Upload folder using huggingface_hub
Browse files
scaling_mha_nh/mha_nh_32_seed_1339/args.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"hellaswag": true, "attention_kind": "dense", "log_dir": "scaling_mha_nh/mha_nh_32_seed_1339", "resume_checkpoint": null, "resume_optimizer": false, "add_a_head": false, "add_head_to_start": true, "new_head_init": "normal", "n_heads": 32, "n_embd": 256, "head_dim": 32, "protect_bos_token": true, "prevent_from_masking_myself": true, "max_steps": 4375, "warmup_steps": 250, "group": "scaling_mha_nh", "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, "att_conv": false, "att_conv_init": "none", "att_conv_scale": 1.0, "att_conv_weight_decay": false, "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": 32, "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, "mup_zero_init": true, "l1_loss": false, "debugpy": false, "key": "sdpa_nh_32", "latent_mask_precision": "bfloat16", "att_conv_precision": "bfloat16", "attn_precision": "bfloat16", "profile_kind": "none", "dense_attention_kind": "mha", "head_dim_value": 32, "sdpa_iter_size": null, "stabilize_attn_scores": true, "override_use_sdpa": false, "autocast_precision": "bfloat16", "simulate_micro_bs": null, "simulate_micro_bs_2": null}
|
scaling_mha_nh/mha_nh_32_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
|
scaling_mha_nh/mha_nh_32_seed_1339/log2.txt
ADDED
|
@@ -0,0 +1,529 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
max_steps: 4375
|
| 2 |
+
0 val loss 11.2854
|
| 3 |
+
0 val perplexity 79650.6250
|
| 4 |
+
0 train 11.284809 (lr=4.1958e-06) (hash(x)=44992657)
|
| 5 |
+
10 train 9.830477 (lr=4.6154e-05) (hash(x)=33468021)
|
| 6 |
+
20 train 9.166265 (lr=8.8112e-05) (hash(x)=40941803)
|
| 7 |
+
30 train 8.425510 (lr=1.3007e-04) (hash(x)=36715902)
|
| 8 |
+
40 train 7.683677 (lr=1.7203e-04) (hash(x)=32710993)
|
| 9 |
+
50 train 7.406013 (lr=2.1399e-04) (hash(x)=43839896)
|
| 10 |
+
60 train 7.368805 (lr=2.5594e-04) (hash(x)=40675468)
|
| 11 |
+
70 train 6.982787 (lr=2.9790e-04) (hash(x)=34592631)
|
| 12 |
+
80 train 6.996358 (lr=3.3986e-04) (hash(x)=44444845)
|
| 13 |
+
90 train 6.820381 (lr=3.8182e-04) (hash(x)=41965258)
|
| 14 |
+
100 val loss 6.7531
|
| 15 |
+
100 val perplexity 856.6851
|
| 16 |
+
100 train 6.615732 (lr=4.2378e-04) (hash(x)=41284750)
|
| 17 |
+
110 train 6.607637 (lr=4.6573e-04) (hash(x)=41118734)
|
| 18 |
+
120 train 6.502601 (lr=5.0769e-04) (hash(x)=37537547)
|
| 19 |
+
130 train 6.385648 (lr=5.4965e-04) (hash(x)=43625179)
|
| 20 |
+
140 train 6.412642 (lr=5.9161e-04) (hash(x)=41940760)
|
| 21 |
+
150 train 6.484372 (lr=6.3357e-04) (hash(x)=39210431)
|
| 22 |
+
160 train 6.455165 (lr=6.7552e-04) (hash(x)=41128294)
|
| 23 |
+
170 train 6.369158 (lr=7.1748e-04) (hash(x)=41590227)
|
| 24 |
+
180 train 6.311815 (lr=7.5944e-04) (hash(x)=38084871)
|
| 25 |
+
190 train 6.246418 (lr=8.0140e-04) (hash(x)=34534333)
|
| 26 |
+
200 val loss 6.1655
|
| 27 |
+
200 val perplexity 476.0399
|
| 28 |
+
200 train 6.198333 (lr=8.4336e-04) (hash(x)=41614299)
|
| 29 |
+
210 train 6.119189 (lr=8.8531e-04) (hash(x)=35551597)
|
| 30 |
+
220 train 6.073170 (lr=9.2727e-04) (hash(x)=44615432)
|
| 31 |
+
230 train 6.197480 (lr=9.6923e-04) (hash(x)=37078768)
|
| 32 |
+
240 train 5.917676 (lr=1.0112e-03) (hash(x)=40644078)
|
| 33 |
+
250 train 5.969959 (lr=1.0531e-03) (hash(x)=44636066)
|
| 34 |
+
260 train 5.864920 (lr=1.0951e-03) (hash(x)=43247952)
|
| 35 |
+
270 train 5.808851 (lr=1.1371e-03) (hash(x)=37989370)
|
| 36 |
+
280 train 5.745636 (lr=1.1790e-03) (hash(x)=36229730)
|
| 37 |
+
290 train 5.654582 (lr=1.2210e-03) (hash(x)=35032996)
|
| 38 |
+
300 val loss 5.7576
|
| 39 |
+
300 val perplexity 316.5827
|
| 40 |
+
300 train 5.513432 (lr=1.2629e-03) (hash(x)=33868636)
|
| 41 |
+
310 train 5.471865 (lr=1.3049e-03) (hash(x)=39782705)
|
| 42 |
+
320 train 5.543752 (lr=1.3469e-03) (hash(x)=43525765)
|
| 43 |
+
330 train 5.639886 (lr=1.3888e-03) (hash(x)=38074926)
|
| 44 |
+
340 train 5.665572 (lr=1.4308e-03) (hash(x)=50125702)
|
| 45 |
+
350 train 5.634722 (lr=1.4727e-03) (hash(x)=35062222)
|
| 46 |
+
360 train 5.577233 (lr=1.5147e-03) (hash(x)=39671001)
|
| 47 |
+
370 train 5.596127 (lr=1.5566e-03) (hash(x)=35748364)
|
| 48 |
+
380 train 5.399237 (lr=1.5986e-03) (hash(x)=39962666)
|
| 49 |
+
390 train 5.384563 (lr=1.6406e-03) (hash(x)=39213360)
|
| 50 |
+
400 val loss 5.3793
|
| 51 |
+
400 val perplexity 216.8602
|
| 52 |
+
400 train 5.363951 (lr=1.6825e-03) (hash(x)=37919584)
|
| 53 |
+
410 train 5.321618 (lr=1.7245e-03) (hash(x)=42001527)
|
| 54 |
+
420 train 5.308980 (lr=1.7664e-03) (hash(x)=33940005)
|
| 55 |
+
430 train 5.344046 (lr=1.8084e-03) (hash(x)=41749217)
|
| 56 |
+
440 train 5.223440 (lr=1.8503e-03) (hash(x)=45219739)
|
| 57 |
+
450 train 5.228865 (lr=1.8923e-03) (hash(x)=38515090)
|
| 58 |
+
460 train 5.131749 (lr=1.9343e-03) (hash(x)=37518607)
|
| 59 |
+
470 train 4.933170 (lr=1.9762e-03) (hash(x)=46830716)
|
| 60 |
+
480 train 5.091573 (lr=2.0182e-03) (hash(x)=49262598)
|
| 61 |
+
490 train 4.899391 (lr=2.0601e-03) (hash(x)=41370559)
|
| 62 |
+
500 val loss 5.1961
|
| 63 |
+
500 val perplexity 180.5627
|
| 64 |
+
500 train 4.983279 (lr=2.1021e-03) (hash(x)=35237459)
|
| 65 |
+
510 train 5.097660 (lr=2.1441e-03) (hash(x)=35070579)
|
| 66 |
+
520 train 5.145703 (lr=2.1860e-03) (hash(x)=37355110)
|
| 67 |
+
530 train 5.235869 (lr=2.2280e-03) (hash(x)=39398044)
|
| 68 |
+
540 train 5.171753 (lr=2.2699e-03) (hash(x)=39156676)
|
| 69 |
+
550 train 5.116210 (lr=2.3119e-03) (hash(x)=38553909)
|
| 70 |
+
560 train 5.169305 (lr=2.3538e-03) (hash(x)=36555371)
|
| 71 |
+
570 train 5.132010 (lr=2.3958e-03) (hash(x)=42048991)
|
| 72 |
+
580 train 5.062787 (lr=2.4378e-03) (hash(x)=37359399)
|
| 73 |
+
590 train 5.074041 (lr=2.4797e-03) (hash(x)=39254663)
|
| 74 |
+
600 val loss 5.0317
|
| 75 |
+
600 val perplexity 153.1981
|
| 76 |
+
600 train 5.058948 (lr=2.5217e-03) (hash(x)=38982539)
|
| 77 |
+
610 train 4.978674 (lr=2.5636e-03) (hash(x)=35859145)
|
| 78 |
+
620 train 4.948539 (lr=2.6056e-03) (hash(x)=35593542)
|
| 79 |
+
630 train 4.852872 (lr=2.6476e-03) (hash(x)=39124418)
|
| 80 |
+
640 train 4.859599 (lr=2.6895e-03) (hash(x)=39107374)
|
| 81 |
+
650 train 4.819085 (lr=2.7315e-03) (hash(x)=34529151)
|
| 82 |
+
660 train 4.728781 (lr=2.7734e-03) (hash(x)=39483931)
|
| 83 |
+
670 train 4.735322 (lr=2.8154e-03) (hash(x)=43008010)
|
| 84 |
+
680 train 4.627130 (lr=2.8573e-03) (hash(x)=33469328)
|
| 85 |
+
690 train 4.666843 (lr=2.8993e-03) (hash(x)=35025253)
|
| 86 |
+
700 val loss 4.9032
|
| 87 |
+
700 val perplexity 134.7164
|
| 88 |
+
700 train 4.635866 (lr=2.9413e-03) (hash(x)=43280188)
|
| 89 |
+
710 train 4.936527 (lr=2.9832e-03) (hash(x)=40999263)
|
| 90 |
+
720 train 4.884318 (lr=3.0000e-03) (hash(x)=38304742)
|
| 91 |
+
730 train 4.898561 (lr=2.9999e-03) (hash(x)=34705188)
|
| 92 |
+
740 train 4.854183 (lr=2.9997e-03) (hash(x)=40028964)
|
| 93 |
+
750 train 4.803912 (lr=2.9994e-03) (hash(x)=39051591)
|
| 94 |
+
760 train 4.832304 (lr=2.9990e-03) (hash(x)=41286198)
|
| 95 |
+
770 train 4.784498 (lr=2.9985e-03) (hash(x)=44512967)
|
| 96 |
+
780 train 4.796700 (lr=2.9979e-03) (hash(x)=45171627)
|
| 97 |
+
790 train 4.769855 (lr=2.9972e-03) (hash(x)=46345673)
|
| 98 |
+
800 val loss 4.7534
|
| 99 |
+
800 val perplexity 115.9759
|
| 100 |
+
800 train 4.715972 (lr=2.9964e-03) (hash(x)=35529545)
|
| 101 |
+
810 train 4.658166 (lr=2.9955e-03) (hash(x)=42321793)
|
| 102 |
+
820 train 4.644793 (lr=2.9945e-03) (hash(x)=49327569)
|
| 103 |
+
830 train 4.715639 (lr=2.9934e-03) (hash(x)=41325057)
|
| 104 |
+
840 train 4.660299 (lr=2.9922e-03) (hash(x)=37305614)
|
| 105 |
+
850 train 4.607965 (lr=2.9909e-03) (hash(x)=38000800)
|
| 106 |
+
860 train 4.722398 (lr=2.9896e-03) (hash(x)=44047697)
|
| 107 |
+
870 train 4.722746 (lr=2.9881e-03) (hash(x)=41576206)
|
| 108 |
+
880 train 4.700004 (lr=2.9865e-03) (hash(x)=44685458)
|
| 109 |
+
890 train 4.680209 (lr=2.9848e-03) (hash(x)=40398597)
|
| 110 |
+
900 val loss 4.6428
|
| 111 |
+
900 val perplexity 103.8307
|
| 112 |
+
900 train 4.610592 (lr=2.9830e-03) (hash(x)=39312673)
|
| 113 |
+
910 train 4.619824 (lr=2.9811e-03) (hash(x)=38881875)
|
| 114 |
+
920 train 4.650215 (lr=2.9792e-03) (hash(x)=39140166)
|
| 115 |
+
930 train 4.547155 (lr=2.9771e-03) (hash(x)=39075781)
|
| 116 |
+
940 train 4.566493 (lr=2.9749e-03) (hash(x)=46887071)
|
| 117 |
+
950 train 4.549455 (lr=2.9726e-03) (hash(x)=39740603)
|
| 118 |
+
960 train 4.610165 (lr=2.9703e-03) (hash(x)=39590268)
|
| 119 |
+
970 train 4.553783 (lr=2.9678e-03) (hash(x)=43580052)
|
| 120 |
+
980 train 4.491550 (lr=2.9652e-03) (hash(x)=40226735)
|
| 121 |
+
990 train 4.594784 (lr=2.9626e-03) (hash(x)=40723896)
|
| 122 |
+
1000 val loss 4.5423
|
| 123 |
+
1000 val perplexity 93.9031
|
| 124 |
+
1000 train 4.551541 (lr=2.9598e-03) (hash(x)=39408050)
|
| 125 |
+
1010 train 4.724638 (lr=2.9570e-03) (hash(x)=32247950)
|
| 126 |
+
1020 train 4.628192 (lr=2.9540e-03) (hash(x)=39049938)
|
| 127 |
+
1030 train 4.588807 (lr=2.9510e-03) (hash(x)=37453736)
|
| 128 |
+
1040 train 4.497368 (lr=2.9478e-03) (hash(x)=36649666)
|
| 129 |
+
1050 train 4.572933 (lr=2.9446e-03) (hash(x)=41404097)
|
| 130 |
+
1060 train 4.554344 (lr=2.9412e-03) (hash(x)=40701599)
|
| 131 |
+
1070 train 4.569267 (lr=2.9378e-03) (hash(x)=41787393)
|
| 132 |
+
1080 train 4.460709 (lr=2.9343e-03) (hash(x)=42325465)
|
| 133 |
+
1090 train 4.490216 (lr=2.9307e-03) (hash(x)=45018925)
|
| 134 |
+
1100 val loss 4.4794
|
| 135 |
+
1100 val perplexity 88.1858
|
| 136 |
+
1100 train 4.514963 (lr=2.9270e-03) (hash(x)=53751982)
|
| 137 |
+
1110 train 4.439380 (lr=2.9231e-03) (hash(x)=43118376)
|
| 138 |
+
1120 train 4.478083 (lr=2.9192e-03) (hash(x)=41489942)
|
| 139 |
+
1130 train 4.484215 (lr=2.9152e-03) (hash(x)=48020487)
|
| 140 |
+
1140 train 4.352102 (lr=2.9112e-03) (hash(x)=43266072)
|
| 141 |
+
1150 train 4.567564 (lr=2.9070e-03) (hash(x)=41790132)
|
| 142 |
+
1160 train 4.504201 (lr=2.9027e-03) (hash(x)=33739090)
|
| 143 |
+
1170 train 4.488771 (lr=2.8983e-03) (hash(x)=43885030)
|
| 144 |
+
1180 train 4.433151 (lr=2.8939e-03) (hash(x)=39046714)
|
| 145 |
+
1190 train 4.458859 (lr=2.8893e-03) (hash(x)=44387023)
|
| 146 |
+
1200 val loss 4.4022
|
| 147 |
+
1200 val perplexity 81.6340
|
| 148 |
+
1200 train 4.323317 (lr=2.8847e-03) (hash(x)=36667141)
|
| 149 |
+
1210 train 4.412539 (lr=2.8800e-03) (hash(x)=45487179)
|
| 150 |
+
1220 train 4.379907 (lr=2.8751e-03) (hash(x)=43456413)
|
| 151 |
+
1230 train 4.312606 (lr=2.8702e-03) (hash(x)=42473632)
|
| 152 |
+
1240 train 4.382661 (lr=2.8652e-03) (hash(x)=42966315)
|
| 153 |
+
1250 train 4.234878 (lr=2.8601e-03) (hash(x)=42930524)
|
| 154 |
+
1260 train 4.385152 (lr=2.8550e-03) (hash(x)=43467676)
|
| 155 |
+
1270 train 4.301885 (lr=2.8497e-03) (hash(x)=40872580)
|
| 156 |
+
1280 train 4.377190 (lr=2.8443e-03) (hash(x)=40619157)
|
| 157 |
+
1290 train 4.392077 (lr=2.8389e-03) (hash(x)=44406165)
|
| 158 |
+
1300 val loss 4.3610
|
| 159 |
+
1300 val perplexity 78.3372
|
| 160 |
+
1300 train 4.353482 (lr=2.8333e-03) (hash(x)=40808029)
|
| 161 |
+
1310 train 4.543764 (lr=2.8277e-03) (hash(x)=46125736)
|
| 162 |
+
1320 train 4.290177 (lr=2.8220e-03) (hash(x)=41698487)
|
| 163 |
+
1330 train 4.324107 (lr=2.8162e-03) (hash(x)=39302878)
|
| 164 |
+
1340 train 4.437029 (lr=2.8103e-03) (hash(x)=41826369)
|
| 165 |
+
1350 train 4.260354 (lr=2.8044e-03) (hash(x)=37061413)
|
| 166 |
+
1360 train 4.376945 (lr=2.7983e-03) (hash(x)=40770974)
|
| 167 |
+
1370 train 4.294771 (lr=2.7922e-03) (hash(x)=38738718)
|
| 168 |
+
1380 train 4.289461 (lr=2.7860e-03) (hash(x)=34961558)
|
| 169 |
+
1390 train 4.343825 (lr=2.7797e-03) (hash(x)=38559342)
|
| 170 |
+
1400 val loss 4.3425
|
| 171 |
+
1400 val perplexity 76.9031
|
| 172 |
+
1400 train 4.294889 (lr=2.7733e-03) (hash(x)=38840020)
|
| 173 |
+
1410 train 4.447262 (lr=2.7668e-03) (hash(x)=44448100)
|
| 174 |
+
1420 train 4.509228 (lr=2.7603e-03) (hash(x)=39199838)
|
| 175 |
+
1430 train 4.299708 (lr=2.7536e-03) (hash(x)=41669873)
|
| 176 |
+
1440 train 4.435995 (lr=2.7469e-03) (hash(x)=53253836)
|
| 177 |
+
1450 train 4.349401 (lr=2.7401e-03) (hash(x)=41320631)
|
| 178 |
+
1460 train 4.268394 (lr=2.7333e-03) (hash(x)=48462803)
|
| 179 |
+
1470 train 4.308894 (lr=2.7263e-03) (hash(x)=42356864)
|
| 180 |
+
1480 train 4.353677 (lr=2.7193e-03) (hash(x)=41118068)
|
| 181 |
+
1490 train 4.268279 (lr=2.7121e-03) (hash(x)=37242854)
|
| 182 |
+
1500 val loss 4.2913
|
| 183 |
+
1500 val perplexity 73.0616
|
| 184 |
+
1500 train 4.315035 (lr=2.7050e-03) (hash(x)=40571031)
|
| 185 |
+
1510 train 4.270278 (lr=2.6977e-03) (hash(x)=35384608)
|
| 186 |
+
1520 train 4.294670 (lr=2.6903e-03) (hash(x)=35341056)
|
| 187 |
+
1530 train 4.182796 (lr=2.6829e-03) (hash(x)=36471779)
|
| 188 |
+
1540 train 4.220690 (lr=2.6754e-03) (hash(x)=40356153)
|
| 189 |
+
1550 train 4.254799 (lr=2.6678e-03) (hash(x)=40133377)
|
| 190 |
+
1560 train 4.197288 (lr=2.6602e-03) (hash(x)=33722160)
|
| 191 |
+
1570 train 4.341755 (lr=2.6525e-03) (hash(x)=39555295)
|
| 192 |
+
1580 train 4.247889 (lr=2.6447e-03) (hash(x)=35950666)
|
| 193 |
+
1590 train 4.398032 (lr=2.6368e-03) (hash(x)=36528535)
|
| 194 |
+
1600 val loss 4.2720
|
| 195 |
+
1600 val perplexity 71.6642
|
| 196 |
+
1600 train 4.289515 (lr=2.6289e-03) (hash(x)=42613017)
|
| 197 |
+
1610 train 4.429567 (lr=2.6208e-03) (hash(x)=42845965)
|
| 198 |
+
1620 train 4.334587 (lr=2.6128e-03) (hash(x)=36137074)
|
| 199 |
+
1630 train 4.272837 (lr=2.6046e-03) (hash(x)=37560569)
|
| 200 |
+
1640 train 4.367486 (lr=2.5964e-03) (hash(x)=43680614)
|
| 201 |
+
1650 train 4.227739 (lr=2.5881e-03) (hash(x)=37281218)
|
| 202 |
+
1660 train 4.264148 (lr=2.5797e-03) (hash(x)=36165222)
|
| 203 |
+
1670 train 4.209046 (lr=2.5713e-03) (hash(x)=50222513)
|
| 204 |
+
1680 train 4.099841 (lr=2.5628e-03) (hash(x)=35239173)
|
| 205 |
+
1690 train 4.150151 (lr=2.5542e-03) (hash(x)=40287349)
|
| 206 |
+
1700 val loss 4.2696
|
| 207 |
+
1700 val perplexity 71.4966
|
| 208 |
+
1700 train 4.338011 (lr=2.5455e-03) (hash(x)=37444780)
|
| 209 |
+
1710 train 4.327559 (lr=2.5368e-03) (hash(x)=38962723)
|
| 210 |
+
1720 train 4.322629 (lr=2.5281e-03) (hash(x)=33882305)
|
| 211 |
+
1730 train 4.302654 (lr=2.5192e-03) (hash(x)=36501540)
|
| 212 |
+
1740 train 4.299812 (lr=2.5103e-03) (hash(x)=38723266)
|
| 213 |
+
1750 train 4.232088 (lr=2.5014e-03) (hash(x)=38772789)
|
| 214 |
+
1760 train 4.306221 (lr=2.4924e-03) (hash(x)=37707898)
|
| 215 |
+
1770 train 4.172493 (lr=2.4833e-03) (hash(x)=30529327)
|
| 216 |
+
1780 train 4.251302 (lr=2.4741e-03) (hash(x)=37885464)
|
| 217 |
+
1790 train 4.273821 (lr=2.4649e-03) (hash(x)=39013967)
|
| 218 |
+
1800 val loss 4.2305
|
| 219 |
+
1800 val perplexity 68.7526
|
| 220 |
+
1800 train 4.310595 (lr=2.4556e-03) (hash(x)=34906955)
|
| 221 |
+
1810 train 4.177126 (lr=2.4463e-03) (hash(x)=40790682)
|
| 222 |
+
1820 train 4.048768 (lr=2.4369e-03) (hash(x)=39912223)
|
| 223 |
+
1830 train 4.171343 (lr=2.4275e-03) (hash(x)=39304486)
|
| 224 |
+
1840 train 4.151830 (lr=2.4180e-03) (hash(x)=37247975)
|
| 225 |
+
1850 train 4.138722 (lr=2.4084e-03) (hash(x)=41340614)
|
| 226 |
+
1860 train 4.299340 (lr=2.3988e-03) (hash(x)=39924627)
|
| 227 |
+
1870 train 4.284440 (lr=2.3891e-03) (hash(x)=40235993)
|
| 228 |
+
1880 train 4.165924 (lr=2.3794e-03) (hash(x)=37693476)
|
| 229 |
+
1890 train 4.278731 (lr=2.3696e-03) (hash(x)=40449864)
|
| 230 |
+
1900 val loss 4.1953
|
| 231 |
+
1900 val perplexity 66.3715
|
| 232 |
+
1900 train 4.188219 (lr=2.3598e-03) (hash(x)=37907749)
|
| 233 |
+
1910 train 4.163033 (lr=2.3499e-03) (hash(x)=34547901)
|
| 234 |
+
1920 train 4.192661 (lr=2.3400e-03) (hash(x)=37771788)
|
| 235 |
+
1930 train 4.072543 (lr=2.3300e-03) (hash(x)=36049454)
|
| 236 |
+
1940 train 4.160332 (lr=2.3200e-03) (hash(x)=29564336)
|
| 237 |
+
1950 train 4.154293 (lr=2.3099e-03) (hash(x)=41689281)
|
| 238 |
+
1960 train 4.068491 (lr=2.2998e-03) (hash(x)=41976699)
|
| 239 |
+
1970 train 4.261730 (lr=2.2896e-03) (hash(x)=41538621)
|
| 240 |
+
1980 train 4.144556 (lr=2.2793e-03) (hash(x)=36490902)
|
| 241 |
+
1990 train 4.177210 (lr=2.2691e-03) (hash(x)=42739315)
|
| 242 |
+
2000 val loss 4.1810
|
| 243 |
+
2000 val perplexity 65.4314
|
| 244 |
+
2000 train 4.193404 (lr=2.2588e-03) (hash(x)=38600074)
|
| 245 |
+
2010 train 4.206016 (lr=2.2484e-03) (hash(x)=34555488)
|
| 246 |
+
2020 train 4.219810 (lr=2.2380e-03) (hash(x)=37227345)
|
| 247 |
+
2030 train 4.177958 (lr=2.2275e-03) (hash(x)=38172386)
|
| 248 |
+
2040 train 4.286770 (lr=2.2170e-03) (hash(x)=39340786)
|
| 249 |
+
2050 train 4.130740 (lr=2.2065e-03) (hash(x)=37237298)
|
| 250 |
+
2060 train 4.364669 (lr=2.1959e-03) (hash(x)=35894782)
|
| 251 |
+
2070 train 4.187634 (lr=2.1853e-03) (hash(x)=44161861)
|
| 252 |
+
2080 train 4.204561 (lr=2.1746e-03) (hash(x)=38062558)
|
| 253 |
+
2090 train 4.053796 (lr=2.1639e-03) (hash(x)=38608301)
|
| 254 |
+
2100 val loss 4.1512
|
| 255 |
+
2100 val perplexity 63.5081
|
| 256 |
+
2100 train 4.081339 (lr=2.1532e-03) (hash(x)=47611537)
|
| 257 |
+
2110 train 4.078694 (lr=2.1424e-03) (hash(x)=38076651)
|
| 258 |
+
2120 train 4.091169 (lr=2.1316e-03) (hash(x)=39870893)
|
| 259 |
+
2130 train 4.242083 (lr=2.1208e-03) (hash(x)=48123321)
|
| 260 |
+
2140 train 4.267059 (lr=2.1099e-03) (hash(x)=44124977)
|
| 261 |
+
2150 train 4.167246 (lr=2.0990e-03) (hash(x)=42181374)
|
| 262 |
+
2160 train 4.199125 (lr=2.0881e-03) (hash(x)=41837878)
|
| 263 |
+
2170 train 4.216421 (lr=2.0771e-03) (hash(x)=42161375)
|
| 264 |
+
2180 train 4.182087 (lr=2.0661e-03) (hash(x)=41792986)
|
| 265 |
+
2190 train 4.153755 (lr=2.0550e-03) (hash(x)=41872416)
|
| 266 |
+
2200 val loss 4.1313
|
| 267 |
+
2200 val perplexity 62.2577
|
| 268 |
+
2200 train 4.122929 (lr=2.0440e-03) (hash(x)=37395985)
|
| 269 |
+
2210 train 4.179051 (lr=2.0329e-03) (hash(x)=40616096)
|
| 270 |
+
2220 train 4.080559 (lr=2.0217e-03) (hash(x)=40118423)
|
| 271 |
+
2230 train 3.986907 (lr=2.0106e-03) (hash(x)=39641586)
|
| 272 |
+
2240 train 4.080826 (lr=1.9994e-03) (hash(x)=34720116)
|
| 273 |
+
2250 train 4.016824 (lr=1.9882e-03) (hash(x)=43641508)
|
| 274 |
+
2260 train 4.043573 (lr=1.9770e-03) (hash(x)=31179786)
|
| 275 |
+
2270 train 4.011892 (lr=1.9657e-03) (hash(x)=42721932)
|
| 276 |
+
2280 train 4.330641 (lr=1.9544e-03) (hash(x)=38474505)
|
| 277 |
+
2290 train 4.165168 (lr=1.9431e-03) (hash(x)=41398545)
|
| 278 |
+
2300 val loss 4.1098
|
| 279 |
+
2300 val perplexity 60.9365
|
| 280 |
+
2300 train 4.098065 (lr=1.9318e-03) (hash(x)=44131094)
|
| 281 |
+
2310 train 4.176984 (lr=1.9205e-03) (hash(x)=41667948)
|
| 282 |
+
2320 train 4.254792 (lr=1.9091e-03) (hash(x)=36339232)
|
| 283 |
+
2330 train 4.129395 (lr=1.8977e-03) (hash(x)=41630644)
|
| 284 |
+
2340 train 4.220294 (lr=1.8863e-03) (hash(x)=41278955)
|
| 285 |
+
2350 train 4.091653 (lr=1.8749e-03) (hash(x)=39491976)
|
| 286 |
+
2360 train 4.154227 (lr=1.8635e-03) (hash(x)=41771238)
|
| 287 |
+
2370 train 4.251866 (lr=1.8520e-03) (hash(x)=45374570)
|
| 288 |
+
2380 train 4.123452 (lr=1.8406e-03) (hash(x)=41419267)
|
| 289 |
+
2390 train 4.162728 (lr=1.8291e-03) (hash(x)=38869169)
|
| 290 |
+
2400 val loss 4.0984
|
| 291 |
+
2400 val perplexity 60.2433
|
| 292 |
+
2400 train 4.064785 (lr=1.8176e-03) (hash(x)=38619293)
|
| 293 |
+
2410 train 3.887310 (lr=1.8061e-03) (hash(x)=41600240)
|
| 294 |
+
2420 train 4.013243 (lr=1.7946e-03) (hash(x)=40891045)
|
| 295 |
+
2430 train 4.004417 (lr=1.7830e-03) (hash(x)=41023249)
|
| 296 |
+
2440 train 3.975204 (lr=1.7715e-03) (hash(x)=33813452)
|
| 297 |
+
2450 train 3.922843 (lr=1.7600e-03) (hash(x)=38464119)
|
| 298 |
+
2460 train 3.820163 (lr=1.7484e-03) (hash(x)=40699982)
|
| 299 |
+
2470 train 3.831806 (lr=1.7368e-03) (hash(x)=38254854)
|
| 300 |
+
2480 train 4.122474 (lr=1.7253e-03) (hash(x)=42736069)
|
| 301 |
+
2490 train 4.131282 (lr=1.7137e-03) (hash(x)=41928525)
|
| 302 |
+
2500 val loss 4.0755
|
| 303 |
+
2500 val perplexity 58.8826
|
| 304 |
+
2500 train 4.089149 (lr=1.7021e-03) (hash(x)=45864011)
|
| 305 |
+
2510 train 4.101380 (lr=1.6906e-03) (hash(x)=41380960)
|
| 306 |
+
2520 train 4.097755 (lr=1.6790e-03) (hash(x)=38366545)
|
| 307 |
+
2530 train 4.096876 (lr=1.6674e-03) (hash(x)=41045176)
|
| 308 |
+
2540 train 4.127111 (lr=1.6558e-03) (hash(x)=43435705)
|
| 309 |
+
2550 train 4.083611 (lr=1.6442e-03) (hash(x)=40693090)
|
| 310 |
+
2560 train 3.992281 (lr=1.6326e-03) (hash(x)=40502478)
|
| 311 |
+
2570 train 4.024837 (lr=1.6210e-03) (hash(x)=38610920)
|
| 312 |
+
2580 train 3.948712 (lr=1.6094e-03) (hash(x)=38333499)
|
| 313 |
+
2590 train 4.029912 (lr=1.5979e-03) (hash(x)=41982736)
|
| 314 |
+
2600 val loss 4.0569
|
| 315 |
+
2600 val perplexity 57.7929
|
| 316 |
+
2600 train 4.009383 (lr=1.5863e-03) (hash(x)=37724702)
|
| 317 |
+
2610 train 3.948321 (lr=1.5747e-03) (hash(x)=39564630)
|
| 318 |
+
2620 train 3.894463 (lr=1.5632e-03) (hash(x)=38465803)
|
| 319 |
+
2630 train 3.810352 (lr=1.5516e-03) (hash(x)=36955007)
|
| 320 |
+
2640 train 3.848025 (lr=1.5400e-03) (hash(x)=32994568)
|
| 321 |
+
2650 train 4.036009 (lr=1.5285e-03) (hash(x)=41165765)
|
| 322 |
+
2660 train 4.045570 (lr=1.5170e-03) (hash(x)=42934000)
|
| 323 |
+
2670 train 4.242981 (lr=1.5054e-03) (hash(x)=42726316)
|
| 324 |
+
2680 train 4.133905 (lr=1.4939e-03) (hash(x)=35880318)
|
| 325 |
+
2690 train 4.007298 (lr=1.4824e-03) (hash(x)=41747329)
|
| 326 |
+
2700 val loss 4.0456
|
| 327 |
+
2700 val perplexity 57.1464
|
| 328 |
+
2700 train 4.143479 (lr=1.4709e-03) (hash(x)=40259630)
|
| 329 |
+
2710 train 4.260278 (lr=1.4594e-03) (hash(x)=39895798)
|
| 330 |
+
2720 train 3.987914 (lr=1.4480e-03) (hash(x)=36146683)
|
| 331 |
+
2730 train 3.968343 (lr=1.4365e-03) (hash(x)=36181984)
|
| 332 |
+
2740 train 3.941955 (lr=1.4251e-03) (hash(x)=43700349)
|
| 333 |
+
2750 train 4.011082 (lr=1.4137e-03) (hash(x)=41610597)
|
| 334 |
+
2760 train 3.996949 (lr=1.4023e-03) (hash(x)=31183639)
|
| 335 |
+
2770 train 3.993346 (lr=1.3909e-03) (hash(x)=37722489)
|
| 336 |
+
2780 train 3.943166 (lr=1.3795e-03) (hash(x)=47290688)
|
| 337 |
+
2790 train 3.972713 (lr=1.3682e-03) (hash(x)=41205574)
|
| 338 |
+
2800 val loss 4.0298
|
| 339 |
+
2800 val perplexity 56.2506
|
| 340 |
+
2800 train 3.824418 (lr=1.3569e-03) (hash(x)=40257962)
|
| 341 |
+
2810 train 3.797223 (lr=1.3456e-03) (hash(x)=39529014)
|
| 342 |
+
2820 train 3.819875 (lr=1.3343e-03) (hash(x)=42244749)
|
| 343 |
+
2830 train 3.875522 (lr=1.3230e-03) (hash(x)=32820090)
|
| 344 |
+
2840 train 3.835784 (lr=1.3118e-03) (hash(x)=40315769)
|
| 345 |
+
2850 train 4.079268 (lr=1.3006e-03) (hash(x)=41524462)
|
| 346 |
+
2860 train 4.068507 (lr=1.2894e-03) (hash(x)=38365734)
|
| 347 |
+
2870 train 4.200030 (lr=1.2783e-03) (hash(x)=37682602)
|
| 348 |
+
2880 train 4.054609 (lr=1.2671e-03) (hash(x)=39162991)
|
| 349 |
+
2890 train 3.912656 (lr=1.2560e-03) (hash(x)=33316384)
|
| 350 |
+
2900 val loss 4.0030
|
| 351 |
+
2900 val perplexity 54.7646
|
| 352 |
+
2900 train 3.957451 (lr=1.2450e-03) (hash(x)=37271132)
|
| 353 |
+
2910 train 4.147655 (lr=1.2339e-03) (hash(x)=35586242)
|
| 354 |
+
2920 train 4.035683 (lr=1.2229e-03) (hash(x)=33320586)
|
| 355 |
+
2930 train 4.052052 (lr=1.2119e-03) (hash(x)=43531361)
|
| 356 |
+
2940 train 3.936853 (lr=1.2010e-03) (hash(x)=37368286)
|
| 357 |
+
2950 train 3.974908 (lr=1.1901e-03) (hash(x)=40363394)
|
| 358 |
+
2960 train 3.935876 (lr=1.1792e-03) (hash(x)=42253792)
|
| 359 |
+
2970 train 3.882682 (lr=1.1684e-03) (hash(x)=38072598)
|
| 360 |
+
2980 train 3.954075 (lr=1.1576e-03) (hash(x)=41470557)
|
| 361 |
+
2990 train 3.880297 (lr=1.1468e-03) (hash(x)=42600033)
|
| 362 |
+
3000 val loss 4.0015
|
| 363 |
+
3000 val perplexity 54.6774
|
| 364 |
+
3000 train 3.831636 (lr=1.1361e-03) (hash(x)=46890983)
|
| 365 |
+
3010 train 3.704132 (lr=1.1254e-03) (hash(x)=37986759)
|
| 366 |
+
3020 train 3.827353 (lr=1.1147e-03) (hash(x)=36270703)
|
| 367 |
+
3030 train 4.093076 (lr=1.1041e-03) (hash(x)=38228599)
|
| 368 |
+
3040 train 3.940425 (lr=1.0935e-03) (hash(x)=44344296)
|
| 369 |
+
3050 train 4.028215 (lr=1.0830e-03) (hash(x)=39962297)
|
| 370 |
+
3060 train 4.001305 (lr=1.0725e-03) (hash(x)=39817394)
|
| 371 |
+
3070 train 4.009473 (lr=1.0620e-03) (hash(x)=41763868)
|
| 372 |
+
3080 train 4.088186 (lr=1.0516e-03) (hash(x)=42343051)
|
| 373 |
+
3090 train 4.089188 (lr=1.0412e-03) (hash(x)=39204893)
|
| 374 |
+
3100 val loss 3.9790
|
| 375 |
+
3100 val perplexity 53.4627
|
| 376 |
+
3100 train 3.987189 (lr=1.0309e-03) (hash(x)=39353599)
|
| 377 |
+
3110 train 4.055553 (lr=1.0207e-03) (hash(x)=36118073)
|
| 378 |
+
3120 train 3.762566 (lr=1.0104e-03) (hash(x)=44070271)
|
| 379 |
+
3130 train 3.783309 (lr=1.0002e-03) (hash(x)=43368151)
|
| 380 |
+
3140 train 3.861881 (lr=9.9011e-04) (hash(x)=40612203)
|
| 381 |
+
3150 train 3.998831 (lr=9.8002e-04) (hash(x)=39598447)
|
| 382 |
+
3160 train 4.055288 (lr=9.6999e-04) (hash(x)=43897682)
|
| 383 |
+
3170 train 4.031361 (lr=9.6000e-04) (hash(x)=39029243)
|
| 384 |
+
3180 train 4.040031 (lr=9.5007e-04) (hash(x)=38586684)
|
| 385 |
+
3190 train 3.992250 (lr=9.4019e-04) (hash(x)=39299115)
|
| 386 |
+
3200 val loss 3.9572
|
| 387 |
+
3200 val perplexity 52.3127
|
| 388 |
+
3200 train 4.020811 (lr=9.3036e-04) (hash(x)=41064949)
|
| 389 |
+
3210 train 4.009590 (lr=9.2058e-04) (hash(x)=41752822)
|
| 390 |
+
3220 train 4.012295 (lr=9.1085e-04) (hash(x)=37484805)
|
| 391 |
+
3230 train 3.976330 (lr=9.0118e-04) (hash(x)=40514919)
|
| 392 |
+
3240 train 4.031167 (lr=8.9157e-04) (hash(x)=42440343)
|
| 393 |
+
3250 train 3.928448 (lr=8.8201e-04) (hash(x)=32464136)
|
| 394 |
+
3260 train 3.949047 (lr=8.7251e-04) (hash(x)=47270558)
|
| 395 |
+
3270 train 3.887468 (lr=8.6307e-04) (hash(x)=39307546)
|
| 396 |
+
3280 train 3.921851 (lr=8.5368e-04) (hash(x)=40410299)
|
| 397 |
+
3290 train 3.920793 (lr=8.4435e-04) (hash(x)=40541497)
|
| 398 |
+
3300 val loss 3.9516
|
| 399 |
+
3300 val perplexity 52.0173
|
| 400 |
+
3300 train 3.816418 (lr=8.3508e-04) (hash(x)=38089139)
|
| 401 |
+
3310 train 3.853703 (lr=8.2588e-04) (hash(x)=42006576)
|
| 402 |
+
3320 train 3.944588 (lr=8.1673e-04) (hash(x)=40422556)
|
| 403 |
+
3330 train 3.736684 (lr=8.0764e-04) (hash(x)=39021643)
|
| 404 |
+
3340 train 4.005380 (lr=7.9862e-04) (hash(x)=40434305)
|
| 405 |
+
3350 train 4.029197 (lr=7.8966e-04) (hash(x)=40833559)
|
| 406 |
+
3360 train 4.026943 (lr=7.8076e-04) (hash(x)=40045231)
|
| 407 |
+
3370 train 3.988830 (lr=7.7192e-04) (hash(x)=34668317)
|
| 408 |
+
3380 train 3.983141 (lr=7.6315e-04) (hash(x)=42738568)
|
| 409 |
+
3390 train 4.081949 (lr=7.5445e-04) (hash(x)=38980166)
|
| 410 |
+
3400 val loss 3.9306
|
| 411 |
+
3400 val perplexity 50.9350
|
| 412 |
+
3400 train 3.989050 (lr=7.4581e-04) (hash(x)=43878176)
|
| 413 |
+
3410 train 4.033318 (lr=7.3724e-04) (hash(x)=38201991)
|
| 414 |
+
3420 train 4.009029 (lr=7.2874e-04) (hash(x)=34905889)
|
| 415 |
+
3430 train 4.008903 (lr=7.2030e-04) (hash(x)=41680405)
|
| 416 |
+
3440 train 3.973709 (lr=7.1193e-04) (hash(x)=39948184)
|
| 417 |
+
3450 train 4.015001 (lr=7.0363e-04) (hash(x)=33623728)
|
| 418 |
+
3460 train 3.891722 (lr=6.9541e-04) (hash(x)=41027647)
|
| 419 |
+
3470 train 3.922876 (lr=6.8725e-04) (hash(x)=42779482)
|
| 420 |
+
3480 train 3.938257 (lr=6.7916e-04) (hash(x)=41579288)
|
| 421 |
+
3490 train 3.794620 (lr=6.7114e-04) (hash(x)=37174846)
|
| 422 |
+
3500 val loss 3.9247
|
| 423 |
+
3500 val perplexity 50.6382
|
| 424 |
+
3500 train 3.674345 (lr=6.6320e-04) (hash(x)=37082090)
|
| 425 |
+
3510 train 3.871007 (lr=6.5533e-04) (hash(x)=43131340)
|
| 426 |
+
3520 train 3.781874 (lr=6.4753e-04) (hash(x)=42457250)
|
| 427 |
+
3530 train 4.016878 (lr=6.3981e-04) (hash(x)=39504492)
|
| 428 |
+
3540 train 3.964962 (lr=6.3216e-04) (hash(x)=36225997)
|
| 429 |
+
3550 train 4.089664 (lr=6.2458e-04) (hash(x)=40912087)
|
| 430 |
+
3560 train 3.930863 (lr=6.1708e-04) (hash(x)=41545925)
|
| 431 |
+
3570 train 4.015086 (lr=6.0966e-04) (hash(x)=32553193)
|
| 432 |
+
3580 train 3.857550 (lr=6.0231e-04) (hash(x)=38799856)
|
| 433 |
+
3590 train 3.910937 (lr=5.9504e-04) (hash(x)=40772075)
|
| 434 |
+
3600 val loss 3.9057
|
| 435 |
+
3600 val perplexity 49.6844
|
| 436 |
+
3600 train 3.951592 (lr=5.8785e-04) (hash(x)=39299903)
|
| 437 |
+
3610 train 3.996492 (lr=5.8074e-04) (hash(x)=39503247)
|
| 438 |
+
3620 train 3.884790 (lr=5.7370e-04) (hash(x)=38730668)
|
| 439 |
+
3630 train 3.911533 (lr=5.6675e-04) (hash(x)=40503276)
|
| 440 |
+
3640 train 3.918066 (lr=5.5987e-04) (hash(x)=36504573)
|
| 441 |
+
3650 train 3.937652 (lr=5.5308e-04) (hash(x)=44903075)
|
| 442 |
+
3660 train 3.842866 (lr=5.4636e-04) (hash(x)=33800148)
|
| 443 |
+
3670 train 3.759050 (lr=5.3973e-04) (hash(x)=39404541)
|
| 444 |
+
3680 train 3.697270 (lr=5.3318e-04) (hash(x)=44777432)
|
| 445 |
+
3690 train 3.795732 (lr=5.2671e-04) (hash(x)=45649913)
|
| 446 |
+
3700 val loss 3.9089
|
| 447 |
+
3700 val perplexity 49.8420
|
| 448 |
+
3700 train 3.921203 (lr=5.2033e-04) (hash(x)=33812537)
|
| 449 |
+
3710 train 4.070670 (lr=5.1402e-04) (hash(x)=36161222)
|
| 450 |
+
3720 train 3.949003 (lr=5.0780e-04) (hash(x)=43690616)
|
| 451 |
+
3730 train 3.956997 (lr=5.0167e-04) (hash(x)=35462313)
|
| 452 |
+
3740 train 4.250108 (lr=4.9562e-04) (hash(x)=44674254)
|
| 453 |
+
3750 train 3.938120 (lr=4.8965e-04) (hash(x)=44276297)
|
| 454 |
+
3760 train 3.973969 (lr=4.8377e-04) (hash(x)=38492178)
|
| 455 |
+
3770 train 3.981700 (lr=4.7798e-04) (hash(x)=44347531)
|
| 456 |
+
3780 train 3.895275 (lr=4.7227e-04) (hash(x)=33792059)
|
| 457 |
+
3790 train 3.993541 (lr=4.6665e-04) (hash(x)=38549318)
|
| 458 |
+
3800 val loss 3.8894
|
| 459 |
+
3800 val perplexity 48.8822
|
| 460 |
+
3800 train 3.871576 (lr=4.6112e-04) (hash(x)=41437448)
|
| 461 |
+
3810 train 3.771946 (lr=4.5567e-04) (hash(x)=36656768)
|
| 462 |
+
3820 train 3.930652 (lr=4.5031e-04) (hash(x)=46387893)
|
| 463 |
+
3830 train 3.936323 (lr=4.4504e-04) (hash(x)=44312129)
|
| 464 |
+
3840 train 3.854645 (lr=4.3986e-04) (hash(x)=40341896)
|
| 465 |
+
3850 train 3.903691 (lr=4.3477e-04) (hash(x)=43589160)
|
| 466 |
+
3860 train 3.888968 (lr=4.2977e-04) (hash(x)=40388945)
|
| 467 |
+
3870 train 3.789804 (lr=4.2486e-04) (hash(x)=33336679)
|
| 468 |
+
3880 train 3.877457 (lr=4.2004e-04) (hash(x)=37643311)
|
| 469 |
+
3890 train 3.905501 (lr=4.1530e-04) (hash(x)=38272927)
|
| 470 |
+
3900 val loss 3.8805
|
| 471 |
+
3900 val perplexity 48.4479
|
| 472 |
+
3900 train 3.849802 (lr=4.1066e-04) (hash(x)=41925748)
|
| 473 |
+
3910 train 3.942454 (lr=4.0611e-04) (hash(x)=38377253)
|
| 474 |
+
3920 train 3.940849 (lr=4.0166e-04) (hash(x)=41616611)
|
| 475 |
+
3930 train 3.870841 (lr=3.9729e-04) (hash(x)=32301827)
|
| 476 |
+
3940 train 3.922612 (lr=3.9302e-04) (hash(x)=47697363)
|
| 477 |
+
3950 train 3.876311 (lr=3.8884e-04) (hash(x)=37867767)
|
| 478 |
+
3960 train 3.869706 (lr=3.8475e-04) (hash(x)=37187295)
|
| 479 |
+
3970 train 3.927860 (lr=3.8076e-04) (hash(x)=41952752)
|
| 480 |
+
3980 train 3.902506 (lr=3.7685e-04) (hash(x)=38358660)
|
| 481 |
+
3990 train 3.767617 (lr=3.7305e-04) (hash(x)=40207878)
|
| 482 |
+
4000 val loss 3.8753
|
| 483 |
+
4000 val perplexity 48.1952
|
| 484 |
+
4000 train 3.855196 (lr=3.6933e-04) (hash(x)=39134015)
|
| 485 |
+
4010 train 3.924716 (lr=3.6572e-04) (hash(x)=38313135)
|
| 486 |
+
4020 train 3.855072 (lr=3.6219e-04) (hash(x)=40710513)
|
| 487 |
+
4030 train 3.983963 (lr=3.5876e-04) (hash(x)=35530165)
|
| 488 |
+
4040 train 3.904366 (lr=3.5543e-04) (hash(x)=39494066)
|
| 489 |
+
4050 train 3.897036 (lr=3.5219e-04) (hash(x)=40861777)
|
| 490 |
+
4060 train 3.877889 (lr=3.4905e-04) (hash(x)=36151939)
|
| 491 |
+
4070 train 3.838426 (lr=3.4600e-04) (hash(x)=39008241)
|
| 492 |
+
4080 train 3.929643 (lr=3.4305e-04) (hash(x)=40395746)
|
| 493 |
+
4090 train 3.966282 (lr=3.4019e-04) (hash(x)=38855480)
|
| 494 |
+
4100 val loss 3.8616
|
| 495 |
+
4100 val perplexity 47.5401
|
| 496 |
+
4100 train 3.900739 (lr=3.3744e-04) (hash(x)=45013254)
|
| 497 |
+
4110 train 3.908433 (lr=3.3477e-04) (hash(x)=39934818)
|
| 498 |
+
4120 train 3.907582 (lr=3.3221e-04) (hash(x)=33417705)
|
| 499 |
+
4130 train 3.841565 (lr=3.2974e-04) (hash(x)=41267978)
|
| 500 |
+
4140 train 3.844404 (lr=3.2737e-04) (hash(x)=38961609)
|
| 501 |
+
4150 train 3.843416 (lr=3.2510e-04) (hash(x)=42173878)
|
| 502 |
+
4160 train 3.803875 (lr=3.2292e-04) (hash(x)=36867405)
|
| 503 |
+
4170 train 3.783822 (lr=3.2085e-04) (hash(x)=38991954)
|
| 504 |
+
4180 train 3.850165 (lr=3.1887e-04) (hash(x)=39032592)
|
| 505 |
+
4190 train 3.823670 (lr=3.1699e-04) (hash(x)=61329402)
|
| 506 |
+
4200 val loss 3.8625
|
| 507 |
+
4200 val perplexity 47.5819
|
| 508 |
+
4200 train 3.834893 (lr=3.1520e-04) (hash(x)=40455200)
|
| 509 |
+
4210 train 3.864929 (lr=3.1352e-04) (hash(x)=40913417)
|
| 510 |
+
4220 train 3.900008 (lr=3.1193e-04) (hash(x)=38809704)
|
| 511 |
+
4230 train 3.884638 (lr=3.1044e-04) (hash(x)=39560204)
|
| 512 |
+
4240 train 3.883267 (lr=3.0905e-04) (hash(x)=41496549)
|
| 513 |
+
4250 train 3.922148 (lr=3.0776e-04) (hash(x)=33994768)
|
| 514 |
+
4260 train 3.906499 (lr=3.0657e-04) (hash(x)=39658687)
|
| 515 |
+
4270 train 3.785110 (lr=3.0548e-04) (hash(x)=42583064)
|
| 516 |
+
4280 train 3.851497 (lr=3.0449e-04) (hash(x)=40635811)
|
| 517 |
+
4290 train 3.752874 (lr=3.0359e-04) (hash(x)=35313841)
|
| 518 |
+
4300 val loss 3.8585
|
| 519 |
+
4300 val perplexity 47.3939
|
| 520 |
+
4300 train 3.931639 (lr=3.0280e-04) (hash(x)=39332245)
|
| 521 |
+
4310 train 3.830856 (lr=3.0210e-04) (hash(x)=43732999)
|
| 522 |
+
4320 train 3.834610 (lr=3.0150e-04) (hash(x)=40831835)
|
| 523 |
+
4330 train 3.968698 (lr=3.0101e-04) (hash(x)=41813368)
|
| 524 |
+
4340 train 3.874768 (lr=3.0061e-04) (hash(x)=36952586)
|
| 525 |
+
4350 train 3.895426 (lr=3.0031e-04) (hash(x)=42470745)
|
| 526 |
+
4360 train 3.824138 (lr=3.0011e-04) (hash(x)=39581495)
|
| 527 |
+
4370 train 4.014685 (lr=3.0001e-04) (hash(x)=35914538)
|
| 528 |
+
4374 val loss 3.8519
|
| 529 |
+
4374 val perplexity 47.0802
|
scaling_mha_nh/mha_nh_32_seed_1339/model_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e17b100f13f6a01b5d89bc6ff4c7dc6a42c04b81baadda0db4e2fdf10ad7d0a1
|
| 3 |
+
size 130706626
|
scaling_mha_nh/mha_nh_32_seed_1339/optimizer_04374.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:603c653cab95603af54886d4e0c1912b85e1d3eddc6f419bc7974c9997e155bc
|
| 3 |
+
size 255125126
|