
+
390 train 5.194472 (lr=1.6406e-03) (hash(x)=39213360)
|
| 52 |
+
400 val loss 5.2069
|
| 53 |
+
400 val perplexity 182.5361
|
| 54 |
+
400 train 5.183053 (lr=1.6825e-03) (hash(x)=37919584)
|
| 55 |
+
410 train 5.169826 (lr=1.7245e-03) (hash(x)=42001527)
|
| 56 |
+
420 train 5.141800 (lr=1.7664e-03) (hash(x)=33940005)
|
| 57 |
+
430 train 5.173197 (lr=1.8084e-03) (hash(x)=41749217)
|
| 58 |
+
440 train 5.060253 (lr=1.8503e-03) (hash(x)=45219739)
|
| 59 |
+
450 train 5.049481 (lr=1.8923e-03) (hash(x)=38515090)
|
| 60 |
+
460 train 4.974513 (lr=1.9343e-03) (hash(x)=37518607)
|
| 61 |
+
470 train 4.746111 (lr=1.9762e-03) (hash(x)=46830716)
|
| 62 |
+
480 train 4.868115 (lr=2.0182e-03) (hash(x)=49262598)
|
| 63 |
+
490 train 4.667449 (lr=2.0601e-03) (hash(x)=41370559)
|
| 64 |
+
500 val loss 4.9739
|
| 65 |
+
500 val perplexity 144.5879
|
| 66 |
+
500 hella 0.2422
|
| 67 |
+
500 train 4.750510 (lr=2.1021e-03) (hash(x)=35237459)
|
| 68 |
+
510 train 4.841918 (lr=2.1441e-03) (hash(x)=35070579)
|
| 69 |
+
520 train 4.941739 (lr=2.1860e-03) (hash(x)=37355110)
|
| 70 |
+
530 train 4.982738 (lr=2.2280e-03) (hash(x)=39398044)
|
| 71 |
+
540 train 4.883339 (lr=2.2699e-03) (hash(x)=39156676)
|
| 72 |
+
550 train 4.819093 (lr=2.3119e-03) (hash(x)=38553909)
|
| 73 |
+
560 train 4.886508 (lr=2.3538e-03) (hash(x)=36555371)
|
| 74 |
+
570 train 4.770768 (lr=2.3958e-03) (hash(x)=42048991)
|
| 75 |
+
580 train 4.726003 (lr=2.4378e-03) (hash(x)=37359399)
|
| 76 |
+
590 train 4.850314 (lr=2.4797e-03) (hash(x)=39254663)
|
| 77 |
+
600 val loss 4.7721
|
| 78 |
+
600 val perplexity 118.1667
|
| 79 |
+
600 train 4.777578 (lr=2.5217e-03) (hash(x)=38982539)
|
| 80 |
+
610 train 4.768059 (lr=2.5636e-03) (hash(x)=35859145)
|
| 81 |
+
620 train 4.686257 (lr=2.6056e-03) (hash(x)=35593542)
|
| 82 |
+
630 train 4.585622 (lr=2.6476e-03) (hash(x)=39124418)
|
| 83 |
+
640 train 4.575212 (lr=2.6895e-03) (hash(x)=39107374)
|
| 84 |
+
650 train 4.492075 (lr=2.7315e-03) (hash(x)=34529151)
|
| 85 |
+
660 train 4.448810 (lr=2.7734e-03) (hash(x)=39483931)
|
| 86 |
+
670 train 4.403851 (lr=2.8154e-03) (hash(x)=43008010)
|
| 87 |
+
680 train 4.318262 (lr=2.8573e-03) (hash(x)=33469328)
|
| 88 |
+
690 train 4.340974 (lr=2.8993e-03) (hash(x)=35025253)
|
| 89 |
+
700 val loss 4.5832
|
| 90 |
+
700 val perplexity 97.8317
|
| 91 |
+
700 train 4.303390 (lr=2.9413e-03) (hash(x)=43280188)
|
| 92 |
+
710 train 4.620908 (lr=2.9832e-03) (hash(x)=40999263)
|
| 93 |
+
720 train 4.625040 (lr=3.0000e-03) (hash(x)=38304742)
|
| 94 |
+
730 train 4.604007 (lr=2.9999e-03) (hash(x)=34705188)
|
| 95 |
+
740 train 4.524267 (lr=2.9997e-03) (hash(x)=40028964)
|
| 96 |
+
750 hella 0.2549
|
| 97 |
+
750 train 4.472753 (lr=2.9994e-03) (hash(x)=39051591)
|
| 98 |
+
760 train 4.552953 (lr=2.9990e-03) (hash(x)=41286198)
|
| 99 |
+
770 train 4.479056 (lr=2.9985e-03) (hash(x)=44512967)
|
| 100 |
+
780 train 4.518661 (lr=2.9979e-03) (hash(x)=45171627)
|
| 101 |
+
790 train 4.476600 (lr=2.9972e-03) (hash(x)=46345673)
|
| 102 |
+
800 val loss 4.4628
|
| 103 |
+
800 val perplexity 86.7257
|
| 104 |
+
800 train 4.428338 (lr=2.9964e-03) (hash(x)=35529545)
|
| 105 |
+
810 train 4.331046 (lr=2.9955e-03) (hash(x)=42321793)
|
| 106 |
+
820 train 4.342955 (lr=2.9945e-03) (hash(x)=49327569)
|
| 107 |
+
830 train 4.367137 (lr=2.9934e-03) (hash(x)=41325057)
|
| 108 |
+
840 train 4.330583 (lr=2.9922e-03) (hash(x)=37305614)
|
| 109 |
+
850 train 4.319045 (lr=2.9909e-03) (hash(x)=38000800)
|
| 110 |
+
860 train 4.439419 (lr=2.9896e-03) (hash(x)=44047697)
|
| 111 |
+
870 train 4.417713 (lr=2.9881e-03) (hash(x)=41576206)
|
| 112 |
+
880 train 4.385619 (lr=2.9865e-03) (hash(x)=44685458)
|
| 113 |
+
890 train 4.354717 (lr=2.9848e-03) (hash(x)=40398597)
|
| 114 |
+
900 val loss 4.3530
|
| 115 |
+
900 val perplexity 77.7150
|
| 116 |
+
900 train 4.326511 (lr=2.9830e-03) (hash(x)=39312673)
|
| 117 |
+
910 train 4.336624 (lr=2.9811e-03) (hash(x)=38881875)
|
| 118 |
+
920 train 4.332971 (lr=2.9792e-03) (hash(x)=39140166)
|
| 119 |
+
930 train 4.247931 (lr=2.9771e-03) (hash(x)=39075781)
|
| 120 |
+
940 train 4.250366 (lr=2.9749e-03) (hash(x)=46887071)
|
| 121 |
+
950 train 4.255210 (lr=2.9726e-03) (hash(x)=39740603)
|
| 122 |
+
960 train 4.281078 (lr=2.9703e-03) (hash(x)=39590268)
|
| 123 |
+
970 train 4.253692 (lr=2.9678e-03) (hash(x)=43580052)
|
| 124 |
+
980 train 4.194341 (lr=2.9652e-03) (hash(x)=40226735)
|
| 125 |
+
990 train 4.274463 (lr=2.9626e-03) (hash(x)=40723896)
|
| 126 |
+
1000 val loss 4.2450
|
| 127 |
+
1000 val perplexity 69.7565
|
| 128 |
+
1000 hella 0.2532
|
| 129 |
+
1000 train 4.226531 (lr=2.9598e-03) (hash(x)=39408050)
|
| 130 |
+
1010 train 4.451937 (lr=2.9570e-03) (hash(x)=32247950)
|
| 131 |
+
1020 train 4.326344 (lr=2.9540e-03) (hash(x)=39049938)
|
| 132 |
+
1030 train 4.295628 (lr=2.9510e-03) (hash(x)=37453736)
|
| 133 |
+
1040 train 4.184593 (lr=2.9478e-03) (hash(x)=36649666)
|
| 134 |
+
1050 train 4.240912 (lr=2.9446e-03) (hash(x)=41404097)
|
| 135 |
+
1060 train 4.227542 (lr=2.9412e-03) (hash(x)=40701599)
|
| 136 |
+
1070 train 4.269124 (lr=2.9378e-03) (hash(x)=41787393)
|
| 137 |
+
1080 train 4.151515 (lr=2.9343e-03) (hash(x)=42325465)
|
| 138 |
+
1090 train 4.177477 (lr=2.9307e-03) (hash(x)=45018925)
|
| 139 |
+
1100 val loss 4.1750
|
| 140 |
+
1100 val perplexity 65.0380
|
| 141 |
+
1100 train 4.189117 (lr=2.9270e-03) (hash(x)=53751982)
|
| 142 |
+
1110 train 4.141866 (lr=2.9231e-03) (hash(x)=43118376)
|
| 143 |
+
1120 train 4.160814 (lr=2.9192e-03) (hash(x)=41489942)
|
| 144 |
+
1130 train 4.166775 (lr=2.9152e-03) (hash(x)=48020487)
|
| 145 |
+
1140 train 4.030068 (lr=2.9112e-03) (hash(x)=43266072)
|
| 146 |
+
1150 train 4.233179 (lr=2.9070e-03) (hash(x)=41790132)
|
| 147 |
+
1160 train 4.209124 (lr=2.9027e-03) (hash(x)=33739090)
|
| 148 |
+
1170 train 4.174531 (lr=2.8983e-03) (hash(x)=43885030)
|
| 149 |
+
1180 train 4.104410 (lr=2.8939e-03) (hash(x)=39046714)
|
| 150 |
+
1190 train 4.138717 (lr=2.8893e-03) (hash(x)=44387023)
|
| 151 |
+
1200 val loss 4.0876
|
| 152 |
+
1200 val perplexity 59.5975
|
| 153 |
+
1200 train 3.998727 (lr=2.8847e-03) (hash(x)=36667141)
|
| 154 |
+
1210 train 4.068636 (lr=2.8800e-03) (hash(x)=45487179)
|
| 155 |
+
1220 train 4.051069 (lr=2.8751e-03) (hash(x)=43456413)
|
| 156 |
+
1230 train 3.982053 (lr=2.8702e-03) (hash(x)=42473632)
|
| 157 |
+
1240 train 4.061913 (lr=2.8652e-03) (hash(x)=42966315)
|
| 158 |
+
1250 hella 0.2509
|
| 159 |
+
1250 train 3.921831 (lr=2.8601e-03) (hash(x)=42930524)
|
| 160 |
+
1260 train 4.055679 (lr=2.8550e-03) (hash(x)=43467676)
|
| 161 |
+
1270 train 3.993620 (lr=2.8497e-03) (hash(x)=40872580)
|
| 162 |
+
1280 train 4.045918 (lr=2.8443e-03) (hash(x)=40619157)
|
| 163 |
+
1290 train 4.034846 (lr=2.8389e-03) (hash(x)=44406165)
|
| 164 |
+
1300 val loss 4.0380
|
| 165 |
+
1300 val perplexity 56.7133
|
| 166 |
+
1300 train 4.019643 (lr=2.8333e-03) (hash(x)=40808029)
|
| 167 |
+
1310 train 4.189798 (lr=2.8277e-03) (hash(x)=46125736)
|
| 168 |
+
1320 train 3.957351 (lr=2.8220e-03) (hash(x)=41698487)
|
| 169 |
+
1330 train 4.002064 (lr=2.8162e-03) (hash(x)=39302878)
|
| 170 |
+
1340 train 4.108329 (lr=2.8103e-03) (hash(x)=41826369)
|
| 171 |
+
1350 train 3.927212 (lr=2.8044e-03) (hash(x)=37061413)
|
| 172 |
+
1360 train 4.057703 (lr=2.7983e-03) (hash(x)=40770974)
|
| 173 |
+
1370 train 3.953366 (lr=2.7922e-03) (hash(x)=38738718)
|
| 174 |
+
1380 train 3.958253 (lr=2.7860e-03) (hash(x)=34961558)
|
| 175 |
+
1390 train 4.018566 (lr=2.7797e-03) (hash(x)=38559342)
|
| 176 |
+
1400 val loss 4.0117
|
| 177 |
+
1400 val perplexity 55.2403
|
| 178 |
+
1400 train 3.964569 (lr=2.7733e-03) (hash(x)=38840020)
|
| 179 |
+
1410 train 4.040871 (lr=2.7668e-03) (hash(x)=44448100)
|
| 180 |
+
1420 train 4.162654 (lr=2.7603e-03) (hash(x)=39199838)
|
| 181 |
+
1430 train 3.951420 (lr=2.7536e-03) (hash(x)=41669873)
|
| 182 |
+
1440 train 4.069230 (lr=2.7469e-03) (hash(x)=53253836)
|
| 183 |
+
1450 train 3.994077 (lr=2.7401e-03) (hash(x)=41320631)
|
| 184 |
+
1460 train 3.921625 (lr=2.7333e-03) (hash(x)=48462803)
|
| 185 |
+
1470 train 3.979478 (lr=2.7263e-03) (hash(x)=42356864)
|
| 186 |
+
1480 train 4.022914 (lr=2.7193e-03) (hash(x)=41118068)
|
| 187 |
+
1490 train 3.932044 (lr=2.7121e-03) (hash(x)=37242854)
|
| 188 |
+
1500 val loss 3.9581
|
| 189 |
+
1500 val perplexity 52.3590
|
| 190 |
+
1500 hella 0.2555
|
| 191 |
+
1500 train 3.964175 (lr=2.7050e-03) (hash(x)=40571031)
|
| 192 |
+
1510 train 3.952404 (lr=2.6977e-03) (hash(x)=35384608)
|
| 193 |
+
1520 train 3.927574 (lr=2.6903e-03) (hash(x)=35341056)
|
| 194 |
+
1530 train 3.848863 (lr=2.6829e-03) (hash(x)=36471779)
|
| 195 |
+
1540 train 3.890480 (lr=2.6754e-03) (hash(x)=40356153)
|
| 196 |
+
1550 train 3.938684 (lr=2.6678e-03) (hash(x)=40133377)
|
| 197 |
+
1560 train 3.848088 (lr=2.6602e-03) (hash(x)=33722160)
|
| 198 |
+
1570 train 3.968737 (lr=2.6525e-03) (hash(x)=39555295)
|
| 199 |
+
1580 train 3.888092 (lr=2.6447e-03) (hash(x)=35950666)
|
| 200 |
+
1590 train 4.028816 (lr=2.6368e-03) (hash(x)=36528535)
|
| 201 |
+
1600 val loss 3.9319
|
| 202 |
+
1600 val perplexity 51.0063
|
| 203 |
+
1600 train 3.935637 (lr=2.6289e-03) (hash(x)=42613017)
|
| 204 |
+
1610 train 4.038793 (lr=2.6208e-03) (hash(x)=42845965)
|
| 205 |
+
1620 train 3.970455 (lr=2.6128e-03) (hash(x)=36137074)
|
| 206 |
+
1630 train 3.926505 (lr=2.6046e-03) (hash(x)=37560569)
|
| 207 |
+
1640 train 4.019643 (lr=2.5964e-03) (hash(x)=43680614)
|
| 208 |
+
1650 train 3.890693 (lr=2.5881e-03) (hash(x)=37281218)
|
| 209 |
+
1660 train 3.903826 (lr=2.5797e-03) (hash(x)=36165222)
|
| 210 |
+
1670 train 3.856676 (lr=2.5713e-03) (hash(x)=50222513)
|
| 211 |
+
1680 train 3.750254 (lr=2.5628e-03) (hash(x)=35239173)
|
| 212 |
+
1690 train 3.784851 (lr=2.5542e-03) (hash(x)=40287349)
|
| 213 |
+
1700 val loss 3.9194
|
| 214 |
+
1700 val perplexity 50.3712
|
| 215 |
+
1700 train 3.992966 (lr=2.5455e-03) (hash(x)=37444780)
|
| 216 |
+
1710 train 3.937276 (lr=2.5368e-03) (hash(x)=38962723)
|
| 217 |
+
1720 train 3.962173 (lr=2.5281e-03) (hash(x)=33882305)
|
| 218 |
+
1730 train 3.943981 (lr=2.5192e-03) (hash(x)=36501540)
|
| 219 |
+
1740 train 3.931055 (lr=2.5103e-03) (hash(x)=38723266)
|
| 220 |
+
1750 hella 0.2567
|
| 221 |
+
1750 train 3.886911 (lr=2.5014e-03) (hash(x)=38772789)
|
| 222 |
+
1760 train 3.952771 (lr=2.4924e-03) (hash(x)=37707898)
|
| 223 |
+
1770 train 3.822035 (lr=2.4833e-03) (hash(x)=30529327)
|
| 224 |
+
1780 train 3.890036 (lr=2.4741e-03) (hash(x)=37885464)
|
| 225 |
+
1790 train 3.932748 (lr=2.4649e-03) (hash(x)=39013967)
|
| 226 |
+
1800 val loss 3.8809
|
| 227 |
+
1800 val perplexity 48.4694
|
| 228 |
+
1800 train 3.910031 (lr=2.4556e-03) (hash(x)=34906955)
|
| 229 |
+
1810 train 3.829335 (lr=2.4463e-03) (hash(x)=40790682)
|
| 230 |
+
1820 train 3.703560 (lr=2.4369e-03) (hash(x)=39912223)
|
| 231 |
+
1830 train 3.810068 (lr=2.4275e-03) (hash(x)=39304486)
|
| 232 |
+
1840 train 3.812281 (lr=2.4180e-03) (hash(x)=37247975)
|
| 233 |
+
1850 train 3.774720 (lr=2.4084e-03) (hash(x)=41340614)
|
| 234 |
+
1860 train 3.931303 (lr=2.3988e-03) (hash(x)=39924627)
|
| 235 |
+
1870 train 3.894182 (lr=2.3891e-03) (hash(x)=40235993)
|
| 236 |
+
1880 train 3.797786 (lr=2.3794e-03) (hash(x)=37693476)
|
| 237 |
+
1890 train 3.921664 (lr=2.3696e-03) (hash(x)=40449864)
|
| 238 |
+
1900 val loss 3.8404
|
| 239 |
+
1900 val perplexity 46.5446
|
| 240 |
+
1900 train 3.808528 (lr=2.3598e-03) (hash(x)=37907749)
|
| 241 |
+
1910 train 3.796353 (lr=2.3499e-03) (hash(x)=34547901)
|
| 242 |
+
1920 train 3.840460 (lr=2.3400e-03) (hash(x)=37771788)
|
| 243 |
+
1930 train 3.713171 (lr=2.3300e-03) (hash(x)=36049454)
|
| 244 |
+
1940 train 3.795187 (lr=2.3200e-03) (hash(x)=29564336)
|
| 245 |
+
1950 train 3.798561 (lr=2.3099e-03) (hash(x)=41689281)
|
| 246 |
+
1960 train 3.703263 (lr=2.2998e-03) (hash(x)=41976699)
|
| 247 |
+
1970 train 3.900806 (lr=2.2896e-03) (hash(x)=41538621)
|
| 248 |
+
1980 train 3.772907 (lr=2.2793e-03) (hash(x)=36490902)
|
| 249 |
+
1990 train 3.790200 (lr=2.2691e-03) (hash(x)=42739315)
|
| 250 |
+
2000 val loss 3.8173
|
| 251 |
+
2000 val perplexity 45.4827
|
| 252 |
+
2000 hella 0.2609
|
| 253 |
+
2000 train 3.793734 (lr=2.2588e-03) (hash(x)=38600074)
|
| 254 |
+
2010 train 3.818375 (lr=2.2484e-03) (hash(x)=34555488)
|
| 255 |
+
2020 train 3.836847 (lr=2.2380e-03) (hash(x)=37227345)
|
| 256 |
+
2030 train 3.806353 (lr=2.2275e-03) (hash(x)=38172386)
|
| 257 |
+
2040 train 3.907732 (lr=2.2170e-03) (hash(x)=39340786)
|
| 258 |
+
2050 train 3.774353 (lr=2.2065e-03) (hash(x)=37237298)
|
| 259 |
+
2060 train 4.009146 (lr=2.1959e-03) (hash(x)=35894782)
|
| 260 |
+
2070 train 3.801398 (lr=2.1853e-03) (hash(x)=44161861)
|
| 261 |
+
2080 train 3.838992 (lr=2.1746e-03) (hash(x)=38062558)
|
| 262 |
+
2090 train 3.686403 (lr=2.1639e-03) (hash(x)=38608301)
|
| 263 |
+
2100 val loss 3.7821
|
| 264 |
+
2100 val perplexity 43.9078
|
| 265 |
+
2100 train 3.712766 (lr=2.1532e-03) (hash(x)=47611537)
|
| 266 |
+
2110 train 3.736415 (lr=2.1424e-03) (hash(x)=38076651)
|
| 267 |
+
2120 train 3.711239 (lr=2.1316e-03) (hash(x)=39870893)
|
| 268 |
+
2130 train 3.858063 (lr=2.1208e-03) (hash(x)=48123321)
|
| 269 |
+
2140 train 3.863752 (lr=2.1099e-03) (hash(x)=44124977)
|
| 270 |
+
2150 train 3.784827 (lr=2.0990e-03) (hash(x)=42181374)
|
| 271 |
+
2160 train 3.812240 (lr=2.0881e-03) (hash(x)=41837878)
|
| 272 |
+
2170 train 3.837761 (lr=2.0771e-03) (hash(x)=42161375)
|
| 273 |
+
2180 train 3.793222 (lr=2.0661e-03) (hash(x)=41792986)
|
| 274 |
+
2190 train 3.770051 (lr=2.0550e-03) (hash(x)=41872416)
|
| 275 |
+
2200 val loss 3.7609
|
| 276 |
+
2200 val perplexity 42.9869
|
| 277 |
+
2200 train 3.772211 (lr=2.0440e-03) (hash(x)=37395985)
|
| 278 |
+
2210 train 3.747992 (lr=2.0329e-03) (hash(x)=40616096)
|
| 279 |
+
2220 train 3.707936 (lr=2.0217e-03) (hash(x)=40118423)
|
| 280 |
+
2230 train 3.635084 (lr=2.0106e-03) (hash(x)=39641586)
|
| 281 |
+
2240 train 3.707739 (lr=1.9994e-03) (hash(x)=34720116)
|
| 282 |
+
2250 hella 0.2570
|
| 283 |
+
2250 train 3.666792 (lr=1.9882e-03) (hash(x)=43641508)
|
| 284 |
+
2260 train 3.677280 (lr=1.9770e-03) (hash(x)=31179786)
|
| 285 |
+
2270 train 3.645961 (lr=1.9657e-03) (hash(x)=42721932)
|
| 286 |
+
2280 train 3.950494 (lr=1.9544e-03) (hash(x)=38474505)
|
| 287 |
+
2290 train 3.767933 (lr=1.9431e-03) (hash(x)=41398545)
|
| 288 |
+
2300 val loss 3.7371
|
| 289 |
+
2300 val perplexity 41.9772
|
| 290 |
+
2300 train 3.697423 (lr=1.9318e-03) (hash(x)=44131094)
|
| 291 |
+
2310 train 3.779630 (lr=1.9205e-03) (hash(x)=41667948)
|
| 292 |
+
2320 train 3.871509 (lr=1.9091e-03) (hash(x)=36339232)
|
| 293 |
+
2330 train 3.714457 (lr=1.8977e-03) (hash(x)=41630644)
|
| 294 |
+
2340 train 3.832514 (lr=1.8863e-03) (hash(x)=41278955)
|
| 295 |
+
2350 train 3.696287 (lr=1.8749e-03) (hash(x)=39491976)
|
| 296 |
+
2360 train 3.764062 (lr=1.8635e-03) (hash(x)=41771238)
|
| 297 |
+
2370 train 3.845859 (lr=1.8520e-03) (hash(x)=45374570)
|
| 298 |
+
2380 train 3.751914 (lr=1.8406e-03) (hash(x)=41419267)
|
| 299 |
+
2390 train 3.758378 (lr=1.8291e-03) (hash(x)=38869169)
|
| 300 |
+
2400 val loss 3.7187
|
| 301 |
+
2400 val perplexity 41.2107
|
| 302 |
+
2400 train 3.684544 (lr=1.8176e-03) (hash(x)=38619293)
|
| 303 |
+
2410 train 3.478309 (lr=1.8061e-03) (hash(x)=41600240)
|
| 304 |
+
2420 train 3.643728 (lr=1.7946e-03) (hash(x)=40891045)
|
| 305 |
+
2430 train 3.611424 (lr=1.7830e-03) (hash(x)=41023249)
|
| 306 |
+
2440 train 3.603184 (lr=1.7715e-03) (hash(x)=33813452)
|
| 307 |
+
2450 train 3.556242 (lr=1.7600e-03) (hash(x)=38464119)
|
| 308 |
+
2460 train 3.436129 (lr=1.7484e-03) (hash(x)=40699982)
|
| 309 |
+
2470 train 3.444068 (lr=1.7368e-03) (hash(x)=38254854)
|
| 310 |
+
2480 train 3.726636 (lr=1.7253e-03) (hash(x)=42736069)
|
| 311 |
+
2490 train 3.720722 (lr=1.7137e-03) (hash(x)=41928525)
|
| 312 |
+
2500 val loss 3.6964
|
| 313 |
+
2500 val perplexity 40.3011
|
| 314 |
+
2500 hella 0.2639
|
| 315 |
+
2500 train 3.674309 (lr=1.7021e-03) (hash(x)=45864011)
|
| 316 |
+
2510 train 3.709013 (lr=1.6906e-03) (hash(x)=41380960)
|
| 317 |
+
2520 train 3.692203 (lr=1.6790e-03) (hash(x)=38366545)
|
| 318 |
+
2530 train 3.729811 (lr=1.6674e-03) (hash(x)=41045176)
|
| 319 |
+
2540 train 3.738357 (lr=1.6558e-03) (hash(x)=43435705)
|
| 320 |
+
2550 train 3.710927 (lr=1.6442e-03) (hash(x)=40693090)
|
| 321 |
+
2560 train 3.613956 (lr=1.6326e-03) (hash(x)=40502478)
|
| 322 |
+
2570 train 3.670545 (lr=1.6210e-03) (hash(x)=38610920)
|
| 323 |
+
2580 train 3.577209 (lr=1.6094e-03) (hash(x)=38333499)
|
| 324 |
+
2590 train 3.641250 (lr=1.5979e-03) (hash(x)=41982736)
|
| 325 |
+
2600 val loss 3.6743
|
| 326 |
+
2600 val perplexity 39.4207
|
| 327 |
+
2600 train 3.623257 (lr=1.5863e-03) (hash(x)=37724702)
|
| 328 |
+
2610 train 3.564488 (lr=1.5747e-03) (hash(x)=39564630)
|
| 329 |
+
2620 train 3.485075 (lr=1.5632e-03) (hash(x)=38465803)
|
| 330 |
+
2630 train 3.394421 (lr=1.5516e-03) (hash(x)=36955007)
|
| 331 |
+
2640 train 3.468785 (lr=1.5400e-03) (hash(x)=32994568)
|
| 332 |
+
2650 train 3.640986 (lr=1.5285e-03) (hash(x)=41165765)
|
| 333 |
+
2660 train 3.649980 (lr=1.5170e-03) (hash(x)=42934000)
|
| 334 |
+
2670 train 3.813231 (lr=1.5054e-03) (hash(x)=42726316)
|
| 335 |
+
2680 train 3.684217 (lr=1.4939e-03) (hash(x)=35880318)
|
| 336 |
+
2690 train 3.621198 (lr=1.4824e-03) (hash(x)=41747329)
|
| 337 |
+
2700 val loss 3.6573
|
| 338 |
+
2700 val perplexity 38.7560
|
| 339 |
+
2700 train 3.709858 (lr=1.4709e-03) (hash(x)=40259630)
|
| 340 |
+
2710 train 3.845279 (lr=1.4594e-03) (hash(x)=39895798)
|
| 341 |
+
2720 train 3.614001 (lr=1.4480e-03) (hash(x)=36146683)
|
| 342 |
+
2730 train 3.576298 (lr=1.4365e-03) (hash(x)=36181984)
|
| 343 |
+
2740 train 3.563178 (lr=1.4251e-03) (hash(x)=43700349)
|
| 344 |
+
2750 hella 0.2630
|
| 345 |
+
2750 train 3.637477 (lr=1.4137e-03) (hash(x)=41610597)
|
| 346 |
+
2760 train 3.634203 (lr=1.4023e-03) (hash(x)=31183639)
|
| 347 |
+
2770 train 3.591848 (lr=1.3909e-03) (hash(x)=37722489)
|
| 348 |
+
2780 train 3.562665 (lr=1.3795e-03) (hash(x)=47290688)
|
| 349 |
+
2790 train 3.598109 (lr=1.3682e-03) (hash(x)=41205574)
|
| 350 |
+
2800 val loss 3.6377
|
| 351 |
+
2800 val perplexity 38.0032
|
| 352 |
+
2800 train 3.433228 (lr=1.3569e-03) (hash(x)=40257962)
|
| 353 |
+
2810 train 3.399247 (lr=1.3456e-03) (hash(x)=39529014)
|
| 354 |
+
2820 train 3.416475 (lr=1.3343e-03) (hash(x)=42244749)
|
| 355 |
+
2830 train 3.486704 (lr=1.3230e-03) (hash(x)=32820090)
|
| 356 |
+
2840 train 3.435577 (lr=1.3118e-03) (hash(x)=40315769)
|
| 357 |
+
2850 train 3.662295 (lr=1.3006e-03) (hash(x)=41524462)
|
| 358 |
+
2860 train 3.663531 (lr=1.2894e-03) (hash(x)=38365734)
|
| 359 |
+
2870 train 3.784638 (lr=1.2783e-03) (hash(x)=37682602)
|
| 360 |
+
2880 train 3.655509 (lr=1.2671e-03) (hash(x)=39162991)
|
| 361 |
+
2890 train 3.517651 (lr=1.2560e-03) (hash(x)=33316384)
|
| 362 |
+
2900 val loss 3.6133
|
| 363 |
+
2900 val perplexity 37.0894
|
| 364 |
+
2900 train 3.562923 (lr=1.2450e-03) (hash(x)=37271132)
|
| 365 |
+
2910 train 3.752937 (lr=1.2339e-03) (hash(x)=35586242)
|
| 366 |
+
2920 train 3.672555 (lr=1.2229e-03) (hash(x)=33320586)
|
| 367 |
+
2930 train 3.638790 (lr=1.2119e-03) (hash(x)=43531361)
|
| 368 |
+
2940 train 3.568851 (lr=1.2010e-03) (hash(x)=37368286)
|
| 369 |
+
2950 train 3.576710 (lr=1.1901e-03) (hash(x)=40363394)
|
| 370 |
+
2960 train 3.552164 (lr=1.1792e-03) (hash(x)=42253792)
|
| 371 |
+
2970 train 3.478312 (lr=1.1684e-03) (hash(x)=38072598)
|
| 372 |
+
2980 train 3.582322 (lr=1.1576e-03) (hash(x)=41470557)
|
| 373 |
+
2990 train 3.480475 (lr=1.1468e-03) (hash(x)=42600033)
|
| 374 |
+
3000 val loss 3.6060
|
| 375 |
+
3000 val perplexity 36.8168
|
| 376 |
+
3000 hella 0.2613
|
| 377 |
+
3000 train 3.435587 (lr=1.1361e-03) (hash(x)=46890983)
|
| 378 |
+
3010 train 3.301002 (lr=1.1254e-03) (hash(x)=37986759)
|
| 379 |
+
3020 train 3.410786 (lr=1.1147e-03) (hash(x)=36270703)
|
| 380 |
+
3030 train 3.681029 (lr=1.1041e-03) (hash(x)=38228599)
|
| 381 |
+
3040 train 3.514694 (lr=1.0935e-03) (hash(x)=44344296)
|
| 382 |
+
3050 train 3.622338 (lr=1.0830e-03) (hash(x)=39962297)
|
| 383 |
+
3060 train 3.612547 (lr=1.0725e-03) (hash(x)=39817394)
|
| 384 |
+
3070 train 3.590033 (lr=1.0620e-03) (hash(x)=41763868)
|
| 385 |
+
3080 train 3.699841 (lr=1.0516e-03) (hash(x)=42343051)
|
| 386 |
+
3090 train 3.674312 (lr=1.0412e-03) (hash(x)=39204893)
|
| 387 |
+
3100 val loss 3.5812
|
| 388 |
+
3100 val perplexity 35.9177
|
| 389 |
+
3100 train 3.582209 (lr=1.0309e-03) (hash(x)=39353599)
|
| 390 |
+
3110 train 3.671150 (lr=1.0207e-03) (hash(x)=36118073)
|
| 391 |
+
3120 train 3.345946 (lr=1.0104e-03) (hash(x)=44070271)
|
| 392 |
+
3130 train 3.407874 (lr=1.0002e-03) (hash(x)=43368151)
|
| 393 |
+
3140 train 3.471335 (lr=9.9011e-04) (hash(x)=40612203)
|
| 394 |
+
3150 train 3.591015 (lr=9.8002e-04) (hash(x)=39598447)
|
| 395 |
+
3160 train 3.614233 (lr=9.6999e-04) (hash(x)=43897682)
|
| 396 |
+
3170 train 3.628272 (lr=9.6000e-04) (hash(x)=39029243)
|
| 397 |
+
3180 train 3.622953 (lr=9.5007e-04) (hash(x)=38586684)
|
| 398 |
+
3190 train 3.598965 (lr=9.4019e-04) (hash(x)=39299115)
|
| 399 |
+
3200 val loss 3.5578
|
| 400 |
+
3200 val perplexity 35.0848
|
| 401 |
+
3200 train 3.598619 (lr=9.3036e-04) (hash(x)=41064949)
|
| 402 |
+
3210 train 3.601455 (lr=9.2058e-04) (hash(x)=41752822)
|
| 403 |
+
3220 train 3.616739 (lr=9.1085e-04) (hash(x)=37484805)
|
| 404 |
+
3230 train 3.579964 (lr=9.0118e-04) (hash(x)=40514919)
|
| 405 |
+
3240 train 3.603373 (lr=8.9157e-04) (hash(x)=42440343)
|
| 406 |
+
3250 hella 0.2646
|
| 407 |
+
3250 train 3.521297 (lr=8.8201e-04) (hash(x)=32464136)
|
| 408 |
+
3260 train 3.536327 (lr=8.7251e-04) (hash(x)=47270558)
|
| 409 |
+
3270 train 3.500509 (lr=8.6307e-04) (hash(x)=39307546)
|
| 410 |
+
3280 train 3.525887 (lr=8.5368e-04) (hash(x)=40410299)
|
| 411 |
+
3290 train 3.529232 (lr=8.4435e-04) (hash(x)=40541497)
|
| 412 |
+
3300 val loss 3.5479
|
| 413 |
+
3300 val perplexity 34.7391
|
| 414 |
+
3300 train 3.429287 (lr=8.3508e-04) (hash(x)=38089139)
|
| 415 |
+
3310 train 3.445024 (lr=8.2588e-04) (hash(x)=42006576)
|
| 416 |
+
3320 train 3.554448 (lr=8.1673e-04) (hash(x)=40422556)
|
| 417 |
+
3330 train 3.317200 (lr=8.0764e-04) (hash(x)=39021643)
|
| 418 |
+
3340 train 3.606389 (lr=7.9862e-04) (hash(x)=40434305)
|
| 419 |
+
3350 train 3.590425 (lr=7.8966e-04) (hash(x)=40833559)
|
| 420 |
+
3360 train 3.616912 (lr=7.8076e-04) (hash(x)=40045231)
|
| 421 |
+
3370 train 3.558201 (lr=7.7192e-04) (hash(x)=34668317)
|
| 422 |
+
3380 train 3.561643 (lr=7.6315e-04) (hash(x)=42738568)
|
| 423 |
+
3390 train 3.665979 (lr=7.5445e-04) (hash(x)=38980166)
|
| 424 |
+
3400 val loss 3.5271
|
| 425 |
+
3400 val perplexity 34.0239
|
| 426 |
+
3400 train 3.596722 (lr=7.4581e-04) (hash(x)=43878176)
|
| 427 |
+
3410 train 3.620165 (lr=7.3724e-04) (hash(x)=38201991)
|
| 428 |
+
3420 train 3.603835 (lr=7.2874e-04) (hash(x)=34905889)
|
| 429 |
+
3430 train 3.599052 (lr=7.2030e-04) (hash(x)=41680405)
|
| 430 |
+
3440 train 3.563881 (lr=7.1193e-04) (hash(x)=39948184)
|
| 431 |
+
3450 train 3.608748 (lr=7.0363e-04) (hash(x)=33623728)
|
| 432 |
+
3460 train 3.495873 (lr=6.9541e-04) (hash(x)=41027647)
|
| 433 |
+
3470 train 3.507893 (lr=6.8725e-04) (hash(x)=42779482)
|
| 434 |
+
3480 train 3.512772 (lr=6.7916e-04) (hash(x)=41579288)
|
| 435 |
+
3490 train 3.382997 (lr=6.7114e-04) (hash(x)=37174846)
|
| 436 |
+
3500 val loss 3.5193
|
| 437 |
+
3500 val perplexity 33.7620
|
| 438 |
+
3500 hella 0.2645
|
| 439 |
+
3500 train 3.237833 (lr=6.6320e-04) (hash(x)=37082090)
|
| 440 |
+
3510 train 3.463198 (lr=6.5533e-04) (hash(x)=43131340)
|
| 441 |
+
3520 train 3.361223 (lr=6.4753e-04) (hash(x)=42457250)
|
| 442 |
+
3530 train 3.611265 (lr=6.3981e-04) (hash(x)=39504492)
|
| 443 |
+
3540 train 3.530120 (lr=6.3216e-04) (hash(x)=36225997)
|
| 444 |
+
3550 train 3.660761 (lr=6.2458e-04) (hash(x)=40912087)
|
| 445 |
+
3560 train 3.502155 (lr=6.1708e-04) (hash(x)=41545925)
|
| 446 |
+
3570 train 3.627885 (lr=6.0966e-04) (hash(x)=32553193)
|
| 447 |
+
3580 train 3.464769 (lr=6.0231e-04) (hash(x)=38799856)
|
| 448 |
+
3590 train 3.471333 (lr=5.9504e-04) (hash(x)=40772075)
|
| 449 |
+
3600 val loss 3.4999
|
| 450 |
+
3600 val perplexity 33.1137
|
| 451 |
+
3600 train 3.543131 (lr=5.8785e-04) (hash(x)=39299903)
|
| 452 |
+
3610 train 3.582037 (lr=5.8074e-04) (hash(x)=39503247)
|
| 453 |
+
3620 train 3.493942 (lr=5.7370e-04) (hash(x)=38730668)
|
| 454 |
+
3630 train 3.513918 (lr=5.6675e-04) (hash(x)=40503276)
|
| 455 |
+
3640 train 3.514417 (lr=5.5987e-04) (hash(x)=36504573)
|
| 456 |
+
3650 train 3.524462 (lr=5.5308e-04) (hash(x)=44903075)
|
| 457 |
+
3660 train 3.435188 (lr=5.4636e-04) (hash(x)=33800148)
|
| 458 |
+
3670 train 3.347687 (lr=5.3973e-04) (hash(x)=39404541)
|
| 459 |
+
3680 train 3.292408 (lr=5.3318e-04) (hash(x)=44777432)
|
| 460 |
+
3690 train 3.374566 (lr=5.2671e-04) (hash(x)=45649913)
|
| 461 |
+
3700 val loss 3.4962
|
| 462 |
+
3700 val perplexity 32.9883
|
| 463 |
+
3700 train 3.497900 (lr=5.2033e-04) (hash(x)=33812537)
|
| 464 |
+
3710 train 3.631359 (lr=5.1402e-04) (hash(x)=36161222)
|
| 465 |
+
3720 train 3.520831 (lr=5.0780e-04) (hash(x)=43690616)
|
| 466 |
+
3730 train 3.496832 (lr=5.0167e-04) (hash(x)=35462313)
|
| 467 |
+
3740 train 3.801235 (lr=4.9562e-04) (hash(x)=44674254)
|
| 468 |
+
3750 hella 0.2636
|
| 469 |
+
3750 train 3.514151 (lr=4.8965e-04) (hash(x)=44276297)
|
| 470 |
+
3760 train 3.542507 (lr=4.8377e-04) (hash(x)=38492178)
|
| 471 |
+
3770 train 3.567749 (lr=4.7798e-04) (hash(x)=44347531)
|
| 472 |
+
3780 train 3.484370 (lr=4.7227e-04) (hash(x)=33792059)
|
| 473 |
+
3790 train 3.542005 (lr=4.6665e-04) (hash(x)=38549318)
|
| 474 |
+
3800 val loss 3.4792
|
| 475 |
+
3800 val perplexity 32.4328
|
| 476 |
+
3800 train 3.458347 (lr=4.6112e-04) (hash(x)=41437448)
|
| 477 |
+
3810 train 3.364571 (lr=4.5567e-04) (hash(x)=36656768)
|
| 478 |
+
3820 train 3.529590 (lr=4.5031e-04) (hash(x)=46387893)
|
| 479 |
+
3830 train 3.513766 (lr=4.4504e-04) (hash(x)=44312129)
|
| 480 |
+
3840 train 3.425236 (lr=4.3986e-04) (hash(x)=40341896)
|
| 481 |
+
3850 train 3.467984 (lr=4.3477e-04) (hash(x)=43589160)
|
| 482 |
+
3860 train 3.473388 (lr=4.2977e-04) (hash(x)=40388945)
|
| 483 |
+
3870 train 3.355441 (lr=4.2486e-04) (hash(x)=33336679)
|
| 484 |
+
3880 train 3.478225 (lr=4.2004e-04) (hash(x)=37643311)
|
| 485 |
+
3890 train 3.469771 (lr=4.1530e-04) (hash(x)=38272927)
|
| 486 |
+
3900 val loss 3.4705
|
| 487 |
+
3900 val perplexity 32.1532
|
| 488 |
+
3900 train 3.424189 (lr=4.1066e-04) (hash(x)=41925748)
|
| 489 |
+
3910 train 3.516947 (lr=4.0611e-04) (hash(x)=38377253)
|
| 490 |
+
3920 train 3.512939 (lr=4.0166e-04) (hash(x)=41616611)
|
| 491 |
+
3930 train 3.445050 (lr=3.9729e-04) (hash(x)=32301827)
|
| 492 |
+
3940 train 3.513631 (lr=3.9302e-04) (hash(x)=47697363)
|
| 493 |
+
3950 train 3.453310 (lr=3.8884e-04) (hash(x)=37867767)
|
| 494 |
+
3960 train 3.462173 (lr=3.8475e-04) (hash(x)=37187295)
|
| 495 |
+
3970 train 3.480918 (lr=3.8076e-04) (hash(x)=41952752)
|
| 496 |
+
3980 train 3.455257 (lr=3.7685e-04) (hash(x)=38358660)
|
| 497 |
+
3990 train 3.363779 (lr=3.7305e-04) (hash(x)=40207878)
|
| 498 |
+
4000 val loss 3.4621
|
| 499 |
+
4000 val perplexity 31.8846
|
| 500 |
+
4000 hella 0.2639
|
| 501 |
+
4000 train 3.443136 (lr=3.6933e-04) (hash(x)=39134015)
|
| 502 |
+
4010 train 3.506713 (lr=3.6572e-04) (hash(x)=38313135)
|
| 503 |
+
4020 train 3.431940 (lr=3.6219e-04) (hash(x)=40710513)
|
| 504 |
+
4030 train 3.533258 (lr=3.5876e-04) (hash(x)=35530165)
|
| 505 |
+
4040 train 3.480905 (lr=3.5543e-04) (hash(x)=39494066)
|
| 506 |
+
4050 train 3.446665 (lr=3.5219e-04) (hash(x)=40861777)
|
| 507 |
+
4060 train 3.454400 (lr=3.4905e-04) (hash(x)=36151939)
|
| 508 |
+
4070 train 3.408767 (lr=3.4600e-04) (hash(x)=39008241)
|
| 509 |
+
4080 train 3.507400 (lr=3.4305e-04) (hash(x)=40395746)
|
| 510 |
+
4090 train 3.562262 (lr=3.4019e-04) (hash(x)=38855480)
|
| 511 |
+
4100 val loss 3.4504
|
| 512 |
+
4100 val perplexity 31.5131
|
| 513 |
+
4100 train 3.483893 (lr=3.3744e-04) (hash(x)=45013254)
|
| 514 |
+
4110 train 3.494424 (lr=3.3477e-04) (hash(x)=39934818)
|
| 515 |
+
4120 train 3.503133 (lr=3.3221e-04) (hash(x)=33417705)
|
| 516 |
+
4130 train 3.418663 (lr=3.2974e-04) (hash(x)=41267978)
|
| 517 |
+
4140 train 3.424051 (lr=3.2737e-04) (hash(x)=38961609)
|
| 518 |
+
4150 train 3.407336 (lr=3.2510e-04) (hash(x)=42173878)
|
| 519 |
+
4160 train 3.390195 (lr=3.2292e-04) (hash(x)=36867405)
|
| 520 |
+
4170 train 3.378022 (lr=3.2085e-04) (hash(x)=38991954)
|
| 521 |
+
4180 train 3.425336 (lr=3.1887e-04) (hash(x)=39032592)
|
| 522 |
+
4190 train 3.394394 (lr=3.1699e-04) (hash(x)=61329402)
|
| 523 |
+
4200 val loss 3.4463
|
| 524 |
+
4200 val perplexity 31.3837
|
| 525 |
+
4200 train 3.384862 (lr=3.1520e-04) (hash(x)=40455200)
|
| 526 |
+
4210 train 3.424593 (lr=3.1352e-04) (hash(x)=40913417)
|
| 527 |
+
4220 train 3.463577 (lr=3.1193e-04) (hash(x)=38809704)
|
| 528 |
+
4230 train 3.451466 (lr=3.1044e-04) (hash(x)=39560204)
|
| 529 |
+
4240 train 3.449147 (lr=3.0905e-04) (hash(x)=41496549)
|
| 530 |
+
4250 hella 0.2672
|
| 531 |
+
4250 train 3.513999 (lr=3.0776e-04) (hash(x)=33994768)
|
| 532 |
+
4260 train 3.478728 (lr=3.0657e-04) (hash(x)=39658687)
|
| 533 |
+
4270 train 3.387655 (lr=3.0548e-04) (hash(x)=42583064)
|
| 534 |
+
4280 train 3.453792 (lr=3.0449e-04) (hash(x)=40635811)
|
| 535 |
+
4290 train 3.355410 (lr=3.0359e-04) (hash(x)=35313841)
|
| 536 |
+
4300 val loss 3.4426
|
| 537 |
+
4300 val perplexity 31.2684
|
| 538 |
+
4300 train 3.518710 (lr=3.0280e-04) (hash(x)=39332245)
|
| 539 |
+
4310 train 3.399336 (lr=3.0210e-04) (hash(x)=43732999)
|
| 540 |
+
4320 train 3.388139 (lr=3.0150e-04) (hash(x)=40831835)
|
| 541 |
+
4330 train 3.546009 (lr=3.0101e-04) (hash(x)=41813368)
|
| 542 |
+
4340 train 3.448143 (lr=3.0061e-04) (hash(x)=36952586)
|
| 543 |
+
4350 train 3.438565 (lr=3.0031e-04) (hash(x)=42470745)
|
| 544 |
+
4360 train 3.427847 (lr=3.0011e-04) (hash(x)=39581495)
|
| 545 |
+
4370 train 3.570997 (lr=3.0001e-04) (hash(x)=35914538)
|
| 546 |
+
4374 val loss 3.4365
|
| 547 |
+
4374 val perplexity 31.0768
|
| 548 |
+
4374 hella 0.2667
|
logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/model_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:66b4efac2ef767fc14c26c0aff95b4b09c9e348380ca5e702e34facd530139ed
|
| 3 |
+
size 557599490
|
logs/64_heads_bump_comparison/13_heads_no_compile_seed_1339/optimizer_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e34daee76fbcbaa72de9b6203f849d20538bb2de973c8a4703e9bfd14a5c402f
|
| 3 |
+
size 1014545926
|