Banaxi-Tech commited on
Commit
dc2ef7f
·
verified ·
1 Parent(s): 2d06c38

Save 5% checkpoint at 1,000,341,504 tokens

Browse files
README.md CHANGED
@@ -1,4 +1,3 @@
1
  5M transformer parameters, partially looped model.
2
 
3
  40M in n-gram embeddings.
4
-
 
1
  5M transformer parameters, partially looped model.
2
 
3
  40M in n-gram embeddings.
 
checkpoint_metadata.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "parameters": 44996036,
3
+ "neural_core_parameters": 4996035,
4
+ "ngram_parameters": 40000001,
5
+ "architecture": {
6
+ "vocab_size": 2048,
7
+ "hidden_size": 256,
8
+ "intermediate_size": 1488,
9
+ "num_hidden_layers": 3,
10
+ "num_attention_heads": 8,
11
+ "num_key_value_heads": 1,
12
+ "head_dim": 32,
13
+ "ngram_buckets": 156250,
14
+ "loop_schedule": [
15
+ 0,
16
+ 1,
17
+ 1,
18
+ 2
19
+ ],
20
+ "refresh_kernel_size": 9,
21
+ "max_position_embeddings": 4096,
22
+ "rope_theta": 100000.0,
23
+ "rms_norm_eps": 1e-05,
24
+ "initializer_range": 0.02,
25
+ "use_cache": false,
26
+ "transformers_version": "5.16.1",
27
+ "architectures": null,
28
+ "output_hidden_states": false,
29
+ "return_dict": true,
30
+ "dtype": null,
31
+ "chunk_size_feed_forward": 0,
32
+ "is_encoder_decoder": false,
33
+ "id2label": {
34
+ "0": "LABEL_0",
35
+ "1": "LABEL_1"
36
+ },
37
+ "label2id": {
38
+ "LABEL_0": 0,
39
+ "LABEL_1": 1
40
+ },
41
+ "problem_type": null,
42
+ "_name_or_path": "",
43
+ "tie_word_embeddings": true,
44
+ "bos_token_id": 1,
45
+ "eos_token_id": 2,
46
+ "pad_token_id": 0,
47
+ "unk_token_id": 3,
48
+ "model_type": "cherry_alpha",
49
+ "output_attentions": false
50
+ },
51
+ "training_percent": 5,
52
+ "step": 1908,
53
+ "total_steps": 38147,
54
+ "tokens_seen": 1000341504,
55
+ "target_tokens": 20000000000,
56
+ "trained_token_target": 20000014336,
57
+ "tokens_per_step": 524288,
58
+ "local_batch": 32,
59
+ "effective_global_batch": 128,
60
+ "world_size": 4,
61
+ "gpu_name": "NVIDIA H200",
62
+ "optimizer": "Muon + AdamW",
63
+ "muon_peak_lr": 0.02,
64
+ "adamw_peak_lr": 0.002,
65
+ "ngram_peak_lr": 0.001,
66
+ "architecture_revision": "2d06c38b56820561813f800da2f896ad64824877",
67
+ "tokenizer_revision": "2d06c38b56820561813f800da2f896ad64824877",
68
+ "dataset_id": "HuggingFaceFW/fineweb-edu",
69
+ "dataset_config": "sample-100BT",
70
+ "dataset_revision": "87f09149ef4734204d70ed1d046ddc9ca3f2b8f9",
71
+ "elapsed_seconds": 422.3712775707245
72
+ }
config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "vocab_size": 2048,
3
+ "hidden_size": 256,
4
+ "intermediate_size": 1488,
5
+ "num_hidden_layers": 3,
6
+ "num_attention_heads": 8,
7
+ "num_key_value_heads": 1,
8
+ "head_dim": 32,
9
+ "ngram_buckets": 156250,
10
+ "loop_schedule": [
11
+ 0,
12
+ 1,
13
+ 1,
14
+ 2
15
+ ],
16
+ "refresh_kernel_size": 9,
17
+ "max_position_embeddings": 4096,
18
+ "rope_theta": 100000.0,
19
+ "rms_norm_eps": 1e-05,
20
+ "initializer_range": 0.02,
21
+ "use_cache": false,
22
+ "transformers_version": "5.16.1",
23
+ "architectures": [
24
+ "CherryAlphaForCausalLM"
25
+ ],
26
+ "output_hidden_states": false,
27
+ "return_dict": true,
28
+ "dtype": null,
29
+ "chunk_size_feed_forward": 0,
30
+ "is_encoder_decoder": false,
31
+ "id2label": {
32
+ "0": "LABEL_0",
33
+ "1": "LABEL_1"
34
+ },
35
+ "label2id": {
36
+ "LABEL_0": 0,
37
+ "LABEL_1": 1
38
+ },
39
+ "problem_type": null,
40
+ "_name_or_path": "Banaxi-Tech/cherry-alpha",
41
+ "tie_word_embeddings": true,
42
+ "bos_token_id": 1,
43
+ "eos_token_id": 2,
44
+ "pad_token_id": 0,
45
+ "unk_token_id": 3,
46
+ "model_type": "cherry_alpha",
47
+ "output_attentions": false,
48
+ "auto_map": {
49
+ "AutoConfig": "configuration_cherry_alpha.CherryAlphaConfig",
50
+ "AutoModelForCausalLM": "modeling_cherry_alpha.CherryAlphaForCausalLM"
51
+ },
52
+ "torch_dtype": "float32"
53
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "5"
7
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab23d112fcc5b2533a3ee1fe3b9e89224cf3985a1e50348edd8c33506203ba85
3
+ size 183136688
training_metrics.jsonl ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"step": 1, "total_steps": 38147, "tokens": 524288, "loss": 7.714707016944885, "ce_loss": 7.708728671073914, "z_loss": 59.78388214111328, "muon_lr": 2e-05, "adamw_lr": 2e-06, "ngram_lr": 1e-06, "grad_norm": 0.35557520389556885, "tokens_per_second": 14162.988352189155, "eta_seconds": 1412095.353796482, "data_wait_seconds": 7.027749240398407}
2
+ {"step": 10, "total_steps": 38147, "tokens": 5242880, "loss": 7.707979202270508, "ce_loss": 7.7020018100738525, "z_loss": 59.7735595703125, "muon_lr": 0.0002, "adamw_lr": 2e-05, "ngram_lr": 1e-05, "grad_norm": 0.34472015500068665, "tokens_per_second": 7246137.722379678, "eta_seconds": 2759.369504425261, "data_wait_seconds": 1.8477439880371094e-05}
3
+ {"step": 20, "total_steps": 38147, "tokens": 10485760, "loss": 7.669735670089722, "ce_loss": 7.663762331008911, "z_loss": 59.7340030670166, "muon_lr": 0.0004, "adamw_lr": 4e-05, "ngram_lr": 2e-05, "grad_norm": 0.36804264783859253, "tokens_per_second": 7551904.648345621, "eta_seconds": 2646.9519289255145, "data_wait_seconds": 1.5437602996826172e-05}
4
+ {"step": 30, "total_steps": 38147, "tokens": 15728640, "loss": 7.607112526893616, "ce_loss": 7.601147294044495, "z_loss": 59.652320861816406, "muon_lr": 0.0006, "adamw_lr": 5.9999999999999995e-05, "ngram_lr": 2.9999999999999997e-05, "grad_norm": 0.45154738426208496, "tokens_per_second": 2129095.2511240807, "eta_seconds": 9386.280714988708, "data_wait_seconds": 1.3232231140136719e-05}
5
+ {"step": 40, "total_steps": 38147, "tokens": 20971520, "loss": 7.497680306434631, "ce_loss": 7.491732835769653, "z_loss": 59.47525501251221, "muon_lr": 0.0008, "adamw_lr": 8e-05, "ngram_lr": 4e-05, "grad_norm": 0.5982406735420227, "tokens_per_second": 7419126.824741767, "eta_seconds": 2692.9102693557743, "data_wait_seconds": 1.2874603271484375e-05}
6
+ {"step": 50, "total_steps": 38147, "tokens": 26214400, "loss": 7.383116602897644, "ce_loss": 7.3772052526474, "z_loss": 59.11469554901123, "muon_lr": 0.001, "adamw_lr": 0.0001, "ngram_lr": 5e-05, "grad_norm": 0.7361624836921692, "tokens_per_second": 7460557.698956279, "eta_seconds": 2677.2529269218444, "data_wait_seconds": 1.519918441772461e-05}
7
+ {"step": 60, "total_steps": 38147, "tokens": 31457280, "loss": 7.178329110145569, "ce_loss": 7.172482132911682, "z_loss": 58.47004413604736, "muon_lr": 0.0012, "adamw_lr": 0.00011999999999999999, "ngram_lr": 5.9999999999999995e-05, "grad_norm": 0.9350511431694031, "tokens_per_second": 2205580.225002911, "eta_seconds": 9053.652562546731, "data_wait_seconds": 1.2874603271484375e-05}
8
+ {"step": 70, "total_steps": 38147, "tokens": 36700160, "loss": 6.938138127326965, "ce_loss": 6.932380676269531, "z_loss": 57.57455253601074, "muon_lr": 0.0014000000000000002, "adamw_lr": 0.00014000000000000001, "ngram_lr": 7.000000000000001e-05, "grad_norm": 1.0273972749710083, "tokens_per_second": 7361676.287022663, "eta_seconds": 2711.7891900777818, "data_wait_seconds": 1.3470649719238281e-05}
9
+ {"step": 80, "total_steps": 38147, "tokens": 41943040, "loss": 6.662042498588562, "ce_loss": 6.656374931335449, "z_loss": 56.67486572265625, "muon_lr": 0.0016, "adamw_lr": 0.00016, "ngram_lr": 8e-05, "grad_norm": 1.019739031791687, "tokens_per_second": 2353892.6813465664, "eta_seconds": 8478.751582074166, "data_wait_seconds": 2.0503997802734375e-05}
10
+ {"step": 90, "total_steps": 38147, "tokens": 47185920, "loss": 6.427657723426819, "ce_loss": 6.42205536365509, "z_loss": 56.02447700500488, "muon_lr": 0.0018, "adamw_lr": 0.00017999999999999998, "ngram_lr": 8.999999999999999e-05, "grad_norm": 0.8781282901763916, "tokens_per_second": 2638210.3543826463, "eta_seconds": 7563.016490650177, "data_wait_seconds": 1.3172626495361328e-05}
11
+ {"step": 100, "total_steps": 38147, "tokens": 52428800, "loss": 6.205592393875122, "ce_loss": 6.200012445449829, "z_loss": 55.79957962036133, "muon_lr": 0.002, "adamw_lr": 0.0002, "ngram_lr": 0.0001, "grad_norm": 0.7427249550819397, "tokens_per_second": 7551533.798686961, "eta_seconds": 2641.5276773929595, "data_wait_seconds": 1.2814998626708984e-05}
12
+ {"step": 110, "total_steps": 38147, "tokens": 57671680, "loss": 6.050146698951721, "ce_loss": 6.044565916061401, "z_loss": 55.80892086029053, "muon_lr": 0.0022, "adamw_lr": 0.00022, "ngram_lr": 0.00011, "grad_norm": 0.6575266122817993, "tokens_per_second": 2207060.043996208, "eta_seconds": 9035.70462899208, "data_wait_seconds": 1.4603137969970703e-05}
13
+ {"step": 120, "total_steps": 38147, "tokens": 62914560, "loss": 5.873514175415039, "ce_loss": 5.867932319641113, "z_loss": 55.81760787963867, "muon_lr": 0.0024, "adamw_lr": 0.00023999999999999998, "ngram_lr": 0.00011999999999999999, "grad_norm": 0.5747721791267395, "tokens_per_second": 2687349.944807866, "eta_seconds": 7418.869959425927, "data_wait_seconds": 1.2993812561035156e-05}
14
+ {"step": 130, "total_steps": 38147, "tokens": 68157440, "loss": 5.708503246307373, "ce_loss": 5.702885389328003, "z_loss": 56.17943096160889, "muon_lr": 0.0026, "adamw_lr": 0.00026000000000000003, "ngram_lr": 0.00013000000000000002, "grad_norm": 0.45857009291648865, "tokens_per_second": 6663422.125151622, "eta_seconds": 2991.2343119859697, "data_wait_seconds": 1.2218952178955078e-05}
15
+ {"step": 140, "total_steps": 38147, "tokens": 73400320, "loss": 5.536954045295715, "ce_loss": 5.53123676776886, "z_loss": 57.1729679107666, "muon_lr": 0.0028000000000000004, "adamw_lr": 0.00028000000000000003, "ngram_lr": 0.00014000000000000001, "grad_norm": 0.35234829783439636, "tokens_per_second": 1530247.7552442267, "eta_seconds": 13021.822085809708, "data_wait_seconds": 1.3530254364013672e-05}
16
+ {"step": 150, "total_steps": 38147, "tokens": 78643200, "loss": 5.3865156173706055, "ce_loss": 5.380668520927429, "z_loss": 58.470916748046875, "muon_lr": 0.003, "adamw_lr": 0.0003, "ngram_lr": 0.00015, "grad_norm": 0.28744885325431824, "tokens_per_second": 7539024.084002833, "eta_seconds": 2642.4336776256564, "data_wait_seconds": 1.3232231140136719e-05}
17
+ {"step": 160, "total_steps": 38147, "tokens": 83886080, "loss": 5.2631731033325195, "ce_loss": 5.2571715116500854, "z_loss": 60.01557445526123, "muon_lr": 0.0032, "adamw_lr": 0.00032, "ngram_lr": 0.00016, "grad_norm": 0.23084913194179535, "tokens_per_second": 2494452.2058846215, "eta_seconds": 7984.169112968445, "data_wait_seconds": 0.6214737296104431}
18
+ {"step": 170, "total_steps": 38147, "tokens": 89128960, "loss": 5.1601094007492065, "ce_loss": 5.153930187225342, "z_loss": 61.79266929626465, "muon_lr": 0.0034, "adamw_lr": 0.00034, "ngram_lr": 0.00017, "grad_norm": 0.2026042491197586, "tokens_per_second": 2445405.9776388393, "eta_seconds": 8142.159444308281, "data_wait_seconds": 1.5437602996826172e-05}
19
+ {"step": 180, "total_steps": 38147, "tokens": 94371840, "loss": 5.056058883666992, "ce_loss": 5.049705266952515, "z_loss": 63.53691291809082, "muon_lr": 0.0036, "adamw_lr": 0.00035999999999999997, "ngram_lr": 0.00017999999999999998, "grad_norm": 0.2568941116333008, "tokens_per_second": 7539223.106709906, "eta_seconds": 2640.2776803731917, "data_wait_seconds": 1.3768672943115234e-05}
20
+ {"step": 190, "total_steps": 38147, "tokens": 99614720, "loss": 4.928068041801453, "ce_loss": 4.921524405479431, "z_loss": 65.43530082702637, "muon_lr": 0.0038000000000000004, "adamw_lr": 0.00038, "ngram_lr": 0.00019, "grad_norm": 0.3614652156829834, "tokens_per_second": 2216051.394467085, "eta_seconds": 8980.116465568542, "data_wait_seconds": 1.615285873413086e-05}
21
+ {"step": 200, "total_steps": 38147, "tokens": 104857600, "loss": 4.847650766372681, "ce_loss": 4.8409271240234375, "z_loss": 67.23649215698242, "muon_lr": 0.004, "adamw_lr": 0.0004, "ngram_lr": 0.0002, "grad_norm": 0.7364533543586731, "tokens_per_second": 2949841.3629276603, "eta_seconds": 6744.483613944054, "data_wait_seconds": 1.3768672943115234e-05}
22
+ {"step": 210, "total_steps": 38147, "tokens": 110100480, "loss": 4.786519527435303, "ce_loss": 4.779638051986694, "z_loss": 68.8156681060791, "muon_lr": 0.004200000000000001, "adamw_lr": 0.00041999999999999996, "ngram_lr": 0.00020999999999999998, "grad_norm": 0.8606784343719482, "tokens_per_second": 7014842.201380305, "eta_seconds": 2835.404316306114, "data_wait_seconds": 1.5735626220703125e-05}
23
+ {"step": 220, "total_steps": 38147, "tokens": 115343360, "loss": 4.655158758163452, "ce_loss": 4.6481239795684814, "z_loss": 70.34732055664062, "muon_lr": 0.0044, "adamw_lr": 0.00044, "ngram_lr": 0.00022, "grad_norm": 1.1180042028427124, "tokens_per_second": 1507641.6918316989, "eta_seconds": 13189.25516834259, "data_wait_seconds": 1.6570091247558594e-05}
24
+ {"step": 230, "total_steps": 38147, "tokens": 120586240, "loss": 4.5521851778030396, "ce_loss": 4.544991970062256, "z_loss": 71.93116569519043, "muon_lr": 0.004600000000000001, "adamw_lr": 0.00046, "ngram_lr": 0.00023, "grad_norm": 1.2257267236709595, "tokens_per_second": 7543091.919962432, "eta_seconds": 2635.448209691048, "data_wait_seconds": 1.3709068298339844e-05}
25
+ {"step": 240, "total_steps": 38147, "tokens": 125829120, "loss": 4.478403449058533, "ce_loss": 4.471078157424927, "z_loss": 73.25160026550293, "muon_lr": 0.0048, "adamw_lr": 0.00047999999999999996, "ngram_lr": 0.00023999999999999998, "grad_norm": 1.0167913436889648, "tokens_per_second": 2604596.425133865, "eta_seconds": 7630.42789440155, "data_wait_seconds": 1.9788742065429688e-05}
26
+ {"step": 250, "total_steps": 38147, "tokens": 131072000, "loss": 4.410106658935547, "ce_loss": 4.402639150619507, "z_loss": 74.67479133605957, "muon_lr": 0.005, "adamw_lr": 0.0005, "ngram_lr": 0.00025, "grad_norm": 0.9511222243309021, "tokens_per_second": 2588768.578017055, "eta_seconds": 7675.055431652068, "data_wait_seconds": 1.3589859008789062e-05}
27
+ {"step": 260, "total_steps": 38147, "tokens": 136314880, "loss": 4.349528193473816, "ce_loss": 4.3419681787490845, "z_loss": 75.60078048706055, "muon_lr": 0.0052, "adamw_lr": 0.0005200000000000001, "ngram_lr": 0.00026000000000000003, "grad_norm": 0.8766444325447083, "tokens_per_second": 7540601.045564089, "eta_seconds": 2634.2329127311705, "data_wait_seconds": 1.4007091522216797e-05}
28
+ {"step": 270, "total_steps": 38147, "tokens": 141557760, "loss": 4.287362098693848, "ce_loss": 4.279715657234192, "z_loss": 76.46497917175293, "muon_lr": 0.0054, "adamw_lr": 0.00054, "ngram_lr": 0.00027, "grad_norm": 0.8805499076843262, "tokens_per_second": 1556964.3294337417, "eta_seconds": 12754.599575972557, "data_wait_seconds": 0.32283496856689453}
29
+ {"step": 280, "total_steps": 38147, "tokens": 146800640, "loss": 4.193420052528381, "ce_loss": 4.185634732246399, "z_loss": 77.8543815612793, "muon_lr": 0.005600000000000001, "adamw_lr": 0.0005600000000000001, "ngram_lr": 0.00028000000000000003, "grad_norm": 0.8973211050033569, "tokens_per_second": 6226640.097926259, "eta_seconds": 3188.431221938133, "data_wait_seconds": 1.3887882232666016e-05}
30
+ {"step": 290, "total_steps": 38147, "tokens": 152043520, "loss": 4.199685096740723, "ce_loss": 4.191836476325989, "z_loss": 78.48650932312012, "muon_lr": 0.0058, "adamw_lr": 0.00058, "ngram_lr": 0.00029, "grad_norm": 0.8038012385368347, "tokens_per_second": 2554586.730615463, "eta_seconds": 7769.542751526832, "data_wait_seconds": 0.3058410882949829}
31
+ {"step": 300, "total_steps": 38147, "tokens": 157286400, "loss": 4.135829210281372, "ce_loss": 4.127879738807678, "z_loss": 79.49372673034668, "muon_lr": 0.006, "adamw_lr": 0.0006, "ngram_lr": 0.0003, "grad_norm": 0.7599166035652161, "tokens_per_second": 2332856.182000322, "eta_seconds": 8505.765631461143, "data_wait_seconds": 1.5616416931152344e-05}
32
+ {"step": 310, "total_steps": 38147, "tokens": 162529280, "loss": 4.115949630737305, "ce_loss": 4.107944846153259, "z_loss": 80.0470962524414, "muon_lr": 0.0062, "adamw_lr": 0.00062, "ngram_lr": 0.00031, "grad_norm": 0.864853024482727, "tokens_per_second": 7541004.440030617, "eta_seconds": 2630.615750694275, "data_wait_seconds": 1.430511474609375e-05}
33
+ {"step": 320, "total_steps": 38147, "tokens": 167772160, "loss": 3.997064709663391, "ce_loss": 3.9888452887535095, "z_loss": 82.19391441345215, "muon_lr": 0.0064, "adamw_lr": 0.00064, "ngram_lr": 0.00032, "grad_norm": 0.9067529439926147, "tokens_per_second": 2442902.0787132718, "eta_seconds": 8118.312374782563, "data_wait_seconds": 1.3232231140136719e-05}
34
+ {"step": 330, "total_steps": 38147, "tokens": 173015040, "loss": 3.9806066751480103, "ce_loss": 3.9723057746887207, "z_loss": 83.00942039489746, "muon_lr": 0.006600000000000001, "adamw_lr": 0.00066, "ngram_lr": 0.00033, "grad_norm": 0.7168138027191162, "tokens_per_second": 1590830.3565939835, "eta_seconds": 12463.302082347871, "data_wait_seconds": 1.519918441772461e-05}
35
+ {"step": 340, "total_steps": 38147, "tokens": 178257920, "loss": 3.9310680627822876, "ce_loss": 3.922678768634796, "z_loss": 83.89345359802246, "muon_lr": 0.0068, "adamw_lr": 0.00068, "ngram_lr": 0.00034, "grad_norm": 0.6740385890007019, "tokens_per_second": 6959418.869516232, "eta_seconds": 2848.1913199424744, "data_wait_seconds": 1.3649463653564453e-05}
36
+ {"step": 350, "total_steps": 38147, "tokens": 183500800, "loss": 3.908783257007599, "ce_loss": 3.9003064036369324, "z_loss": 84.76858329772949, "muon_lr": 0.007, "adamw_lr": 0.0007000000000000001, "ngram_lr": 0.00035000000000000005, "grad_norm": 0.7061223387718201, "tokens_per_second": 1474547.0362467857, "eta_seconds": 13439.051484203339, "data_wait_seconds": 1.3887882232666016e-05}
37
+ {"step": 360, "total_steps": 38147, "tokens": 188743680, "loss": 3.90825217962265, "ce_loss": 3.899712562561035, "z_loss": 85.39625930786133, "muon_lr": 0.0072, "adamw_lr": 0.0007199999999999999, "ngram_lr": 0.00035999999999999997, "grad_norm": 0.6888197660446167, "tokens_per_second": 2562862.608651872, "eta_seconds": 7730.133714199065, "data_wait_seconds": 1.5079975128173828e-05}
38
+ {"step": 370, "total_steps": 38147, "tokens": 193986560, "loss": 3.8621427416801453, "ce_loss": 3.853570282459259, "z_loss": 85.72406005859375, "muon_lr": 0.0074, "adamw_lr": 0.00074, "ngram_lr": 0.00037, "grad_norm": 0.6649040579795837, "tokens_per_second": 2347827.1303791525, "eta_seconds": 8435.896970319747, "data_wait_seconds": 2.110004425048828e-05}
39
+ {"step": 380, "total_steps": 38147, "tokens": 199229440, "loss": 3.8472890853881836, "ce_loss": 3.8385685682296753, "z_loss": 87.20517158508301, "muon_lr": 0.007600000000000001, "adamw_lr": 0.00076, "ngram_lr": 0.00038, "grad_norm": 0.6724202632904053, "tokens_per_second": 1641361.985795614, "eta_seconds": 12063.63073310852, "data_wait_seconds": 0.30100560188293457}
40
+ {"step": 390, "total_steps": 38147, "tokens": 204472320, "loss": 3.8528982400894165, "ce_loss": 3.8440893292427063, "z_loss": 88.08908462524414, "muon_lr": 0.0078, "adamw_lr": 0.00078, "ngram_lr": 0.00039, "grad_norm": 0.6090479493141174, "tokens_per_second": 7533755.098855937, "eta_seconds": 2627.579707098007, "data_wait_seconds": 1.4185905456542969e-05}
41
+ {"step": 400, "total_steps": 38147, "tokens": 209715200, "loss": 3.8002134561538696, "ce_loss": 3.7913466095924377, "z_loss": 88.66859817504883, "muon_lr": 0.008, "adamw_lr": 0.0008, "ngram_lr": 0.0004, "grad_norm": 0.6068047881126404, "tokens_per_second": 1520347.7664032138, "eta_seconds": 13016.955444884301, "data_wait_seconds": 1.5139579772949219e-05}
42
+ {"step": 410, "total_steps": 38147, "tokens": 214958080, "loss": 3.7618911266326904, "ce_loss": 3.7529407739639282, "z_loss": 89.50360488891602, "muon_lr": 0.008199999999999999, "adamw_lr": 0.0008200000000000001, "ngram_lr": 0.00041000000000000005, "grad_norm": 0.6481754183769226, "tokens_per_second": 2602001.159767926, "eta_seconds": 7603.784564709664, "data_wait_seconds": 1.3828277587890625e-05}
43
+ {"step": 420, "total_steps": 38147, "tokens": 220200960, "loss": 3.780871868133545, "ce_loss": 3.7718703746795654, "z_loss": 90.01469421386719, "muon_lr": 0.008400000000000001, "adamw_lr": 0.0008399999999999999, "ngram_lr": 0.00041999999999999996, "grad_norm": 0.6722284555435181, "tokens_per_second": 6804740.231408232, "eta_seconds": 2906.7697962522507, "data_wait_seconds": 1.4901161193847656e-05}
44
+ {"step": 430, "total_steps": 38147, "tokens": 225443840, "loss": 3.7115601301193237, "ce_loss": 3.7024566531181335, "z_loss": 91.03459930419922, "muon_lr": 0.0086, "adamw_lr": 0.00086, "ngram_lr": 0.00043, "grad_norm": 0.5772982835769653, "tokens_per_second": 2380381.679234891, "eta_seconds": 8307.31082687378, "data_wait_seconds": 1.3053417205810547e-05}
45
+ {"step": 440, "total_steps": 38147, "tokens": 230686720, "loss": 3.7086883783340454, "ce_loss": 3.6994104981422424, "z_loss": 92.77891731262207, "muon_lr": 0.0088, "adamw_lr": 0.00088, "ngram_lr": 0.00044, "grad_norm": 0.6159934401512146, "tokens_per_second": 2792574.6689005336, "eta_seconds": 7079.247633433341, "data_wait_seconds": 1.704692840576172e-05}
46
+ {"step": 450, "total_steps": 38147, "tokens": 235929600, "loss": 3.644501805305481, "ce_loss": 3.635166585445404, "z_loss": 93.35165023803711, "muon_lr": 0.009, "adamw_lr": 0.0009, "ngram_lr": 0.00045, "grad_norm": 0.6296858787536621, "tokens_per_second": 2388646.949062072, "eta_seconds": 8274.175781297683, "data_wait_seconds": 1.4722347259521484e-05}
47
+ {"step": 460, "total_steps": 38147, "tokens": 241172480, "loss": 3.6783171892166138, "ce_loss": 3.668973684310913, "z_loss": 93.43500709533691, "muon_lr": 0.009200000000000002, "adamw_lr": 0.00092, "ngram_lr": 0.00046, "grad_norm": 0.591005802154541, "tokens_per_second": 2627867.9105253303, "eta_seconds": 7518.9631019353865, "data_wait_seconds": 1.424551010131836e-05}
48
+ {"step": 470, "total_steps": 38147, "tokens": 246415360, "loss": 3.612688660621643, "ce_loss": 3.603201448917389, "z_loss": 94.8715705871582, "muon_lr": 0.0094, "adamw_lr": 0.0009400000000000001, "ngram_lr": 0.00047000000000000004, "grad_norm": 0.6261810064315796, "tokens_per_second": 6894123.051277123, "eta_seconds": 2865.280881857872, "data_wait_seconds": 1.239776611328125e-05}
49
+ {"step": 480, "total_steps": 38147, "tokens": 251658240, "loss": 3.591165840625763, "ce_loss": 3.581607162952423, "z_loss": 95.58747482299805, "muon_lr": 0.0096, "adamw_lr": 0.0009599999999999999, "ngram_lr": 0.00047999999999999996, "grad_norm": 0.5834220051765442, "tokens_per_second": 2382430.8853730257, "eta_seconds": 8289.162223863603, "data_wait_seconds": 1.436471939086914e-05}
50
+ {"step": 490, "total_steps": 38147, "tokens": 256901120, "loss": 3.568116545677185, "ce_loss": 3.558379888534546, "z_loss": 97.36702728271484, "muon_lr": 0.009800000000000001, "adamw_lr": 0.00098, "ngram_lr": 0.00049, "grad_norm": 0.5412684082984924, "tokens_per_second": 2619433.10010262, "eta_seconds": 7537.170243144035, "data_wait_seconds": 1.3768672943115234e-05}
51
+ {"step": 500, "total_steps": 38147, "tokens": 262144000, "loss": 3.566168487071991, "ce_loss": 3.5563835501670837, "z_loss": 97.84905433654785, "muon_lr": 0.01, "adamw_lr": 0.001, "ngram_lr": 0.0005, "grad_norm": 0.6467755436897278, "tokens_per_second": 2408860.0150029096, "eta_seconds": 8193.863575744628, "data_wait_seconds": 1.424551010131836e-05}
52
+ {"step": 510, "total_steps": 38147, "tokens": 267386880, "loss": 3.532553493976593, "ce_loss": 3.522671163082123, "z_loss": 98.82350540161133, "muon_lr": 0.0102, "adamw_lr": 0.00102, "ngram_lr": 0.00051, "grad_norm": 0.6117472052574158, "tokens_per_second": 1786655.051136468, "eta_seconds": 11044.452841329574, "data_wait_seconds": 0.30620819330215454}
53
+ {"step": 520, "total_steps": 38147, "tokens": 272629760, "loss": 3.5175934433937073, "ce_loss": 3.507621645927429, "z_loss": 99.7182445526123, "muon_lr": 0.0104, "adamw_lr": 0.0010400000000000001, "ngram_lr": 0.0005200000000000001, "grad_norm": 0.598330020904541, "tokens_per_second": 2796972.6564179924, "eta_seconds": 7053.1202837228775, "data_wait_seconds": 0.2776179909706116}
54
+ {"step": 530, "total_steps": 38147, "tokens": 277872640, "loss": 3.5138492584228516, "ce_loss": 3.5040327310562134, "z_loss": 98.16528701782227, "muon_lr": 0.0106, "adamw_lr": 0.00106, "ngram_lr": 0.00053, "grad_norm": 0.5971319079399109, "tokens_per_second": 2577936.63096905, "eta_seconds": 7650.359383964539, "data_wait_seconds": 1.436471939086914e-05}
55
+ {"step": 540, "total_steps": 38147, "tokens": 283115520, "loss": 3.5122064352035522, "ce_loss": 3.50213623046875, "z_loss": 100.70252990722656, "muon_lr": 0.0108, "adamw_lr": 0.00108, "ngram_lr": 0.00054, "grad_norm": 0.5612853765487671, "tokens_per_second": 2692535.700885099, "eta_seconds": 7322.799400401116, "data_wait_seconds": 1.4424324035644531e-05}
56
+ {"step": 550, "total_steps": 38147, "tokens": 288358400, "loss": 3.4556495547294617, "ce_loss": 3.445426285266876, "z_loss": 102.23232078552246, "muon_lr": 0.011, "adamw_lr": 0.0011, "ngram_lr": 0.00055, "grad_norm": 0.5357598662376404, "tokens_per_second": 1584479.0822886438, "eta_seconds": 12440.46460211277, "data_wait_seconds": 1.4185905456542969e-05}
57
+ {"step": 560, "total_steps": 38147, "tokens": 293601280, "loss": 3.4608590602874756, "ce_loss": 3.450589895248413, "z_loss": 102.69148254394531, "muon_lr": 0.011200000000000002, "adamw_lr": 0.0011200000000000001, "ngram_lr": 0.0005600000000000001, "grad_norm": 0.6118342280387878, "tokens_per_second": 2643159.72827788, "eta_seconds": 7455.627007770538, "data_wait_seconds": 1.7642974853515625e-05}
58
+ {"step": 570, "total_steps": 38147, "tokens": 298844160, "loss": 3.409735918045044, "ce_loss": 3.399404287338257, "z_loss": 103.31617546081543, "muon_lr": 0.0114, "adamw_lr": 0.0011400000000000002, "ngram_lr": 0.0005700000000000001, "grad_norm": 0.5069886445999146, "tokens_per_second": 1733275.3549153127, "eta_seconds": 11366.439913964272, "data_wait_seconds": 1.519918441772461e-05}
59
+ {"step": 580, "total_steps": 38147, "tokens": 304087040, "loss": 3.4123098850250244, "ce_loss": 3.4019526839256287, "z_loss": 103.57165145874023, "muon_lr": 0.0116, "adamw_lr": 0.00116, "ngram_lr": 0.00058, "grad_norm": 0.6694517135620117, "tokens_per_second": 1577392.5636317662, "eta_seconds": 12486.382749676704, "data_wait_seconds": 2.086162567138672e-05}
60
+ {"step": 590, "total_steps": 38147, "tokens": 309329920, "loss": 3.4019775986671448, "ce_loss": 3.391569674015045, "z_loss": 104.07889938354492, "muon_lr": 0.011800000000000001, "adamw_lr": 0.0011799999999999998, "ngram_lr": 0.0005899999999999999, "grad_norm": 0.53458172082901, "tokens_per_second": 2345532.98612417, "eta_seconds": 8394.972286677359, "data_wait_seconds": 1.7881393432617188e-05}
61
+ {"step": 600, "total_steps": 38147, "tokens": 314572800, "loss": 3.3456825613975525, "ce_loss": 3.3351190090179443, "z_loss": 105.63541984558105, "muon_lr": 0.012, "adamw_lr": 0.0012, "ngram_lr": 0.0006, "grad_norm": 0.5371636152267456, "tokens_per_second": 2549665.1968491157, "eta_seconds": 7720.79469897747, "data_wait_seconds": 1.3887882232666016e-05}
62
+ {"step": 610, "total_steps": 38147, "tokens": 319815680, "loss": 3.378806948661804, "ce_loss": 3.3683334589004517, "z_loss": 104.73417472839355, "muon_lr": 0.0122, "adamw_lr": 0.00122, "ngram_lr": 0.00061, "grad_norm": 0.6403800249099731, "tokens_per_second": 2484238.1395004145, "eta_seconds": 7922.025808668137, "data_wait_seconds": 1.6570091247558594e-05}
63
+ {"step": 620, "total_steps": 38147, "tokens": 325058560, "loss": 3.3543894290924072, "ce_loss": 3.3437212109565735, "z_loss": 106.68229293823242, "muon_lr": 0.0124, "adamw_lr": 0.00124, "ngram_lr": 0.00062, "grad_norm": 0.5302276611328125, "tokens_per_second": 2331266.442651187, "eta_seconds": 8439.599788355827, "data_wait_seconds": 1.4543533325195312e-05}
64
+ {"step": 630, "total_steps": 38147, "tokens": 330301440, "loss": 3.308763027191162, "ce_loss": 3.2979326248168945, "z_loss": 108.30423736572266, "muon_lr": 0.0126, "adamw_lr": 0.00126, "ngram_lr": 0.00063, "grad_norm": 0.5488021373748779, "tokens_per_second": 2881010.2367827357, "eta_seconds": 6827.366541385651, "data_wait_seconds": 1.436471939086914e-05}
65
+ {"step": 640, "total_steps": 38147, "tokens": 335544320, "loss": 3.259060263633728, "ce_loss": 3.2481698989868164, "z_loss": 108.90389060974121, "muon_lr": 0.0128, "adamw_lr": 0.00128, "ngram_lr": 0.00064, "grad_norm": 0.5964745879173279, "tokens_per_second": 1576400.4209367563, "eta_seconds": 12474.286199641227, "data_wait_seconds": 0.29108285903930664}
66
+ {"step": 650, "total_steps": 38147, "tokens": 340787200, "loss": 3.2980770468711853, "ce_loss": 3.2872347831726074, "z_loss": 108.42268180847168, "muon_lr": 0.013, "adamw_lr": 0.0013, "ngram_lr": 0.00065, "grad_norm": 0.6500165462493896, "tokens_per_second": 2435903.6836278653, "eta_seconds": 8070.60938744545, "data_wait_seconds": 1.3768672943115234e-05}
67
+ {"step": 660, "total_steps": 38147, "tokens": 346030080, "loss": 3.295556128025055, "ce_loss": 3.284740626811981, "z_loss": 108.15526008605957, "muon_lr": 0.013200000000000002, "adamw_lr": 0.00132, "ngram_lr": 0.00066, "grad_norm": 0.4578292965888977, "tokens_per_second": 6281518.321906842, "eta_seconds": 3128.8588600397106, "data_wait_seconds": 1.519918441772461e-05}
68
+ {"step": 670, "total_steps": 38147, "tokens": 351272960, "loss": 3.261557459831238, "ce_loss": 3.2503941655158997, "z_loss": 111.6330795288086, "muon_lr": 0.0134, "adamw_lr": 0.00134, "ngram_lr": 0.00067, "grad_norm": 0.5622156262397766, "tokens_per_second": 2550623.0756593333, "eta_seconds": 7703.506473970413, "data_wait_seconds": 1.4603137969970703e-05}
69
+ {"step": 680, "total_steps": 38147, "tokens": 356515840, "loss": 3.223462998867035, "ce_loss": 3.21236652135849, "z_loss": 110.96495246887207, "muon_lr": 0.0136, "adamw_lr": 0.00136, "ngram_lr": 0.00068, "grad_norm": 0.5643494725227356, "tokens_per_second": 2659415.398652916, "eta_seconds": 7386.397215700149, "data_wait_seconds": 2.1636486053466797e-05}
70
+ {"step": 690, "total_steps": 38147, "tokens": 361758720, "loss": 3.1894708275794983, "ce_loss": 3.1781685948371887, "z_loss": 113.0228099822998, "muon_lr": 0.013800000000000002, "adamw_lr": 0.0013800000000000002, "ngram_lr": 0.0006900000000000001, "grad_norm": 0.480710506439209, "tokens_per_second": 2703578.6144791762, "eta_seconds": 7263.800472021103, "data_wait_seconds": 2.0563602447509766e-05}
71
+ {"step": 700, "total_steps": 38147, "tokens": 367001600, "loss": 3.2279842495918274, "ce_loss": 3.216698944568634, "z_loss": 112.8526611328125, "muon_lr": 0.014, "adamw_lr": 0.0014000000000000002, "ngram_lr": 0.0007000000000000001, "grad_norm": 0.5350193381309509, "tokens_per_second": 2349470.419104123, "eta_seconds": 8356.356639504433, "data_wait_seconds": 1.5139579772949219e-05}
72
+ {"step": 710, "total_steps": 38147, "tokens": 372244480, "loss": 3.1822314262390137, "ce_loss": 3.1709017157554626, "z_loss": 113.29736518859863, "muon_lr": 0.0142, "adamw_lr": 0.0014199999999999998, "ngram_lr": 0.0007099999999999999, "grad_norm": 0.4777548611164093, "tokens_per_second": 2840580.231207093, "eta_seconds": 6909.774855279922, "data_wait_seconds": 0.2696620225906372}
73
+ {"step": 720, "total_steps": 38147, "tokens": 377487360, "loss": 3.162253737449646, "ce_loss": 3.150775671005249, "z_loss": 114.78024482727051, "muon_lr": 0.0144, "adamw_lr": 0.0014399999999999999, "ngram_lr": 0.0007199999999999999, "grad_norm": 0.4700380861759186, "tokens_per_second": 2522006.835096101, "eta_seconds": 7780.520933938026, "data_wait_seconds": 0.2681888937950134}
74
+ {"step": 730, "total_steps": 38147, "tokens": 382730240, "loss": 3.1321462392807007, "ce_loss": 3.120709776878357, "z_loss": 114.36466789245605, "muon_lr": 0.0146, "adamw_lr": 0.00146, "ngram_lr": 0.00073, "grad_norm": 0.49616074562072754, "tokens_per_second": 2575030.7976421914, "eta_seconds": 7618.2716392993925, "data_wait_seconds": 1.3768672943115234e-05}
75
+ {"step": 740, "total_steps": 38147, "tokens": 387973120, "loss": 3.1368382573127747, "ce_loss": 3.1254504919052124, "z_loss": 113.87773323059082, "muon_lr": 0.0148, "adamw_lr": 0.00148, "ngram_lr": 0.00074, "grad_norm": 0.46780115365982056, "tokens_per_second": 1676152.6366288147, "eta_seconds": 11700.629636836053, "data_wait_seconds": 1.5616416931152344e-05}
76
+ {"step": 750, "total_steps": 38147, "tokens": 393216000, "loss": 3.1099137663841248, "ce_loss": 3.098430573940277, "z_loss": 114.8317928314209, "muon_lr": 0.015, "adamw_lr": 0.0015, "ngram_lr": 0.00075, "grad_norm": 0.5300443768501282, "tokens_per_second": 1642876.1605366643, "eta_seconds": 11934.43474740982, "data_wait_seconds": 0.3087407350540161}
77
+ {"step": 760, "total_steps": 38147, "tokens": 398458880, "loss": 3.191687226295471, "ce_loss": 3.180337905883789, "z_loss": 113.49361038208008, "muon_lr": 0.015200000000000002, "adamw_lr": 0.00152, "ngram_lr": 0.00076, "grad_norm": 0.44815412163734436, "tokens_per_second": 2737917.067952167, "eta_seconds": 7159.294810438156, "data_wait_seconds": 1.3589859008789062e-05}
78
+ {"step": 770, "total_steps": 38147, "tokens": 403701760, "loss": 3.1594170928001404, "ce_loss": 3.148090422153473, "z_loss": 113.26649856567383, "muon_lr": 0.0154, "adamw_lr": 0.0015400000000000001, "ngram_lr": 0.0007700000000000001, "grad_norm": 0.42371243238449097, "tokens_per_second": 2586341.620891021, "eta_seconds": 7576.846158957482, "data_wait_seconds": 1.3709068298339844e-05}
79
+ {"step": 780, "total_steps": 38147, "tokens": 408944640, "loss": 3.0839158296585083, "ce_loss": 3.07234650850296, "z_loss": 115.6932315826416, "muon_lr": 0.0156, "adamw_lr": 0.00156, "ngram_lr": 0.00078, "grad_norm": 0.4582517147064209, "tokens_per_second": 1689025.5405396791, "eta_seconds": 11599.036974740027, "data_wait_seconds": 0.2677650451660156}
80
+ {"step": 790, "total_steps": 38147, "tokens": 414187520, "loss": 3.108667731285095, "ce_loss": 3.0971415638923645, "z_loss": 115.26179504394531, "muon_lr": 0.0158, "adamw_lr": 0.00158, "ngram_lr": 0.00079, "grad_norm": 0.45045292377471924, "tokens_per_second": 2562223.272678378, "eta_seconds": 7644.074981617928, "data_wait_seconds": 1.990795135498047e-05}
81
+ {"step": 800, "total_steps": 38147, "tokens": 419430400, "loss": 3.2435104846954346, "ce_loss": 3.232246160507202, "z_loss": 112.64291381835938, "muon_lr": 0.016, "adamw_lr": 0.0016, "ngram_lr": 0.0008, "grad_norm": 0.4701556861400604, "tokens_per_second": 6876225.9754491, "eta_seconds": 2847.5771456480024, "data_wait_seconds": 1.3828277587890625e-05}
82
+ {"step": 810, "total_steps": 38147, "tokens": 424673280, "loss": 3.1093117594718933, "ce_loss": 3.097711145877838, "z_loss": 116.00627326965332, "muon_lr": 0.0162, "adamw_lr": 0.0016200000000000001, "ngram_lr": 0.0008100000000000001, "grad_norm": 0.44234150648117065, "tokens_per_second": 2671352.261571577, "eta_seconds": 7327.877097153663, "data_wait_seconds": 1.3768672943115234e-05}
83
+ {"step": 820, "total_steps": 38147, "tokens": 429916160, "loss": 3.072600245475769, "ce_loss": 3.0611568689346313, "z_loss": 114.43404769897461, "muon_lr": 0.016399999999999998, "adamw_lr": 0.0016400000000000002, "ngram_lr": 0.0008200000000000001, "grad_norm": 0.4286942481994629, "tokens_per_second": 2623675.3419863195, "eta_seconds": 7459.039562869072, "data_wait_seconds": 1.430511474609375e-05}
84
+ {"step": 830, "total_steps": 38147, "tokens": 435159040, "loss": 3.040897011756897, "ce_loss": 3.0293624997138977, "z_loss": 115.34466171264648, "muon_lr": 0.0166, "adamw_lr": 0.0016600000000000002, "ngram_lr": 0.0008300000000000001, "grad_norm": 0.36852356791496277, "tokens_per_second": 2638921.7467197133, "eta_seconds": 7413.958113884926, "data_wait_seconds": 1.4781951904296875e-05}
85
+ {"step": 840, "total_steps": 38147, "tokens": 440401920, "loss": 3.0382885932922363, "ce_loss": 3.0266162157058716, "z_loss": 116.72443389892578, "muon_lr": 0.016800000000000002, "adamw_lr": 0.0016799999999999999, "ngram_lr": 0.0008399999999999999, "grad_norm": 0.48568201065063477, "tokens_per_second": 1994549.7898416177, "eta_seconds": 9806.53003280163, "data_wait_seconds": 0.21687018871307373}
86
+ {"step": 850, "total_steps": 38147, "tokens": 445644800, "loss": 3.039895474910736, "ce_loss": 3.0284404158592224, "z_loss": 114.55129432678223, "muon_lr": 0.017, "adamw_lr": 0.0017, "ngram_lr": 0.00085, "grad_norm": 0.3958066999912262, "tokens_per_second": 2872180.885076886, "eta_seconds": 6808.195694637298, "data_wait_seconds": 1.5497207641601562e-05}
87
+ {"step": 860, "total_steps": 38147, "tokens": 450887680, "loss": 3.0325602889060974, "ce_loss": 3.0209288001060486, "z_loss": 116.31522560119629, "muon_lr": 0.0172, "adamw_lr": 0.00172, "ngram_lr": 0.00086, "grad_norm": 0.38858407735824585, "tokens_per_second": 1710538.1927495573, "eta_seconds": 11428.640844655036, "data_wait_seconds": 1.5079975128173828e-05}
88
+ {"step": 870, "total_steps": 38147, "tokens": 456130560, "loss": 3.0295189023017883, "ce_loss": 3.0178542733192444, "z_loss": 116.64595985412598, "muon_lr": 0.017400000000000002, "adamw_lr": 0.00174, "ngram_lr": 0.00087, "grad_norm": 0.41391289234161377, "tokens_per_second": 1643875.2431881751, "eta_seconds": 11888.909366440774, "data_wait_seconds": 0.2788180112838745}
89
+ {"step": 880, "total_steps": 38147, "tokens": 461373440, "loss": 2.9740660190582275, "ce_loss": 2.9625149965286255, "z_loss": 115.50982093811035, "muon_lr": 0.0176, "adamw_lr": 0.00176, "ngram_lr": 0.00088, "grad_norm": 0.44681233167648315, "tokens_per_second": 2808658.6179272267, "eta_seconds": 6956.5737791299825, "data_wait_seconds": 1.8358230590820312e-05}
90
+ {"step": 890, "total_steps": 38147, "tokens": 466616320, "loss": 2.9605724811553955, "ce_loss": 2.9489070177078247, "z_loss": 116.6550064086914, "muon_lr": 0.0178, "adamw_lr": 0.0017800000000000001, "ngram_lr": 0.0008900000000000001, "grad_norm": 0.442884236574173, "tokens_per_second": 6392384.96096733, "eta_seconds": 3055.729299044609, "data_wait_seconds": 1.4066696166992188e-05}
91
+ {"step": 900, "total_steps": 38147, "tokens": 471859200, "loss": 3.0077711939811707, "ce_loss": 2.9963501691818237, "z_loss": 114.21093368530273, "muon_lr": 0.018, "adamw_lr": 0.0018, "ngram_lr": 0.0009, "grad_norm": 0.37169408798217773, "tokens_per_second": 2476922.8535696766, "eta_seconds": 7884.03849875927, "data_wait_seconds": 1.3589859008789062e-05}
92
+ {"step": 910, "total_steps": 38147, "tokens": 477102080, "loss": 3.0009095668792725, "ce_loss": 2.989498198032379, "z_loss": 114.11362838745117, "muon_lr": 0.0182, "adamw_lr": 0.00182, "ngram_lr": 0.00091, "grad_norm": 0.4474586248397827, "tokens_per_second": 2858783.6342958887, "eta_seconds": 6829.097530078888, "data_wait_seconds": 1.4960765838623047e-05}
93
+ {"step": 920, "total_steps": 38147, "tokens": 482344960, "loss": 3.0015029907226562, "ce_loss": 2.990237534046173, "z_loss": 112.65457344055176, "muon_lr": 0.018400000000000003, "adamw_lr": 0.00184, "ngram_lr": 0.00092, "grad_norm": 0.3788028955459595, "tokens_per_second": 2497477.2856354667, "eta_seconds": 7814.953708791732, "data_wait_seconds": 1.4066696166992188e-05}
94
+ {"step": 930, "total_steps": 38147, "tokens": 487587840, "loss": 2.99288272857666, "ce_loss": 2.9814385175704956, "z_loss": 114.44186019897461, "muon_lr": 0.018600000000000002, "adamw_lr": 0.00186, "ngram_lr": 0.00093, "grad_norm": 0.39528924226760864, "tokens_per_second": 2833052.099588909, "eta_seconds": 6887.422401738167, "data_wait_seconds": 1.3828277587890625e-05}
95
+ {"step": 940, "total_steps": 38147, "tokens": 492830720, "loss": 2.9920470118522644, "ce_loss": 2.9807915687561035, "z_loss": 112.55487442016602, "muon_lr": 0.0188, "adamw_lr": 0.0018800000000000002, "ngram_lr": 0.0009400000000000001, "grad_norm": 0.44949570298194885, "tokens_per_second": 6275480.866983186, "eta_seconds": 3108.4763111352922, "data_wait_seconds": 1.3768672943115234e-05}
96
+ {"step": 950, "total_steps": 38147, "tokens": 498073600, "loss": 2.9330687522888184, "ce_loss": 2.921750843524933, "z_loss": 113.1789779663086, "muon_lr": 0.019, "adamw_lr": 0.0019000000000000002, "ngram_lr": 0.0009500000000000001, "grad_norm": 0.37153464555740356, "tokens_per_second": 1686702.6688753876, "eta_seconds": 11562.16865952015, "data_wait_seconds": 1.71661376953125e-05}
97
+ {"step": 960, "total_steps": 38147, "tokens": 503316480, "loss": 3.0376710891723633, "ce_loss": 3.0265387892723083, "z_loss": 111.3227481842041, "muon_lr": 0.0192, "adamw_lr": 0.0019199999999999998, "ngram_lr": 0.0009599999999999999, "grad_norm": 0.4085904359817505, "tokens_per_second": 7540779.464716624, "eta_seconds": 2585.5016642808914, "data_wait_seconds": 1.424551010131836e-05}
98
+ {"step": 970, "total_steps": 38147, "tokens": 508559360, "loss": 2.9695889949798584, "ce_loss": 2.958306908607483, "z_loss": 112.8207950592041, "muon_lr": 0.0194, "adamw_lr": 0.0019399999999999999, "ngram_lr": 0.0009699999999999999, "grad_norm": 0.4627630114555359, "tokens_per_second": 2483076.814170484, "eta_seconds": 7849.718891000747, "data_wait_seconds": 1.33514404296875e-05}
99
+ {"step": 980, "total_steps": 38147, "tokens": 513802240, "loss": 2.946959912776947, "ce_loss": 2.9358288645744324, "z_loss": 111.31134986877441, "muon_lr": 0.019600000000000003, "adamw_lr": 0.00196, "ngram_lr": 0.00098, "grad_norm": 0.3509756028652191, "tokens_per_second": 2547692.561699213, "eta_seconds": 7648.5728258371355, "data_wait_seconds": 1.430511474609375e-05}
100
+ {"step": 990, "total_steps": 38147, "tokens": 519045120, "loss": 2.954591929912567, "ce_loss": 2.943390369415283, "z_loss": 112.01510238647461, "muon_lr": 0.0198, "adamw_lr": 0.00198, "ngram_lr": 0.00099, "grad_norm": 0.49273747205734253, "tokens_per_second": 2666698.4252481647, "eta_seconds": 7305.276454043388, "data_wait_seconds": 1.9848346710205078e-05}
101
+ {"step": 1000, "total_steps": 38147, "tokens": 524288000, "loss": 2.924726188182831, "ce_loss": 2.913919985294342, "z_loss": 108.06197357177734, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.42398208379745483, "tokens_per_second": 1605650.2340449595, "eta_seconds": 12129.494906830789, "data_wait_seconds": 0.32046687602996826}
102
+ {"step": 1010, "total_steps": 38147, "tokens": 529530880, "loss": 2.9289298057556152, "ce_loss": 2.9177311062812805, "z_loss": 111.98730278015137, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.4256858229637146, "tokens_per_second": 7558349.7590290215, "eta_seconds": 2576.023083972931, "data_wait_seconds": 1.800060272216797e-05}
103
+ {"step": 1020, "total_steps": 38147, "tokens": 534773760, "loss": 2.9584363102912903, "ce_loss": 2.947560131549835, "z_loss": 108.76181983947754, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.36685916781425476, "tokens_per_second": 1564545.9115745646, "eta_seconds": 12441.463322997093, "data_wait_seconds": 1.519918441772461e-05}
104
+ {"step": 1030, "total_steps": 38147, "tokens": 540016640, "loss": 2.9317819476127625, "ce_loss": 2.9209842681884766, "z_loss": 107.97616577148438, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.355136513710022, "tokens_per_second": 2788272.8633446293, "eta_seconds": 6979.230028676986, "data_wait_seconds": 1.4007091522216797e-05}
105
+ {"step": 1040, "total_steps": 38147, "tokens": 545259520, "loss": 2.987605035305023, "ce_loss": 2.976965367794037, "z_loss": 106.3967113494873, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.37652018666267395, "tokens_per_second": 2567392.470476758, "eta_seconds": 7577.631795573235, "data_wait_seconds": 1.9729137420654297e-05}
106
+ {"step": 1050, "total_steps": 38147, "tokens": 550502400, "loss": 2.9414061307907104, "ce_loss": 2.9310302734375, "z_loss": 103.75876426696777, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.36180785298347473, "tokens_per_second": 1525853.898815238, "eta_seconds": 12746.641045451164, "data_wait_seconds": 0.32529109716415405}
107
+ {"step": 1060, "total_steps": 38147, "tokens": 555745280, "loss": 2.9102484583854675, "ce_loss": 2.899677038192749, "z_loss": 105.7143325805664, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.4226013123989105, "tokens_per_second": 7556445.970443235, "eta_seconds": 2573.202949118614, "data_wait_seconds": 1.3768672943115234e-05}
108
+ {"step": 1070, "total_steps": 38147, "tokens": 560988160, "loss": 2.8674152493476868, "ce_loss": 2.85694020986557, "z_loss": 104.75015640258789, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.35181882977485657, "tokens_per_second": 2516291.6495953896, "eta_seconds": 7725.26752974987, "data_wait_seconds": 1.436471939086914e-05}
109
+ {"step": 1080, "total_steps": 38147, "tokens": 566231040, "loss": 2.910629391670227, "ce_loss": 2.900367796421051, "z_loss": 102.61604690551758, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.36210769414901733, "tokens_per_second": 1738000.6750800605, "eta_seconds": 11181.689152741434, "data_wait_seconds": 1.4603137969970703e-05}
110
+ {"step": 1090, "total_steps": 38147, "tokens": 571473920, "loss": 2.8911567330360413, "ce_loss": 2.8808059096336365, "z_loss": 103.50761413574219, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.36968934535980225, "tokens_per_second": 2771070.9496948197, "eta_seconds": 7011.202805233002, "data_wait_seconds": 1.6987323760986328e-05}
111
+ {"step": 1100, "total_steps": 38147, "tokens": 576716800, "loss": 2.8918769359588623, "ce_loss": 2.8815014362335205, "z_loss": 103.75422286987305, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.32807397842407227, "tokens_per_second": 2508127.9544066167, "eta_seconds": 7744.141403102874, "data_wait_seconds": 1.5914440155029297e-05}
112
+ {"step": 1110, "total_steps": 38147, "tokens": 581959680, "loss": 2.9024685621261597, "ce_loss": 2.891847848892212, "z_loss": 106.20717239379883, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3861878216266632, "tokens_per_second": 2716498.9128585025, "eta_seconds": 7148.19158000946, "data_wait_seconds": 1.2516975402832031e-05}
113
+ {"step": 1120, "total_steps": 38147, "tokens": 587202560, "loss": 2.8908268213272095, "ce_loss": 2.880602240562439, "z_loss": 102.2458667755127, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.34420809149742126, "tokens_per_second": 2880587.1771238474, "eta_seconds": 6739.185652899742, "data_wait_seconds": 1.6629695892333984e-05}
114
+ {"step": 1130, "total_steps": 38147, "tokens": 592445440, "loss": 2.9057271480560303, "ce_loss": 2.8955677151679993, "z_loss": 101.59448623657227, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.38885369896888733, "tokens_per_second": 2370020.46942291, "eta_seconds": 8188.776909899711, "data_wait_seconds": 1.430511474609375e-05}
115
+ {"step": 1140, "total_steps": 38147, "tokens": 597688320, "loss": 2.9049463868141174, "ce_loss": 2.894855320453644, "z_loss": 100.91032791137695, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3611096441745758, "tokens_per_second": 2722147.2817771435, "eta_seconds": 7127.581283307076, "data_wait_seconds": 1.3887882232666016e-05}
116
+ {"step": 1150, "total_steps": 38147, "tokens": 602931200, "loss": 2.8645296692848206, "ce_loss": 2.854761064052582, "z_loss": 97.6858901977539, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.35328203439712524, "tokens_per_second": 2409901.4400756783, "eta_seconds": 8048.911384272576, "data_wait_seconds": 1.7762184143066406e-05}
117
+ {"step": 1160, "total_steps": 38147, "tokens": 608174080, "loss": 2.8724053502082825, "ce_loss": 2.862437069416046, "z_loss": 99.68237113952637, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3404342532157898, "tokens_per_second": 2643720.905313955, "eta_seconds": 7335.0557606220245, "data_wait_seconds": 1.6033649444580078e-05}
118
+ {"step": 1170, "total_steps": 38147, "tokens": 613416960, "loss": 2.7999804615974426, "ce_loss": 2.7900286316871643, "z_loss": 99.51833534240723, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.36866751313209534, "tokens_per_second": 6360286.477554992, "eta_seconds": 3048.0698384284974, "data_wait_seconds": 1.33514404296875e-05}
119
+ {"step": 1180, "total_steps": 38147, "tokens": 618659840, "loss": 2.864223301410675, "ce_loss": 2.854664981365204, "z_loss": 95.58329582214355, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3295610249042511, "tokens_per_second": 1612433.6128741086, "eta_seconds": 12019.939513325691, "data_wait_seconds": 1.9073486328125e-05}
120
+ {"step": 1190, "total_steps": 38147, "tokens": 623902720, "loss": 2.8554254174232483, "ce_loss": 2.845815062522888, "z_loss": 96.1036434173584, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.35233941674232483, "tokens_per_second": 2815655.4621051224, "eta_seconds": 6881.563414549828, "data_wait_seconds": 1.4603137969970703e-05}
121
+ {"step": 1200, "total_steps": 38147, "tokens": 629145600, "loss": 2.85099458694458, "ce_loss": 2.8413406014442444, "z_loss": 96.53900337219238, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3234858214855194, "tokens_per_second": 2605107.7051399928, "eta_seconds": 7435.726629567146, "data_wait_seconds": 1.33514404296875e-05}
122
+ {"step": 1210, "total_steps": 38147, "tokens": 634388480, "loss": 2.887865722179413, "ce_loss": 2.878342628479004, "z_loss": 95.2308349609375, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3499768078327179, "tokens_per_second": 1599361.989722008, "eta_seconds": 12108.344440126419, "data_wait_seconds": 1.3947486877441406e-05}
123
+ {"step": 1220, "total_steps": 38147, "tokens": 639631360, "loss": 2.8485671281814575, "ce_loss": 2.8389505743980408, "z_loss": 96.16562461853027, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3125127851963043, "tokens_per_second": 6415248.161731393, "eta_seconds": 3017.8696892023086, "data_wait_seconds": 1.3053417205810547e-05}
124
+ {"step": 1230, "total_steps": 38147, "tokens": 644874240, "loss": 2.8873162269592285, "ce_loss": 2.878905177116394, "z_loss": 84.11035537719727, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.36331793665885925, "tokens_per_second": 2533973.210173628, "eta_seconds": 7638.25758626461, "data_wait_seconds": 1.3887882232666016e-05}
125
+ {"step": 1240, "total_steps": 38147, "tokens": 650117120, "loss": 2.871512234210968, "ce_loss": 2.862541913986206, "z_loss": 89.70259284973145, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.335274875164032, "tokens_per_second": 2605862.4952830714, "eta_seconds": 7425.525042486191, "data_wait_seconds": 1.4841556549072266e-05}
126
+ {"step": 1250, "total_steps": 38147, "tokens": 655360000, "loss": 2.8772477507591248, "ce_loss": 2.868080496788025, "z_loss": 91.67267227172852, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3277376592159271, "tokens_per_second": 2534360.4536445714, "eta_seconds": 7632.953042721748, "data_wait_seconds": 1.5079975128173828e-05}
127
+ {"step": 1260, "total_steps": 38147, "tokens": 660602880, "loss": 2.894215226173401, "ce_loss": 2.885343015193939, "z_loss": 88.72207260131836, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3104668855667114, "tokens_per_second": 2784988.4448289997, "eta_seconds": 6944.1621892213825, "data_wait_seconds": 0.2811630964279175}
128
+ {"step": 1270, "total_steps": 38147, "tokens": 665845760, "loss": 2.8565327525138855, "ce_loss": 2.8473777770996094, "z_loss": 91.54939079284668, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3162831664085388, "tokens_per_second": 6707752.376291472, "eta_seconds": 2882.361704993248, "data_wait_seconds": 1.3768672943115234e-05}
129
+ {"step": 1280, "total_steps": 38147, "tokens": 671088640, "loss": 2.848808228969574, "ce_loss": 2.8396589159965515, "z_loss": 91.49283218383789, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3437793552875519, "tokens_per_second": 2565697.023381898, "eta_seconds": 7533.596336531639, "data_wait_seconds": 1.3828277587890625e-05}
130
+ {"step": 1290, "total_steps": 38147, "tokens": 676331520, "loss": 2.835054099559784, "ce_loss": 2.8261046409606934, "z_loss": 89.49443054199219, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.307991623878479, "tokens_per_second": 2711435.819614361, "eta_seconds": 7126.7343583106995, "data_wait_seconds": 1.3470649719238281e-05}
131
+ {"step": 1300, "total_steps": 38147, "tokens": 681574400, "loss": 2.802314579486847, "ce_loss": 2.793455719947815, "z_loss": 88.58807373046875, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3329191505908966, "tokens_per_second": 1631623.7961630016, "eta_seconds": 11840.008696508408, "data_wait_seconds": 0.265525221824646}
132
+ {"step": 1310, "total_steps": 38147, "tokens": 686817280, "loss": 2.9004313349723816, "ce_loss": 2.8922202587127686, "z_loss": 82.11084938049316, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.34262603521347046, "tokens_per_second": 2615056.5198853384, "eta_seconds": 7385.3841816186905, "data_wait_seconds": 2.2172927856445312e-05}
133
+ {"step": 1320, "total_steps": 38147, "tokens": 692060160, "loss": 2.8569279313087463, "ce_loss": 2.8483506441116333, "z_loss": 85.77322006225586, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3144185543060303, "tokens_per_second": 2631772.125593683, "eta_seconds": 7336.4840322732925, "data_wait_seconds": 0.3055676221847534}
134
+ {"step": 1330, "total_steps": 38147, "tokens": 697303040, "loss": 2.856894314289093, "ce_loss": 2.848380982875824, "z_loss": 85.13311958312988, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.29871562123298645, "tokens_per_second": 2730369.552730713, "eta_seconds": 7069.633221149445, "data_wait_seconds": 1.424551010131836e-05}
135
+ {"step": 1340, "total_steps": 38147, "tokens": 702545920, "loss": 2.8031177520751953, "ce_loss": 2.7943984866142273, "z_loss": 87.19232177734375, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27501800656318665, "tokens_per_second": 2587802.5476370105, "eta_seconds": 7457.086876130104, "data_wait_seconds": 1.4066696166992188e-05}
136
+ {"step": 1350, "total_steps": 38147, "tokens": 707788800, "loss": 2.7828128337860107, "ce_loss": 2.774497151374817, "z_loss": 83.15630531311035, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.31010302901268005, "tokens_per_second": 1654307.8281123885, "eta_seconds": 11661.811186623574, "data_wait_seconds": 1.8537044525146484e-05}
137
+ {"step": 1360, "total_steps": 38147, "tokens": 713031680, "loss": 2.8022242188453674, "ce_loss": 2.794496238231659, "z_loss": 77.27964973449707, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.32581764459609985, "tokens_per_second": 2613572.7441498605, "eta_seconds": 7379.546905350685, "data_wait_seconds": 0.3112301826477051}
138
+ {"step": 1370, "total_steps": 38147, "tokens": 718274560, "loss": 2.7481974363327026, "ce_loss": 2.73990935087204, "z_loss": 82.88132858276367, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.342210054397583, "tokens_per_second": 2663308.8283478674, "eta_seconds": 7239.768655729294, "data_wait_seconds": 3.170967102050781e-05}
139
+ {"step": 1380, "total_steps": 38147, "tokens": 723517440, "loss": 2.835219085216522, "ce_loss": 2.827329635620117, "z_loss": 78.8942699432373, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3475613594055176, "tokens_per_second": 2641304.393661897, "eta_seconds": 7298.097463607789, "data_wait_seconds": 1.4901161193847656e-05}
140
+ {"step": 1390, "total_steps": 38147, "tokens": 728760320, "loss": 2.833637773990631, "ce_loss": 2.825924277305603, "z_loss": 77.1344985961914, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2952996790409088, "tokens_per_second": 2561401.9509453033, "eta_seconds": 7523.713335537911, "data_wait_seconds": 1.4185905456542969e-05}
141
+ {"step": 1400, "total_steps": 38147, "tokens": 734003200, "loss": 2.7667983174324036, "ce_loss": 2.75865238904953, "z_loss": 81.45978164672852, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3422465920448303, "tokens_per_second": 1634050.712069265, "eta_seconds": 11790.338570094109, "data_wait_seconds": 1.800060272216797e-05}
142
+ {"step": 1410, "total_steps": 38147, "tokens": 739246080, "loss": 2.786009669303894, "ce_loss": 2.7784125208854675, "z_loss": 75.97151184082031, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2788220942020416, "tokens_per_second": 6957228.070811748, "eta_seconds": 2768.4543412923813, "data_wait_seconds": 1.3113021850585938e-05}
143
+ {"step": 1420, "total_steps": 38147, "tokens": 744488960, "loss": 2.792876660823822, "ce_loss": 2.784794509410858, "z_loss": 80.82183074951172, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.34174618124961853, "tokens_per_second": 1633283.435815648, "eta_seconds": 11789.457331013678, "data_wait_seconds": 1.4662742614746094e-05}
144
+ {"step": 1430, "total_steps": 38147, "tokens": 749731840, "loss": 2.814731776714325, "ce_loss": 2.807329833507538, "z_loss": 74.01954460144043, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.29924944043159485, "tokens_per_second": 2340681.2870438965, "eta_seconds": 8224.221982955933, "data_wait_seconds": 1.436471939086914e-05}
145
+ {"step": 1440, "total_steps": 38147, "tokens": 754974720, "loss": 2.7337063550949097, "ce_loss": 2.726025402545929, "z_loss": 76.80953216552734, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3030129671096802, "tokens_per_second": 2258180.3546671444, "eta_seconds": 8522.366061782837, "data_wait_seconds": 1.704692840576172e-05}
146
+ {"step": 1450, "total_steps": 38147, "tokens": 760217600, "loss": 2.7654570937156677, "ce_loss": 2.758043110370636, "z_loss": 74.13994789123535, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2708876430988312, "tokens_per_second": 3237253.9890408553, "eta_seconds": 5943.24597363472, "data_wait_seconds": 1.3113021850585938e-05}
147
+ {"step": 1460, "total_steps": 38147, "tokens": 765460480, "loss": 2.7607697248458862, "ce_loss": 2.7528892159461975, "z_loss": 78.80474281311035, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.34124332666397095, "tokens_per_second": 2821840.0352527373, "eta_seconds": 6816.3161680698395, "data_wait_seconds": 1.3768672943115234e-05}
148
+ {"step": 1470, "total_steps": 38147, "tokens": 770703360, "loss": 2.8053457736968994, "ce_loss": 2.7979031801223755, "z_loss": 74.42557144165039, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.32462337613105774, "tokens_per_second": 1567415.1420480788, "eta_seconds": 12268.167162704469, "data_wait_seconds": 1.5676021575927734e-05}
149
+ {"step": 1480, "total_steps": 38147, "tokens": 775946240, "loss": 2.7839956283569336, "ce_loss": 2.776710033416748, "z_loss": 72.85579490661621, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.28924083709716797, "tokens_per_second": 2671281.1950498126, "eta_seconds": 7196.572240924835, "data_wait_seconds": 1.5914440155029297e-05}
150
+ {"step": 1490, "total_steps": 38147, "tokens": 781189120, "loss": 2.741193950176239, "ce_loss": 2.7338677644729614, "z_loss": 73.26183891296387, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.28094014525413513, "tokens_per_second": 2352903.3713271315, "eta_seconds": 8168.131955695152, "data_wait_seconds": 1.990795135498047e-05}
151
+ {"step": 1500, "total_steps": 38147, "tokens": 786432000, "loss": 2.951059103012085, "ce_loss": 2.944058418273926, "z_loss": 70.00639247894287, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.365144282579422, "tokens_per_second": 6401784.607559324, "eta_seconds": 3001.2853467941286, "data_wait_seconds": 2.5212764739990234e-05}
152
+ {"step": 1510, "total_steps": 38147, "tokens": 791674880, "loss": 2.7116302847862244, "ce_loss": 2.704595148563385, "z_loss": 70.35105514526367, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.33222249150276184, "tokens_per_second": 2801345.9765211726, "eta_seconds": 6856.825117993355, "data_wait_seconds": 1.52587890625e-05}
153
+ {"step": 1520, "total_steps": 38147, "tokens": 796917760, "loss": 2.7428330779075623, "ce_loss": 2.7355427145957947, "z_loss": 72.90324401855469, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.30114972591400146, "tokens_per_second": 1688998.5569944961, "eta_seconds": 11369.516271328926, "data_wait_seconds": 1.5616416931152344e-05}
154
+ {"step": 1530, "total_steps": 38147, "tokens": 802160640, "loss": 2.781431496143341, "ce_loss": 2.774819791316986, "z_loss": 66.11725425720215, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.31724417209625244, "tokens_per_second": 2708558.666097617, "eta_seconds": 7087.848580241203, "data_wait_seconds": 1.6391277313232422e-05}
155
+ {"step": 1540, "total_steps": 38147, "tokens": 807403520, "loss": 2.746455729007721, "ce_loss": 2.739383816719055, "z_loss": 70.71903038024902, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.25161200761795044, "tokens_per_second": 2486174.692176049, "eta_seconds": 7719.735413765908, "data_wait_seconds": 1.5676021575927734e-05}
156
+ {"step": 1550, "total_steps": 38147, "tokens": 812646400, "loss": 2.726296603679657, "ce_loss": 2.7191177010536194, "z_loss": 71.78890419006348, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2958833873271942, "tokens_per_second": 2842929.0619157287, "eta_seconds": 6749.154663419723, "data_wait_seconds": 1.5139579772949219e-05}
157
+ {"step": 1560, "total_steps": 38147, "tokens": 817889280, "loss": 2.7478440403938293, "ce_loss": 2.741152584552765, "z_loss": 66.91392707824707, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2727183997631073, "tokens_per_second": 2632398.746555712, "eta_seconds": 7286.937467622757, "data_wait_seconds": 1.5497207641601562e-05}
158
+ {"step": 1570, "total_steps": 38147, "tokens": 823132160, "loss": 2.7708719968795776, "ce_loss": 2.7639504075050354, "z_loss": 69.21631240844727, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27389276027679443, "tokens_per_second": 1598911.5995570794, "eta_seconds": 11993.710084605216, "data_wait_seconds": 1.800060272216797e-05}
159
+ {"step": 1580, "total_steps": 38147, "tokens": 828375040, "loss": 2.790256917476654, "ce_loss": 2.7837425470352173, "z_loss": 65.14363670349121, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.26978978514671326, "tokens_per_second": 2590151.093555174, "eta_seconds": 7401.745536661148, "data_wait_seconds": 2.008676528930664e-05}
160
+ {"step": 1590, "total_steps": 38147, "tokens": 833617920, "loss": 2.8182828426361084, "ce_loss": 2.8112958669662476, "z_loss": 69.86994552612305, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3157118260860443, "tokens_per_second": 2464606.545867479, "eta_seconds": 7776.6556483983995, "data_wait_seconds": 1.9609928131103516e-05}
161
+ {"step": 1600, "total_steps": 38147, "tokens": 838860800, "loss": 2.753062665462494, "ce_loss": 2.746344029903412, "z_loss": 67.18671989440918, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.31922388076782227, "tokens_per_second": 2669799.3951092563, "eta_seconds": 7177.0012275457375, "data_wait_seconds": 1.4960765838623047e-05}
162
+ {"step": 1610, "total_steps": 38147, "tokens": 844103680, "loss": 2.793520927429199, "ce_loss": 2.7873623371124268, "z_loss": 61.58553886413574, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27721408009529114, "tokens_per_second": 2806031.226052965, "eta_seconds": 6826.6919049739845, "data_wait_seconds": 1.6629695892333984e-05}
163
+ {"step": 1620, "total_steps": 38147, "tokens": 849346560, "loss": 2.7549352049827576, "ce_loss": 2.7485275268554688, "z_loss": 64.07634258270264, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.28096964955329895, "tokens_per_second": 1565101.116314537, "eta_seconds": 12236.057834458352, "data_wait_seconds": 2.0802021026611328e-05}
164
+ {"step": 1630, "total_steps": 38147, "tokens": 854589440, "loss": 2.7777227759361267, "ce_loss": 2.7717427015304565, "z_loss": 59.801384925842285, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.31133636832237244, "tokens_per_second": 2736263.74112108, "eta_seconds": 6996.922339129448, "data_wait_seconds": 2.0384788513183594e-05}
165
+ {"step": 1640, "total_steps": 38147, "tokens": 859832320, "loss": 2.7790610790252686, "ce_loss": 2.7731428146362305, "z_loss": 59.18235206604004, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3280175030231476, "tokens_per_second": 2499957.658681969, "eta_seconds": 7656.202475881576, "data_wait_seconds": 2.294778823852539e-05}
166
+ {"step": 1650, "total_steps": 38147, "tokens": 865075200, "loss": 2.751917600631714, "ce_loss": 2.7455748915672302, "z_loss": 63.42680072784424, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27850833535194397, "tokens_per_second": 2573581.540946958, "eta_seconds": 7435.140030169487, "data_wait_seconds": 1.4603137969970703e-05}
167
+ {"step": 1660, "total_steps": 38147, "tokens": 870318080, "loss": 2.760525166988373, "ce_loss": 2.75418883562088, "z_loss": 63.36373805999756, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27275946736335754, "tokens_per_second": 2760655.7555824853, "eta_seconds": 6929.402993226051, "data_wait_seconds": 1.531839370727539e-05}
168
+ {"step": 1670, "total_steps": 38147, "tokens": 875560960, "loss": 2.7678259015083313, "ce_loss": 2.7619963884353638, "z_loss": 58.29490661621094, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2893087863922119, "tokens_per_second": 1646448.9679981323, "eta_seconds": 11615.576156759262, "data_wait_seconds": 0.2792210578918457}
169
+ {"step": 1680, "total_steps": 38147, "tokens": 880803840, "loss": 2.7290388345718384, "ce_loss": 2.723233938217163, "z_loss": 58.04855537414551, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.24690788984298706, "tokens_per_second": 2702760.517488974, "eta_seconds": 7073.956561183929, "data_wait_seconds": 1.4901161193847656e-05}
170
+ {"step": 1690, "total_steps": 38147, "tokens": 886046720, "loss": 2.744577169418335, "ce_loss": 2.738496482372284, "z_loss": 60.807148933410645, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2671208083629608, "tokens_per_second": 6014019.58857321, "eta_seconds": 3178.235011458397, "data_wait_seconds": 1.710653305053711e-05}
171
+ {"step": 1700, "total_steps": 38147, "tokens": 891289600, "loss": 2.74891996383667, "ce_loss": 2.742631673812866, "z_loss": 62.88258457183838, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.29851800203323364, "tokens_per_second": 2517344.486709844, "eta_seconds": 7590.826300048828, "data_wait_seconds": 1.4662742614746094e-05}
172
+ {"step": 1710, "total_steps": 38147, "tokens": 896532480, "loss": 2.761611819267273, "ce_loss": 2.755860924720764, "z_loss": 57.50888729095459, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2522546350955963, "tokens_per_second": 2896759.293772547, "eta_seconds": 6594.7771004199985, "data_wait_seconds": 1.4960765838623047e-05}
173
+ {"step": 1720, "total_steps": 38147, "tokens": 901775360, "loss": 2.788254201412201, "ce_loss": 2.782429337501526, "z_loss": 58.24919509887695, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.29173117876052856, "tokens_per_second": 2500542.410185822, "eta_seconds": 7637.63849723339, "data_wait_seconds": 1.6689300537109375e-05}
174
+ {"step": 1730, "total_steps": 38147, "tokens": 907018240, "loss": 2.747836470603943, "ce_loss": 2.7423707842826843, "z_loss": 54.65701389312744, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.26120322942733765, "tokens_per_second": 2727668.95746097, "eta_seconds": 6999.748280954361, "data_wait_seconds": 1.621246337890625e-05}
175
+ {"step": 1740, "total_steps": 38147, "tokens": 912261120, "loss": 2.7551265358924866, "ce_loss": 2.7498756051063538, "z_loss": 52.50991439819336, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2796667218208313, "tokens_per_second": 7560324.686732888, "eta_seconds": 2524.726649570465, "data_wait_seconds": 1.4960765838623047e-05}
176
+ {"step": 1750, "total_steps": 38147, "tokens": 917504000, "loss": 2.7523351311683655, "ce_loss": 2.7468464374542236, "z_loss": 54.88673210144043, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2969740331172943, "tokens_per_second": 2349501.044071323, "eta_seconds": 8121.941628479957, "data_wait_seconds": 1.4066696166992188e-05}
177
+ {"step": 1760, "total_steps": 38147, "tokens": 922746880, "loss": 2.7313899397850037, "ce_loss": 2.725655257701874, "z_loss": 57.34693431854248, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.3038170635700226, "tokens_per_second": 2924595.9047206254, "eta_seconds": 6523.04389307499, "data_wait_seconds": 1.4483928680419922e-05}
178
+ {"step": 1770, "total_steps": 38147, "tokens": 927989760, "loss": 2.707336187362671, "ce_loss": 2.701449930667877, "z_loss": 58.86225986480713, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2818325459957123, "tokens_per_second": 2318070.5503498786, "eta_seconds": 8227.54276099205, "data_wait_seconds": 2.1398067474365234e-05}
179
+ {"step": 1780, "total_steps": 38147, "tokens": 933232640, "loss": 2.7424604892730713, "ce_loss": 2.7373231649398804, "z_loss": 51.373313903808594, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.23135115206241608, "tokens_per_second": 2656015.0865279255, "eta_seconds": 7178.717392349244, "data_wait_seconds": 1.615285873413086e-05}
180
+ {"step": 1790, "total_steps": 38147, "tokens": 938475520, "loss": 2.730637013912201, "ce_loss": 2.725328207015991, "z_loss": 53.087873458862305, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.28773877024650574, "tokens_per_second": 6892491.606394309, "eta_seconds": 2765.5512555599216, "data_wait_seconds": 1.3589859008789062e-05}
181
+ {"step": 1800, "total_steps": 38147, "tokens": 943718400, "loss": 2.7252138257026672, "ce_loss": 2.7199509739875793, "z_loss": 52.62870979309082, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2927887737751007, "tokens_per_second": 1439371.395752649, "eta_seconds": 13239.318213653565, "data_wait_seconds": 2.2351741790771484e-05}
182
+ {"step": 1810, "total_steps": 38147, "tokens": 948961280, "loss": 2.812345862388611, "ce_loss": 2.8074509501457214, "z_loss": 48.94905185699463, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2802093029022217, "tokens_per_second": 7518693.981969685, "eta_seconds": 2533.824770855904, "data_wait_seconds": 1.5556812286376953e-05}
183
+ {"step": 1820, "total_steps": 38147, "tokens": 954204160, "loss": 2.7508572936058044, "ce_loss": 2.7450738549232483, "z_loss": 57.83441638946533, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.28408244252204895, "tokens_per_second": 2387292.8055333365, "eta_seconds": 7977.9950460433965, "data_wait_seconds": 2.9921531677246094e-05}
184
+ {"step": 1830, "total_steps": 38147, "tokens": 959447040, "loss": 2.7739405632019043, "ce_loss": 2.7687235474586487, "z_loss": 52.16995143890381, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.28828728199005127, "tokens_per_second": 1507721.3891651034, "eta_seconds": 12628.7041046381, "data_wait_seconds": 1.9729137420654297e-05}
185
+ {"step": 1840, "total_steps": 38147, "tokens": 964689920, "loss": 2.6930466890335083, "ce_loss": 2.6877512335777283, "z_loss": 52.95441913604736, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.24788732826709747, "tokens_per_second": 6905979.043455076, "eta_seconds": 2756.354210782051, "data_wait_seconds": 1.5556812286376953e-05}
186
+ {"step": 1850, "total_steps": 38147, "tokens": 969932800, "loss": 2.719034492969513, "ce_loss": 2.71403568983078, "z_loss": 49.987897872924805, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.24021324515342712, "tokens_per_second": 2424299.598756752, "eta_seconds": 7849.723501896859, "data_wait_seconds": 1.2695789337158203e-05}
187
+ {"step": 1860, "total_steps": 38147, "tokens": 975175680, "loss": 2.7132816910743713, "ce_loss": 2.70825332403183, "z_loss": 50.28340244293213, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27513521909713745, "tokens_per_second": 2581285.5787858367, "eta_seconds": 7370.295953440666, "data_wait_seconds": 1.5735626220703125e-05}
188
+ {"step": 1870, "total_steps": 38147, "tokens": 980418560, "loss": 2.702106535434723, "ce_loss": 2.6970843076705933, "z_loss": 50.22195529937744, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2953353822231293, "tokens_per_second": 2683922.0920847715, "eta_seconds": 7086.493245124817, "data_wait_seconds": 2.1159648895263672e-05}
189
+ {"step": 1880, "total_steps": 38147, "tokens": 985661440, "loss": 2.74798321723938, "ce_loss": 2.7427735328674316, "z_loss": 52.09687328338623, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.27042293548583984, "tokens_per_second": 2557135.210381743, "eta_seconds": 7435.80269780159, "data_wait_seconds": 1.7344951629638672e-05}
190
+ {"step": 1890, "total_steps": 38147, "tokens": 990904320, "loss": 2.6904376745224, "ce_loss": 2.685773015022278, "z_loss": 46.64657974243164, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.26898521184921265, "tokens_per_second": 6823365.3125744695, "eta_seconds": 2785.8848449707034, "data_wait_seconds": 1.4185905456542969e-05}
191
+ {"step": 1900, "total_steps": 38147, "tokens": 996147200, "loss": 2.773164987564087, "ce_loss": 2.7682440280914307, "z_loss": 49.20980262756348, "muon_lr": 0.02, "adamw_lr": 0.002, "ngram_lr": 0.001, "grad_norm": 0.2981660068035126, "tokens_per_second": 2689600.7918179273, "eta_seconds": 7065.683202433585, "data_wait_seconds": 1.52587890625e-05}