Banaxi-Tech commited on
Commit
f362cf1
·
verified ·
1 Parent(s): 5c111f2

Save 5% checkpoint at 1,500,512,256 tokens

Browse files
README.md CHANGED
@@ -1,4 +1,3 @@
1
  Under 3M parameters, partially looped model.
2
 
3
  1M in n-gram embeddings.
4
-
 
1
  Under 3M parameters, partially looped model.
2
 
3
  1M in n-gram embeddings.
 
checkpoint_metadata.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "parameters": 2873061,
3
+ "neural_core_parameters": 1873060,
4
+ "ngram_parameters": 1000001,
5
+ "architecture": {
6
+ "vocab_size": 2048,
7
+ "hidden_size": 160,
8
+ "intermediate_size": 512,
9
+ "num_hidden_layers": 4,
10
+ "num_attention_heads": 5,
11
+ "num_key_value_heads": 1,
12
+ "head_dim": 32,
13
+ "ngram_buckets": 6250,
14
+ "loop_schedule": [
15
+ 0,
16
+ 1,
17
+ 1,
18
+ 2,
19
+ 3
20
+ ],
21
+ "refresh_kernel_size": 9,
22
+ "max_position_embeddings": 4096,
23
+ "rope_theta": 100000.0,
24
+ "rms_norm_eps": 1e-05,
25
+ "initializer_range": 0.02,
26
+ "use_cache": false,
27
+ "transformers_version": "5.16.1",
28
+ "architectures": null,
29
+ "output_hidden_states": false,
30
+ "return_dict": true,
31
+ "dtype": null,
32
+ "chunk_size_feed_forward": 0,
33
+ "is_encoder_decoder": false,
34
+ "id2label": {
35
+ "0": "LABEL_0",
36
+ "1": "LABEL_1"
37
+ },
38
+ "label2id": {
39
+ "LABEL_0": 0,
40
+ "LABEL_1": 1
41
+ },
42
+ "problem_type": null,
43
+ "_name_or_path": "",
44
+ "tie_word_embeddings": true,
45
+ "bos_token_id": 1,
46
+ "eos_token_id": 2,
47
+ "pad_token_id": 0,
48
+ "unk_token_id": 3,
49
+ "model_type": "bird_alpha",
50
+ "output_attentions": false
51
+ },
52
+ "training_percent": 5,
53
+ "step": 1431,
54
+ "total_steps": 28611,
55
+ "tokens_seen": 1500512256,
56
+ "target_tokens": 30000000000,
57
+ "trained_token_target": 30000807936,
58
+ "tokens_per_step": 1048576,
59
+ "local_batch": 64,
60
+ "effective_global_batch": 256,
61
+ "world_size": 4,
62
+ "gpu_name": "NVIDIA H200",
63
+ "optimizer": "Muon + AdamW",
64
+ "muon_peak_lr": 0.03,
65
+ "adamw_peak_lr": 0.003,
66
+ "architecture_revision": "5c111f242535ea87cb8e51d8ae6e5e203e7dede3",
67
+ "tokenizer_revision": "5c111f242535ea87cb8e51d8ae6e5e203e7dede3",
68
+ "dataset_id": "HuggingFaceFW/fineweb-edu",
69
+ "dataset_config": "sample-100BT",
70
+ "dataset_revision": "87f09149ef4734204d70ed1d046ddc9ca3f2b8f9",
71
+ "elapsed_seconds": 475.5409269332886
72
+ }
config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "vocab_size": 2048,
3
+ "hidden_size": 160,
4
+ "intermediate_size": 512,
5
+ "num_hidden_layers": 4,
6
+ "num_attention_heads": 5,
7
+ "num_key_value_heads": 1,
8
+ "head_dim": 32,
9
+ "ngram_buckets": 6250,
10
+ "loop_schedule": [
11
+ 0,
12
+ 1,
13
+ 1,
14
+ 2,
15
+ 3
16
+ ],
17
+ "refresh_kernel_size": 9,
18
+ "max_position_embeddings": 4096,
19
+ "rope_theta": 100000.0,
20
+ "rms_norm_eps": 1e-05,
21
+ "initializer_range": 0.02,
22
+ "use_cache": false,
23
+ "transformers_version": "5.16.1",
24
+ "architectures": [
25
+ "BirdAlphaForCausalLM"
26
+ ],
27
+ "output_hidden_states": false,
28
+ "return_dict": true,
29
+ "dtype": null,
30
+ "chunk_size_feed_forward": 0,
31
+ "is_encoder_decoder": false,
32
+ "id2label": {
33
+ "0": "LABEL_0",
34
+ "1": "LABEL_1"
35
+ },
36
+ "label2id": {
37
+ "LABEL_0": 0,
38
+ "LABEL_1": 1
39
+ },
40
+ "problem_type": null,
41
+ "_name_or_path": "Banaxi-Tech/bird-alpha",
42
+ "tie_word_embeddings": true,
43
+ "bos_token_id": 1,
44
+ "eos_token_id": 2,
45
+ "pad_token_id": 0,
46
+ "unk_token_id": 3,
47
+ "model_type": "bird_alpha",
48
+ "output_attentions": false,
49
+ "auto_map": {
50
+ "AutoConfig": "configuration_bird_alpha.BirdAlphaConfig",
51
+ "AutoModelForCausalLM": "modeling_bird_alpha.BirdAlphaForCausalLM"
52
+ },
53
+ "torch_dtype": "float32"
54
+ }
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:e6e383c2b757a44f5633c9a06ba85b9018559ad246c83119fbfe892e7e6b537a
3
+ size 13860292
training_metrics.jsonl ADDED
@@ -0,0 +1,144 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"step": 1, "total_steps": 28611, "tokens": 1048576, "loss": 7.658415675163269, "ce_loss": 7.652538180351257, "z_loss": 58.77440071105957, "muon_lr": 2.9999999999999997e-05, "adamw_lr": 3e-06, "grad_norm": 0.2501152455806732, "tokens_per_second": 26121.44329903569, "eta_seconds": 1148472.502708435, "data_wait_seconds": 6.96092289686203}
2
+ {"step": 10, "total_steps": 28611, "tokens": 10485760, "loss": 7.650874495506287, "ce_loss": 7.644997477531433, "z_loss": 58.76986026763916, "muon_lr": 0.0003, "adamw_lr": 2.9999999999999997e-05, "grad_norm": 0.24064870178699493, "tokens_per_second": 9640523.625252537, "eta_seconds": 3110.860295746061, "data_wait_seconds": 1.8358230590820312e-05}
3
+ {"step": 20, "total_steps": 28611, "tokens": 20971520, "loss": 7.625383377075195, "ce_loss": 7.619508504867554, "z_loss": 58.7486572265625, "muon_lr": 0.0006, "adamw_lr": 5.9999999999999995e-05, "grad_norm": 0.2815065085887909, "tokens_per_second": 3985046.9996836823, "eta_seconds": 7523.0822668790815, "data_wait_seconds": 1.6510486602783203e-05}
4
+ {"step": 30, "total_steps": 28611, "tokens": 31457280, "loss": 7.581531047821045, "ce_loss": 7.575660705566406, "z_loss": 58.704139709472656, "muon_lr": 0.0008999999999999999, "adamw_lr": 8.999999999999999e-05, "grad_norm": 0.37200769782066345, "tokens_per_second": 4320569.62453979, "eta_seconds": 6936.435067677498, "data_wait_seconds": 1.615285873413086e-05}
5
+ {"step": 40, "total_steps": 28611, "tokens": 41943040, "loss": 7.501992225646973, "ce_loss": 7.496133327484131, "z_loss": 58.58889102935791, "muon_lr": 0.0012, "adamw_lr": 0.00011999999999999999, "grad_norm": 0.5417554378509521, "tokens_per_second": 3827110.8439003434, "eta_seconds": 7828.0630266428, "data_wait_seconds": 0.5764961242675781}
6
+ {"step": 50, "total_steps": 28611, "tokens": 52428800, "loss": 7.380180835723877, "ce_loss": 7.3743510246276855, "z_loss": 58.29871082305908, "muon_lr": 0.0015, "adamw_lr": 0.00015, "grad_norm": 0.7928475737571716, "tokens_per_second": 10493595.85124161, "eta_seconds": 2853.9672730445864, "data_wait_seconds": 1.3649463653564453e-05}
7
+ {"step": 60, "total_steps": 28611, "tokens": 62914560, "loss": 7.1978737115859985, "ce_loss": 7.19210147857666, "z_loss": 57.72246837615967, "muon_lr": 0.0017999999999999997, "adamw_lr": 0.00017999999999999998, "grad_norm": 1.0268207788467407, "tokens_per_second": 4109971.613744878, "eta_seconds": 7284.209281611443, "data_wait_seconds": 1.3768672943115234e-05}
8
+ {"step": 70, "total_steps": 28611, "tokens": 73400320, "loss": 6.966561555862427, "ce_loss": 6.960871577262878, "z_loss": 56.89952278137207, "muon_lr": 0.0021000000000000003, "adamw_lr": 0.00020999999999999998, "grad_norm": 1.1235384941101074, "tokens_per_second": 3974418.4186084676, "eta_seconds": 7530.0092903852465, "data_wait_seconds": 1.5139579772949219e-05}
9
+ {"step": 80, "total_steps": 28611, "tokens": 83886080, "loss": 6.758980751037598, "ce_loss": 6.7533475160598755, "z_loss": 56.33141231536865, "muon_lr": 0.0024, "adamw_lr": 0.00023999999999999998, "grad_norm": 0.9458981156349182, "tokens_per_second": 4487315.825950716, "eta_seconds": 6666.997157406807, "data_wait_seconds": 0.5646641254425049}
10
+ {"step": 90, "total_steps": 28611, "tokens": 94371840, "loss": 6.4907461404800415, "ce_loss": 6.485217094421387, "z_loss": 55.29051876068115, "muon_lr": 0.0026999999999999997, "adamw_lr": 0.00027, "grad_norm": 0.8544664978981018, "tokens_per_second": 8691441.117994452, "eta_seconds": 3440.9064837455753, "data_wait_seconds": 1.5497207641601562e-05}
11
+ {"step": 100, "total_steps": 28611, "tokens": 104857600, "loss": 6.3015841245651245, "ce_loss": 6.296112775802612, "z_loss": 54.71487236022949, "muon_lr": 0.003, "adamw_lr": 0.0003, "grad_norm": 0.6929007768630981, "tokens_per_second": 4084996.1027654517, "eta_seconds": 7318.47707657814, "data_wait_seconds": 1.5139579772949219e-05}
12
+ {"step": 110, "total_steps": 28611, "tokens": 115343360, "loss": 6.1221970319747925, "ce_loss": 6.116770505905151, "z_loss": 54.265398025512695, "muon_lr": 0.0033, "adamw_lr": 0.00033, "grad_norm": 0.5652133226394653, "tokens_per_second": 2907323.794510571, "eta_seconds": 10279.372608041762, "data_wait_seconds": 0.23885387182235718}
13
+ {"step": 120, "total_steps": 28611, "tokens": 125829120, "loss": 5.953835725784302, "ce_loss": 5.948423385620117, "z_loss": 54.12429904937744, "muon_lr": 0.0035999999999999995, "adamw_lr": 0.00035999999999999997, "grad_norm": 0.4497940242290497, "tokens_per_second": 4086372.353506308, "eta_seconds": 7310.880221271515, "data_wait_seconds": 0.5769467353820801}
14
+ {"step": 130, "total_steps": 28611, "tokens": 136314880, "loss": 5.778643488883972, "ce_loss": 5.773171901702881, "z_loss": 54.71708869934082, "muon_lr": 0.0039, "adamw_lr": 0.00039, "grad_norm": 0.3327091932296753, "tokens_per_second": 10472207.538728662, "eta_seconds": 2851.7858288764955, "data_wait_seconds": 1.4603137969970703e-05}
15
+ {"step": 140, "total_steps": 28611, "tokens": 146800640, "loss": 5.598773717880249, "ce_loss": 5.593178629875183, "z_loss": 55.94973373413086, "muon_lr": 0.004200000000000001, "adamw_lr": 0.00041999999999999996, "grad_norm": 0.2598232328891754, "tokens_per_second": 3034017.482395055, "eta_seconds": 9839.761131644249, "data_wait_seconds": 1.4960765838623047e-05}
16
+ {"step": 150, "total_steps": 28611, "tokens": 157286400, "loss": 5.480322360992432, "ce_loss": 5.474583983421326, "z_loss": 57.38265609741211, "muon_lr": 0.0045, "adamw_lr": 0.00045, "grad_norm": 0.23036696016788483, "tokens_per_second": 2904060.452353806, "eta_seconds": 10276.480819058419, "data_wait_seconds": 1.8358230590820312e-05}
17
+ {"step": 160, "total_steps": 28611, "tokens": 167772160, "loss": 5.325891733169556, "ce_loss": 5.319991827011108, "z_loss": 58.99911594390869, "muon_lr": 0.0048, "adamw_lr": 0.00047999999999999996, "grad_norm": 0.27997156977653503, "tokens_per_second": 4232833.51628002, "eta_seconds": 7048.005942416191, "data_wait_seconds": 2.4557113647460938e-05}
18
+ {"step": 170, "total_steps": 28611, "tokens": 178257920, "loss": 5.1897053718566895, "ce_loss": 5.183643102645874, "z_loss": 60.62161350250244, "muon_lr": 0.0050999999999999995, "adamw_lr": 0.00051, "grad_norm": 0.32591354846954346, "tokens_per_second": 4522172.922345714, "eta_seconds": 6594.738973522187, "data_wait_seconds": 2.2709369659423828e-05}
19
+ {"step": 180, "total_steps": 28611, "tokens": 188743680, "loss": 5.07506799697876, "ce_loss": 5.068827152252197, "z_loss": 62.40785217285156, "muon_lr": 0.005399999999999999, "adamw_lr": 0.00054, "grad_norm": 0.41397807002067566, "tokens_per_second": 2835512.981041012, "eta_seconds": 10513.81688439846, "data_wait_seconds": 1.5079975128173828e-05}
20
+ {"step": 190, "total_steps": 28611, "tokens": 199229440, "loss": 4.964954853057861, "ce_loss": 4.95854389667511, "z_loss": 64.10907936096191, "muon_lr": 0.0057, "adamw_lr": 0.0005700000000000001, "grad_norm": 0.5261039137840271, "tokens_per_second": 2881485.5251134075, "eta_seconds": 10342.435607004165, "data_wait_seconds": 0.2694505453109741}
21
+ {"step": 200, "total_steps": 28611, "tokens": 209715200, "loss": 4.888995289802551, "ce_loss": 4.882405042648315, "z_loss": 65.90203666687012, "muon_lr": 0.006, "adamw_lr": 0.0006, "grad_norm": 0.5600895285606384, "tokens_per_second": 4233518.844883411, "eta_seconds": 7036.957629704476, "data_wait_seconds": 1.5854835510253906e-05}
22
+ {"step": 210, "total_steps": 28611, "tokens": 220200960, "loss": 4.788140773773193, "ce_loss": 4.781434059143066, "z_loss": 67.06870079040527, "muon_lr": 0.0063, "adamw_lr": 0.00063, "grad_norm": 0.6154087781906128, "tokens_per_second": 4705206.6349173775, "eta_seconds": 6329.287805342675, "data_wait_seconds": 1.52587890625e-05}
23
+ {"step": 220, "total_steps": 28611, "tokens": 230686720, "loss": 4.66383957862854, "ce_loss": 4.656960964202881, "z_loss": 68.7861156463623, "muon_lr": 0.0066, "adamw_lr": 0.00066, "grad_norm": 0.585784912109375, "tokens_per_second": 3181187.785021972, "eta_seconds": 9358.177897000312, "data_wait_seconds": 1.4841556549072266e-05}
24
+ {"step": 230, "total_steps": 28611, "tokens": 241172480, "loss": 4.550064325332642, "ce_loss": 4.5429768562316895, "z_loss": 70.87542724609375, "muon_lr": 0.0069, "adamw_lr": 0.0006900000000000001, "grad_norm": 0.5849143266677856, "tokens_per_second": 2907896.8208084726, "eta_seconds": 10234.075446915627, "data_wait_seconds": 2.0265579223632812e-05}
25
+ {"step": 240, "total_steps": 28611, "tokens": 251658240, "loss": 4.48644483089447, "ce_loss": 4.479283332824707, "z_loss": 71.61376190185547, "muon_lr": 0.007199999999999999, "adamw_lr": 0.0007199999999999999, "grad_norm": 0.7232865691184998, "tokens_per_second": 4203794.224541029, "eta_seconds": 7076.737848472594, "data_wait_seconds": 1.2755393981933594e-05}
26
+ {"step": 250, "total_steps": 28611, "tokens": 262144000, "loss": 4.4155155420303345, "ce_loss": 4.4081573486328125, "z_loss": 73.58216667175293, "muon_lr": 0.0075, "adamw_lr": 0.00075, "grad_norm": 0.629396915435791, "tokens_per_second": 2936917.058415517, "eta_seconds": 10125.809937596321, "data_wait_seconds": 3.9577484130859375e-05}
27
+ {"step": 260, "total_steps": 28611, "tokens": 272629760, "loss": 4.356036901473999, "ce_loss": 4.348586201667786, "z_loss": 74.50595092773438, "muon_lr": 0.0078, "adamw_lr": 0.00078, "grad_norm": 0.6383018493652344, "tokens_per_second": 3115016.1961450703, "eta_seconds": 9543.506776237487, "data_wait_seconds": 0.26773279905319214}
28
+ {"step": 270, "total_steps": 28611, "tokens": 283115520, "loss": 4.286056637763977, "ce_loss": 4.278461933135986, "z_loss": 75.94821166992188, "muon_lr": 0.0081, "adamw_lr": 0.0008100000000000001, "grad_norm": 0.5395342111587524, "tokens_per_second": 4037690.692205999, "eta_seconds": 7360.071556091309, "data_wait_seconds": 1.4960765838623047e-05}
29
+ {"step": 280, "total_steps": 28611, "tokens": 293601280, "loss": 4.225488305091858, "ce_loss": 4.217692852020264, "z_loss": 77.95409393310547, "muon_lr": 0.008400000000000001, "adamw_lr": 0.0008399999999999999, "grad_norm": 0.5229359865188599, "tokens_per_second": 2986191.1582815987, "eta_seconds": 9948.193227219583, "data_wait_seconds": 1.6510486602783203e-05}
30
+ {"step": 290, "total_steps": 28611, "tokens": 304087040, "loss": 4.172782778739929, "ce_loss": 4.164864659309387, "z_loss": 79.18205451965332, "muon_lr": 0.0087, "adamw_lr": 0.00087, "grad_norm": 0.6144379377365112, "tokens_per_second": 2816403.198565148, "eta_seconds": 10544.200812983514, "data_wait_seconds": 0.03830081224441528}
31
+ {"step": 300, "total_steps": 28611, "tokens": 314572800, "loss": 4.118071913719177, "ce_loss": 4.110149383544922, "z_loss": 79.22486114501953, "muon_lr": 0.009, "adamw_lr": 0.0009, "grad_norm": 0.6424883008003235, "tokens_per_second": 2928278.516093577, "eta_seconds": 10137.77718644142, "data_wait_seconds": 2.0563602447509766e-05}
32
+ {"step": 310, "total_steps": 28611, "tokens": 325058560, "loss": 4.100507378578186, "ce_loss": 4.092403769493103, "z_loss": 81.03700256347656, "muon_lr": 0.0093, "adamw_lr": 0.00093, "grad_norm": 0.5703659057617188, "tokens_per_second": 2758221.322972699, "eta_seconds": 10759.01673619747, "data_wait_seconds": 1.609325408935547e-05}
33
+ {"step": 320, "total_steps": 28611, "tokens": 335544320, "loss": 4.0376598834991455, "ce_loss": 4.029485702514648, "z_loss": 81.7427806854248, "muon_lr": 0.0096, "adamw_lr": 0.0009599999999999999, "grad_norm": 0.5430964231491089, "tokens_per_second": 4303733.210411089, "eta_seconds": 6892.9141667604445, "data_wait_seconds": 1.4722347259521484e-05}
34
+ {"step": 330, "total_steps": 28611, "tokens": 346030080, "loss": 4.046276569366455, "ce_loss": 4.037989377975464, "z_loss": 82.87249946594238, "muon_lr": 0.0099, "adamw_lr": 0.00099, "grad_norm": 0.5942437052726746, "tokens_per_second": 4630049.981591588, "eta_seconds": 6404.850481939316, "data_wait_seconds": 1.7464160919189453e-05}
35
+ {"step": 340, "total_steps": 28611, "tokens": 356515840, "loss": 4.022267758846283, "ce_loss": 4.01377010345459, "z_loss": 84.97736167907715, "muon_lr": 0.010199999999999999, "adamw_lr": 0.00102, "grad_norm": 0.6505797505378723, "tokens_per_second": 3020554.1876086337, "eta_seconds": 9814.189799213409, "data_wait_seconds": 1.4424324035644531e-05}
36
+ {"step": 350, "total_steps": 28611, "tokens": 367001600, "loss": 3.983468532562256, "ce_loss": 3.9749914407730103, "z_loss": 84.77128410339355, "muon_lr": 0.0105, "adamw_lr": 0.0010500000000000002, "grad_norm": 0.5382428765296936, "tokens_per_second": 3060953.25699869, "eta_seconds": 9681.234520077705, "data_wait_seconds": 1.9431114196777344e-05}
37
+ {"step": 360, "total_steps": 28611, "tokens": 377487360, "loss": 3.956587314605713, "ce_loss": 3.947991669178009, "z_loss": 85.95678329467773, "muon_lr": 0.010799999999999999, "adamw_lr": 0.00108, "grad_norm": 0.5217306017875671, "tokens_per_second": 3196455.4395741127, "eta_seconds": 9267.553118133545, "data_wait_seconds": 0.2250012755393982}
38
+ {"step": 370, "total_steps": 28611, "tokens": 387973120, "loss": 3.9053712487220764, "ce_loss": 3.8966651558876038, "z_loss": 87.06047248840332, "muon_lr": 0.0111, "adamw_lr": 0.00111, "grad_norm": 0.5119828581809998, "tokens_per_second": 3770526.4639239963, "eta_seconds": 7853.766602444649, "data_wait_seconds": 1.4781951904296875e-05}
39
+ {"step": 380, "total_steps": 28611, "tokens": 398458880, "loss": 3.922762453556061, "ce_loss": 3.9140048027038574, "z_loss": 87.57659912109375, "muon_lr": 0.0114, "adamw_lr": 0.0011400000000000002, "grad_norm": 0.4903268814086914, "tokens_per_second": 2973132.1813034415, "eta_seconds": 9956.620577502252, "data_wait_seconds": 1.4424324035644531e-05}
40
+ {"step": 390, "total_steps": 28611, "tokens": 408944640, "loss": 3.863273501396179, "ce_loss": 3.8543195128440857, "z_loss": 89.53996276855469, "muon_lr": 0.011699999999999999, "adamw_lr": 0.00117, "grad_norm": 0.5509487986564636, "tokens_per_second": 3151383.7001267276, "eta_seconds": 9390.11752038002, "data_wait_seconds": 1.5974044799804688e-05}
41
+ {"step": 400, "total_steps": 28611, "tokens": 419430400, "loss": 3.9289989471435547, "ce_loss": 3.920101284980774, "z_loss": 88.97674369812012, "muon_lr": 0.012, "adamw_lr": 0.0012, "grad_norm": 0.5293706655502319, "tokens_per_second": 4416376.239046539, "eta_seconds": 6698.1108345031735, "data_wait_seconds": 1.52587890625e-05}
42
+ {"step": 410, "total_steps": 28611, "tokens": 429916160, "loss": 3.85553377866745, "ce_loss": 3.8464804887771606, "z_loss": 90.53336143493652, "muon_lr": 0.012299999999999998, "adamw_lr": 0.00123, "grad_norm": 0.5603596568107605, "tokens_per_second": 3290579.0857434436, "eta_seconds": 8986.531247377396, "data_wait_seconds": 1.5020370483398438e-05}
43
+ {"step": 420, "total_steps": 28611, "tokens": 440401920, "loss": 3.7968878149986267, "ce_loss": 3.7876104712486267, "z_loss": 92.7734546661377, "muon_lr": 0.0126, "adamw_lr": 0.00126, "grad_norm": 0.4505522549152374, "tokens_per_second": 2994604.465552165, "eta_seconds": 9871.222178435326, "data_wait_seconds": 1.901388168334961e-05}
44
+ {"step": 430, "total_steps": 28611, "tokens": 450887680, "loss": 3.7956278920173645, "ce_loss": 3.786339223384857, "z_loss": 92.88668060302734, "muon_lr": 0.0129, "adamw_lr": 0.0012900000000000001, "grad_norm": 0.4249292016029358, "tokens_per_second": 3504553.2902044626, "eta_seconds": 8431.865007901191, "data_wait_seconds": 2.2470951080322266e-05}
45
+ {"step": 440, "total_steps": 28611, "tokens": 461373440, "loss": 3.7461383938789368, "ce_loss": 3.73668771982193, "z_loss": 94.50640869140625, "muon_lr": 0.0132, "adamw_lr": 0.00132, "grad_norm": 0.5242290496826172, "tokens_per_second": 3040534.0879493183, "eta_seconds": 9715.212407279014, "data_wait_seconds": 0.46789973974227905}
46
+ {"step": 450, "total_steps": 28611, "tokens": 471859200, "loss": 3.7524389028549194, "ce_loss": 3.7428849935531616, "z_loss": 95.53912544250488, "muon_lr": 0.0135, "adamw_lr": 0.00135, "grad_norm": 0.4392189383506775, "tokens_per_second": 4529304.860088389, "eta_seconds": 6519.532168436051, "data_wait_seconds": 1.6391277313232422e-05}
47
+ {"step": 460, "total_steps": 28611, "tokens": 482344960, "loss": 3.7209240794181824, "ce_loss": 3.711251974105835, "z_loss": 96.72179794311523, "muon_lr": 0.0138, "adamw_lr": 0.0013800000000000002, "grad_norm": 0.4613193869590759, "tokens_per_second": 3128710.733053156, "eta_seconds": 9434.705057311057, "data_wait_seconds": 1.4901161193847656e-05}
48
+ {"step": 470, "total_steps": 28611, "tokens": 492830720, "loss": 3.7094306349754333, "ce_loss": 3.699747860431671, "z_loss": 96.82803726196289, "muon_lr": 0.0141, "adamw_lr": 0.00141, "grad_norm": 0.479788213968277, "tokens_per_second": 4969502.527441014, "eta_seconds": 5937.813101625442, "data_wait_seconds": 1.7940998077392578e-05}
49
+ {"step": 480, "total_steps": 28611, "tokens": 503316480, "loss": 3.6889609694480896, "ce_loss": 3.679142713546753, "z_loss": 98.18230628967285, "muon_lr": 0.014399999999999998, "adamw_lr": 0.0014399999999999999, "grad_norm": 0.4321846663951874, "tokens_per_second": 3204067.907384038, "eta_seconds": 9206.262884759903, "data_wait_seconds": 1.4483928680419922e-05}
50
+ {"step": 490, "total_steps": 28611, "tokens": 513802240, "loss": 3.642675280570984, "ce_loss": 3.6328312158584595, "z_loss": 98.44086074829102, "muon_lr": 0.0147, "adamw_lr": 0.00147, "grad_norm": 0.4882560968399048, "tokens_per_second": 3013422.752364635, "eta_seconds": 9785.2203687191, "data_wait_seconds": 1.537799835205078e-05}
51
+ {"step": 500, "total_steps": 28611, "tokens": 524288000, "loss": 3.627257823944092, "ce_loss": 3.6171244382858276, "z_loss": 101.33450317382812, "muon_lr": 0.015, "adamw_lr": 0.0015, "grad_norm": 0.42968258261680603, "tokens_per_second": 2876288.5355830668, "eta_seconds": 10248.10952425003, "data_wait_seconds": 0.29945123195648193}
52
+ {"step": 510, "total_steps": 28611, "tokens": 534773760, "loss": 3.6098544001579285, "ce_loss": 3.5997840762138367, "z_loss": 100.70385551452637, "muon_lr": 0.0153, "adamw_lr": 0.0015300000000000001, "grad_norm": 0.4384866952896118, "tokens_per_second": 4121368.725162058, "eta_seconds": 7149.574847817421, "data_wait_seconds": 1.5497207641601562e-05}
53
+ {"step": 520, "total_steps": 28611, "tokens": 545259520, "loss": 3.6037115454673767, "ce_loss": 3.5936650037765503, "z_loss": 100.46476936340332, "muon_lr": 0.0156, "adamw_lr": 0.00156, "grad_norm": 0.4173238277435303, "tokens_per_second": 3318404.3269965407, "eta_seconds": 8876.419361066819, "data_wait_seconds": 0.2597828507423401}
54
+ {"step": 530, "total_steps": 28611, "tokens": 555745280, "loss": 3.5656059980392456, "ce_loss": 3.5553550124168396, "z_loss": 102.50959777832031, "muon_lr": 0.015899999999999997, "adamw_lr": 0.00159, "grad_norm": 0.3985057473182678, "tokens_per_second": 4411908.728328408, "eta_seconds": 6673.996328830719, "data_wait_seconds": 1.2278556823730469e-05}
55
+ {"step": 540, "total_steps": 28611, "tokens": 566231040, "loss": 3.5345144271850586, "ce_loss": 3.5240917801856995, "z_loss": 104.22616195678711, "muon_lr": 0.0162, "adamw_lr": 0.0016200000000000001, "grad_norm": 0.3384662866592407, "tokens_per_second": 2966848.957822013, "eta_seconds": 9921.157873034477, "data_wait_seconds": 1.3649463653564453e-05}
56
+ {"step": 550, "total_steps": 28611, "tokens": 576716800, "loss": 3.520662307739258, "ce_loss": 3.5101261734962463, "z_loss": 105.36075592041016, "muon_lr": 0.0165, "adamw_lr": 0.0016500000000000002, "grad_norm": 0.42579013109207153, "tokens_per_second": 3056823.801204298, "eta_seconds": 9625.70728623867, "data_wait_seconds": 1.7762184143066406e-05}
57
+ {"step": 560, "total_steps": 28611, "tokens": 587202560, "loss": 3.4658342003822327, "ce_loss": 3.45522940158844, "z_loss": 106.04750442504883, "muon_lr": 0.016800000000000002, "adamw_lr": 0.0016799999999999999, "grad_norm": 0.3602370023727417, "tokens_per_second": 4653193.684445071, "eta_seconds": 6321.16506869793, "data_wait_seconds": 1.2934207916259766e-05}
58
+ {"step": 570, "total_steps": 28611, "tokens": 597688320, "loss": 3.436771512031555, "ce_loss": 3.4259150624275208, "z_loss": 108.56437873840332, "muon_lr": 0.017099999999999997, "adamw_lr": 0.00171, "grad_norm": 0.39852383732795715, "tokens_per_second": 3207857.3374175825, "eta_seconds": 9165.968596243858, "data_wait_seconds": 1.8596649169921875e-05}
59
+ {"step": 580, "total_steps": 28611, "tokens": 608174080, "loss": 3.424259603023529, "ce_loss": 3.4134265184402466, "z_loss": 108.33104515075684, "muon_lr": 0.0174, "adamw_lr": 0.00174, "grad_norm": 0.37904128432273865, "tokens_per_second": 3291341.2456180463, "eta_seconds": 8930.290620923042, "data_wait_seconds": 2.4318695068359375e-05}
60
+ {"step": 590, "total_steps": 28611, "tokens": 618659840, "loss": 3.4032499194145203, "ce_loss": 3.3922666907310486, "z_loss": 109.8320198059082, "muon_lr": 0.0177, "adamw_lr": 0.00177, "grad_norm": 0.38818180561065674, "tokens_per_second": 3185228.6555029857, "eta_seconds": 9224.501997756957, "data_wait_seconds": 0.2531042695045471}
61
+ {"step": 600, "total_steps": 28611, "tokens": 629145600, "loss": 3.3748250007629395, "ce_loss": 3.3638306856155396, "z_loss": 109.94318580627441, "muon_lr": 0.018, "adamw_lr": 0.0018, "grad_norm": 0.4476962685585022, "tokens_per_second": 4285969.271839257, "eta_seconds": 6852.980148267747, "data_wait_seconds": 1.7762184143066406e-05}
62
+ {"step": 610, "total_steps": 28611, "tokens": 639631360, "loss": 3.373824119567871, "ce_loss": 3.362819194793701, "z_loss": 110.04925918579102, "muon_lr": 0.0183, "adamw_lr": 0.00183, "grad_norm": 0.3837887942790985, "tokens_per_second": 4204328.7026657155, "eta_seconds": 6983.558768224715, "data_wait_seconds": 1.7523765563964844e-05}
63
+ {"step": 620, "total_steps": 28611, "tokens": 650117120, "loss": 3.3434019088745117, "ce_loss": 3.3322024941444397, "z_loss": 111.99420928955078, "muon_lr": 0.0186, "adamw_lr": 0.00186, "grad_norm": 0.36157673597335815, "tokens_per_second": 2971066.4631645023, "eta_seconds": 9878.840200948714, "data_wait_seconds": 1.6987323760986328e-05}
64
+ {"step": 630, "total_steps": 28611, "tokens": 660602880, "loss": 3.352968990802765, "ce_loss": 3.3416643142700195, "z_loss": 113.0467472076416, "muon_lr": 0.0189, "adamw_lr": 0.0018900000000000002, "grad_norm": 0.3507070541381836, "tokens_per_second": 3064367.7725730836, "eta_seconds": 9574.635694384575, "data_wait_seconds": 0.2604506015777588}
65
+ {"step": 640, "total_steps": 28611, "tokens": 671088640, "loss": 3.3101269602775574, "ce_loss": 3.298548102378845, "z_loss": 115.78902626037598, "muon_lr": 0.0192, "adamw_lr": 0.0019199999999999998, "grad_norm": 0.3767795264720917, "tokens_per_second": 4317431.43141659, "eta_seconds": 6793.326023101806, "data_wait_seconds": 1.4603137969970703e-05}
66
+ {"step": 650, "total_steps": 28611, "tokens": 681574400, "loss": 3.2799994945526123, "ce_loss": 3.2685241103172302, "z_loss": 114.75420951843262, "muon_lr": 0.0195, "adamw_lr": 0.00195, "grad_norm": 0.383748859167099, "tokens_per_second": 2989026.562979306, "eta_seconds": 9808.957169914245, "data_wait_seconds": 1.9371509552001953e-05}
67
+ {"step": 660, "total_steps": 28611, "tokens": 692060160, "loss": 3.3123960494995117, "ce_loss": 3.3010253310203552, "z_loss": 113.70717239379883, "muon_lr": 0.0198, "adamw_lr": 0.00198, "grad_norm": 0.3841683566570282, "tokens_per_second": 3104387.4250797406, "eta_seconds": 9441.072831058502, "data_wait_seconds": 0.2689374089241028}
68
+ {"step": 670, "total_steps": 28611, "tokens": 702545920, "loss": 3.292060375213623, "ce_loss": 3.280732214450836, "z_loss": 113.28091430664062, "muon_lr": 0.020099999999999996, "adamw_lr": 0.00201, "grad_norm": 0.37805303931236267, "tokens_per_second": 4171408.070654066, "eta_seconds": 7023.590480661392, "data_wait_seconds": 1.3887882232666016e-05}
69
+ {"step": 680, "total_steps": 28611, "tokens": 713031680, "loss": 3.2452685832977295, "ce_loss": 3.2336464524269104, "z_loss": 116.2209300994873, "muon_lr": 0.020399999999999998, "adamw_lr": 0.00204, "grad_norm": 0.39511507749557495, "tokens_per_second": 2938010.4462629007, "eta_seconds": 9968.57458190918, "data_wait_seconds": 0.3025229573249817}
70
+ {"step": 690, "total_steps": 28611, "tokens": 723517440, "loss": 3.258640766143799, "ce_loss": 3.247053802013397, "z_loss": 115.86975479125977, "muon_lr": 0.0207, "adamw_lr": 0.00207, "grad_norm": 0.36698827147483826, "tokens_per_second": 3393294.9348579906, "eta_seconds": 8627.982847952842, "data_wait_seconds": 1.4185905456542969e-05}
71
+ {"step": 700, "total_steps": 28611, "tokens": 734003200, "loss": 3.235472619533539, "ce_loss": 3.223949134349823, "z_loss": 115.23495292663574, "muon_lr": 0.021, "adamw_lr": 0.0021000000000000003, "grad_norm": 0.3743961453437805, "tokens_per_second": 2886434.9565392025, "eta_seconds": 10139.429842233658, "data_wait_seconds": 1.895427703857422e-05}
72
+ {"step": 710, "total_steps": 28611, "tokens": 744488960, "loss": 3.1973748207092285, "ce_loss": 3.185771644115448, "z_loss": 116.0313949584961, "muon_lr": 0.0213, "adamw_lr": 0.00213, "grad_norm": 0.337881863117218, "tokens_per_second": 4317810.792060535, "eta_seconds": 6775.729735493659, "data_wait_seconds": 1.9669532775878906e-05}
73
+ {"step": 720, "total_steps": 28611, "tokens": 754974720, "loss": 3.1727139353752136, "ce_loss": 3.1611295342445374, "z_loss": 115.84384536743164, "muon_lr": 0.021599999999999998, "adamw_lr": 0.00216, "grad_norm": 0.3610634207725525, "tokens_per_second": 2611167.1343052494, "eta_seconds": 11200.29155995846, "data_wait_seconds": 1.811981201171875e-05}
74
+ {"step": 730, "total_steps": 28611, "tokens": 765460480, "loss": 3.197407901287079, "ce_loss": 3.1856467723846436, "z_loss": 117.6107292175293, "muon_lr": 0.0219, "adamw_lr": 0.00219, "grad_norm": 0.32704517245292664, "tokens_per_second": 4876149.562391825, "eta_seconds": 5995.5805460691445, "data_wait_seconds": 1.3649463653564453e-05}
75
+ {"step": 740, "total_steps": 28611, "tokens": 775946240, "loss": 3.1888396739959717, "ce_loss": 3.1771335005760193, "z_loss": 117.06163024902344, "muon_lr": 0.0222, "adamw_lr": 0.00222, "grad_norm": 0.3705799877643585, "tokens_per_second": 3008495.119937335, "eta_seconds": 9714.113046860695, "data_wait_seconds": 1.8596649169921875e-05}
76
+ {"step": 750, "total_steps": 28611, "tokens": 786432000, "loss": 3.289639472961426, "ce_loss": 3.277992010116577, "z_loss": 116.47453498840332, "muon_lr": 0.0225, "adamw_lr": 0.00225, "grad_norm": 0.3481082320213318, "tokens_per_second": 4378446.831706549, "eta_seconds": 6672.314877605438, "data_wait_seconds": 1.811981201171875e-05}
77
+ {"step": 760, "total_steps": 28611, "tokens": 796917760, "loss": 3.1439366936683655, "ce_loss": 3.1319808959960938, "z_loss": 119.55829048156738, "muon_lr": 0.0228, "adamw_lr": 0.0022800000000000003, "grad_norm": 0.28526878356933594, "tokens_per_second": 3053391.5136219533, "eta_seconds": 9564.410605621339, "data_wait_seconds": 1.537799835205078e-05}
78
+ {"step": 770, "total_steps": 28611, "tokens": 807403520, "loss": 3.127338409423828, "ce_loss": 3.1155490279197693, "z_loss": 117.8934326171875, "muon_lr": 0.0231, "adamw_lr": 0.00231, "grad_norm": 0.3212087154388428, "tokens_per_second": 2916315.9528381396, "eta_seconds": 10010.370922803879, "data_wait_seconds": 1.913309097290039e-05}
79
+ {"step": 780, "total_steps": 28611, "tokens": 817889280, "loss": 3.1337884068489075, "ce_loss": 3.121930420398712, "z_loss": 118.57930374145508, "muon_lr": 0.023399999999999997, "adamw_lr": 0.00234, "grad_norm": 0.32087457180023193, "tokens_per_second": 4545692.888984666, "eta_seconds": 6419.905472874641, "data_wait_seconds": 1.3589859008789062e-05}
80
+ {"step": 790, "total_steps": 28611, "tokens": 828375040, "loss": 3.1538994908332825, "ce_loss": 3.142125368118286, "z_loss": 117.74141502380371, "muon_lr": 0.0237, "adamw_lr": 0.00237, "grad_norm": 0.3151688575744629, "tokens_per_second": 2819035.2419290585, "eta_seconds": 10348.374671626092, "data_wait_seconds": 0.3007614016532898}
81
+ {"step": 800, "total_steps": 28611, "tokens": 838860800, "loss": 3.123330235481262, "ce_loss": 3.1117360591888428, "z_loss": 115.94133567810059, "muon_lr": 0.024, "adamw_lr": 0.0024, "grad_norm": 0.3414774537086487, "tokens_per_second": 4332835.602157292, "eta_seconds": 6730.4531751632685, "data_wait_seconds": 1.436471939086914e-05}
82
+ {"step": 810, "total_steps": 28611, "tokens": 849346560, "loss": 3.1211427450180054, "ce_loss": 3.1091283559799194, "z_loss": 120.14446258544922, "muon_lr": 0.024300000000000002, "adamw_lr": 0.0024300000000000003, "grad_norm": 0.3174896538257599, "tokens_per_second": 2960920.511438067, "eta_seconds": 9845.404921674728, "data_wait_seconds": 1.4007091522216797e-05}
83
+ {"step": 820, "total_steps": 28611, "tokens": 859832320, "loss": 3.142742931842804, "ce_loss": 3.1310081481933594, "z_loss": 117.34791946411133, "muon_lr": 0.024599999999999997, "adamw_lr": 0.00246, "grad_norm": 0.338113933801651, "tokens_per_second": 2934101.7007762943, "eta_seconds": 9931.82192978859, "data_wait_seconds": 0.2962721586227417}
84
+ {"step": 830, "total_steps": 28611, "tokens": 870318080, "loss": 3.106025457382202, "ce_loss": 3.0943062901496887, "z_loss": 117.19226455688477, "muon_lr": 0.0249, "adamw_lr": 0.00249, "grad_norm": 0.2847413420677185, "tokens_per_second": 4947502.258979993, "eta_seconds": 5887.918454837799, "data_wait_seconds": 1.4066696166992188e-05}
85
+ {"step": 840, "total_steps": 28611, "tokens": 880803840, "loss": 3.1202743649482727, "ce_loss": 3.1088407039642334, "z_loss": 114.3368034362793, "muon_lr": 0.0252, "adamw_lr": 0.00252, "grad_norm": 0.39518630504608154, "tokens_per_second": 3025256.505129675, "eta_seconds": 9625.631428813933, "data_wait_seconds": 1.8417835235595703e-05}
86
+ {"step": 850, "total_steps": 28611, "tokens": 891289600, "loss": 3.1070111989974976, "ce_loss": 3.095278024673462, "z_loss": 117.33139419555664, "muon_lr": 0.0255, "adamw_lr": 0.00255, "grad_norm": 0.31667178869247437, "tokens_per_second": 4472919.606989316, "eta_seconds": 6507.945792388916, "data_wait_seconds": 1.7940998077392578e-05}
87
+ {"step": 860, "total_steps": 28611, "tokens": 901775360, "loss": 3.1141257882118225, "ce_loss": 3.102661907672882, "z_loss": 114.63846397399902, "muon_lr": 0.0258, "adamw_lr": 0.0025800000000000003, "grad_norm": 0.307193398475647, "tokens_per_second": 3136343.7287295316, "eta_seconds": 9278.011306428909, "data_wait_seconds": 1.806020736694336e-05}
88
+ {"step": 870, "total_steps": 28611, "tokens": 912261120, "loss": 3.125473976135254, "ce_loss": 3.1140242218971252, "z_loss": 114.4969654083252, "muon_lr": 0.026099999999999998, "adamw_lr": 0.00261, "grad_norm": 0.318877249956131, "tokens_per_second": 4727444.947404775, "eta_seconds": 6153.122276329995, "data_wait_seconds": 1.245737075805664e-05}
89
+ {"step": 880, "total_steps": 28611, "tokens": 922746880, "loss": 3.103769898414612, "ce_loss": 3.0922698974609375, "z_loss": 115.00003433227539, "muon_lr": 0.0264, "adamw_lr": 0.00264, "grad_norm": 0.2901608347892761, "tokens_per_second": 3179047.904090628, "eta_seconds": 9146.782915282249, "data_wait_seconds": 1.4603137969970703e-05}
90
+ {"step": 890, "total_steps": 28611, "tokens": 933232640, "loss": 3.112219989299774, "ce_loss": 3.100844442844391, "z_loss": 113.75528335571289, "muon_lr": 0.026699999999999998, "adamw_lr": 0.00267, "grad_norm": 0.3010977804660797, "tokens_per_second": 2872809.007319772, "eta_seconds": 10118.171873569489, "data_wait_seconds": 1.7642974853515625e-05}
91
+ {"step": 900, "total_steps": 28611, "tokens": 943718400, "loss": 3.0875620245933533, "ce_loss": 3.0759623646736145, "z_loss": 115.99650192260742, "muon_lr": 0.027, "adamw_lr": 0.0027, "grad_norm": 0.2923755645751953, "tokens_per_second": 3129632.899013479, "eta_seconds": 9284.504117131233, "data_wait_seconds": 0.2618265748023987}
92
+ {"step": 910, "total_steps": 28611, "tokens": 954204160, "loss": 3.0757806301116943, "ce_loss": 3.064554452896118, "z_loss": 112.26165390014648, "muon_lr": 0.0273, "adamw_lr": 0.00273, "grad_norm": 0.30451148748397827, "tokens_per_second": 3904326.815778178, "eta_seconds": 7439.593340039253, "data_wait_seconds": 1.9609928131103516e-05}
93
+ {"step": 920, "total_steps": 28611, "tokens": 964689920, "loss": 3.0603960156440735, "ce_loss": 3.0487239360809326, "z_loss": 116.72171401977539, "muon_lr": 0.0276, "adamw_lr": 0.0027600000000000003, "grad_norm": 0.32377398014068604, "tokens_per_second": 4586712.20281355, "eta_seconds": 6330.486137366295, "data_wait_seconds": 1.3232231140136719e-05}
94
+ {"step": 930, "total_steps": 28611, "tokens": 975175680, "loss": 3.067907392978668, "ce_loss": 3.056676745414734, "z_loss": 112.30597877502441, "muon_lr": 0.027899999999999998, "adamw_lr": 0.00279, "grad_norm": 0.30490797758102417, "tokens_per_second": 3126801.3388824584, "eta_seconds": 9282.851422333717, "data_wait_seconds": 1.4066696166992188e-05}
95
+ {"step": 940, "total_steps": 28611, "tokens": 985661440, "loss": 3.0771570205688477, "ce_loss": 3.0656723380088806, "z_loss": 114.8468189239502, "muon_lr": 0.0282, "adamw_lr": 0.00282, "grad_norm": 0.32569849491119385, "tokens_per_second": 2987776.1378422235, "eta_seconds": 9711.285302972794, "data_wait_seconds": 1.8358230590820312e-05}
96
+ {"step": 950, "total_steps": 28611, "tokens": 996147200, "loss": 3.0963059663772583, "ce_loss": 3.0853822827339172, "z_loss": 109.23702621459961, "muon_lr": 0.0285, "adamw_lr": 0.00285, "grad_norm": 0.30788472294807434, "tokens_per_second": 4605382.199332614, "eta_seconds": 6297.99210588932, "data_wait_seconds": 1.3768672943115234e-05}
97
+ {"step": 960, "total_steps": 28611, "tokens": 1006632960, "loss": 3.0548954010009766, "ce_loss": 3.043603837490082, "z_loss": 112.91586112976074, "muon_lr": 0.028799999999999996, "adamw_lr": 0.0028799999999999997, "grad_norm": 0.323307603597641, "tokens_per_second": 3283787.2306856294, "eta_seconds": 8829.49257645607, "data_wait_seconds": 1.9073486328125e-05}
98
+ {"step": 970, "total_steps": 28611, "tokens": 1017118720, "loss": 3.0637696385383606, "ce_loss": 3.0526286363601685, "z_loss": 111.40994453430176, "muon_lr": 0.029099999999999997, "adamw_lr": 0.0029100000000000003, "grad_norm": 0.3229497969150543, "tokens_per_second": 4526511.626984805, "eta_seconds": 6403.096159791947, "data_wait_seconds": 1.537799835205078e-05}
99
+ {"step": 980, "total_steps": 28611, "tokens": 1027604480, "loss": 3.072390377521515, "ce_loss": 3.0610828399658203, "z_loss": 113.07608604431152, "muon_lr": 0.0294, "adamw_lr": 0.00294, "grad_norm": 0.2861950993537903, "tokens_per_second": 3032353.386853322, "eta_seconds": 9554.69226694107, "data_wait_seconds": 1.4007091522216797e-05}
100
+ {"step": 990, "total_steps": 28611, "tokens": 1038090240, "loss": 3.027897536754608, "ce_loss": 3.016885280609131, "z_loss": 110.12204360961914, "muon_lr": 0.0297, "adamw_lr": 0.00297, "grad_norm": 0.28617510199546814, "tokens_per_second": 3040820.8325579828, "eta_seconds": 9524.638014149667, "data_wait_seconds": 0.5366204977035522}
101
+ {"step": 1000, "total_steps": 28611, "tokens": 1048576000, "loss": 3.1257240176200867, "ce_loss": 3.1150004863739014, "z_loss": 107.23563957214355, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.30688267946243286, "tokens_per_second": 4300609.325851903, "eta_seconds": 6732.123227739335, "data_wait_seconds": 1.3113021850585938e-05}
102
+ {"step": 1010, "total_steps": 28611, "tokens": 1059061760, "loss": 3.029353082180023, "ce_loss": 3.0183483362197876, "z_loss": 110.04720878601074, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.27792295813560486, "tokens_per_second": 2856398.3400227637, "eta_seconds": 10132.251433730125, "data_wait_seconds": 1.7404556274414062e-05}
103
+ {"step": 1020, "total_steps": 28611, "tokens": 1069547520, "loss": 3.0148768424987793, "ce_loss": 3.004064679145813, "z_loss": 108.12155151367188, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.30622598528862, "tokens_per_second": 3996502.335133982, "eta_seconds": 7239.14512989521, "data_wait_seconds": 1.5079975128173828e-05}
104
+ {"step": 1030, "total_steps": 28611, "tokens": 1080033280, "loss": 3.035790741443634, "ce_loss": 3.024936079978943, "z_loss": 108.54716682434082, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.275447815656662, "tokens_per_second": 2175742.807511626, "eta_seconds": 13292.368268966673, "data_wait_seconds": 0.27960848808288574}
105
+ {"step": 1040, "total_steps": 28611, "tokens": 1090519040, "loss": 3.0180326104164124, "ce_loss": 3.0069708228111267, "z_loss": 110.61770629882812, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2872465252876282, "tokens_per_second": 2734656.1088816873, "eta_seconds": 10571.818811917305, "data_wait_seconds": 1.8656253814697266e-05}
106
+ {"step": 1050, "total_steps": 28611, "tokens": 1101004800, "loss": 3.0059269666671753, "ce_loss": 2.9953866004943848, "z_loss": 105.40345001220703, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.28501302003860474, "tokens_per_second": 2790104.7052822886, "eta_seconds": 10357.963656806945, "data_wait_seconds": 0.29505276679992676}
107
+ {"step": 1060, "total_steps": 28611, "tokens": 1111490560, "loss": 3.0221720337867737, "ce_loss": 3.010904848575592, "z_loss": 112.67137145996094, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.23075471818447113, "tokens_per_second": 4088532.3468503854, "eta_seconds": 7065.938318490982, "data_wait_seconds": 1.7404556274414062e-05}
108
+ {"step": 1070, "total_steps": 28611, "tokens": 1121976320, "loss": 3.01272052526474, "ce_loss": 3.0020349621772766, "z_loss": 106.85589218139648, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.3040335178375244, "tokens_per_second": 2823678.0110833645, "eta_seconds": 10227.381274580956, "data_wait_seconds": 0.25865083932876587}
109
+ {"step": 1080, "total_steps": 28611, "tokens": 1132462080, "loss": 3.0016103982925415, "ce_loss": 2.9906504154205322, "z_loss": 109.60005378723145, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.26074299216270447, "tokens_per_second": 4398737.552773541, "eta_seconds": 6562.870712256431, "data_wait_seconds": 1.2516975402832031e-05}
110
+ {"step": 1090, "total_steps": 28611, "tokens": 1142947840, "loss": 2.9933687448501587, "ce_loss": 2.9823977947235107, "z_loss": 109.70959854125977, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.24118128418922424, "tokens_per_second": 2879741.9961167765, "eta_seconds": 10020.988038134574, "data_wait_seconds": 0.5813084840774536}
111
+ {"step": 1100, "total_steps": 28611, "tokens": 1153433600, "loss": 2.9651313424110413, "ce_loss": 2.954811990261078, "z_loss": 103.19386672973633, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2672273814678192, "tokens_per_second": 4069010.456454323, "eta_seconds": 7089.530647492409, "data_wait_seconds": 1.4483928680419922e-05}
112
+ {"step": 1110, "total_steps": 28611, "tokens": 1163919360, "loss": 2.978721857070923, "ce_loss": 2.968121588230133, "z_loss": 106.00293731689453, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2674412131309509, "tokens_per_second": 3021022.4748534136, "eta_seconds": 9545.40683362484, "data_wait_seconds": 1.722574234008789e-05}
113
+ {"step": 1120, "total_steps": 28611, "tokens": 1174405120, "loss": 2.9888700246810913, "ce_loss": 2.9784131050109863, "z_loss": 104.56944847106934, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.25692218542099, "tokens_per_second": 4450901.410443153, "eta_seconds": 6476.531416392327, "data_wait_seconds": 1.2993812561035156e-05}
114
+ {"step": 1130, "total_steps": 28611, "tokens": 1184890880, "loss": 2.9841817021369934, "ce_loss": 2.9736284017562866, "z_loss": 105.53267669677734, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2695963382720947, "tokens_per_second": 3043270.2444554665, "eta_seconds": 9468.734204101564, "data_wait_seconds": 1.800060272216797e-05}
115
+ {"step": 1140, "total_steps": 28611, "tokens": 1195376640, "loss": 2.960077166557312, "ce_loss": 2.9492769837379456, "z_loss": 108.00103759765625, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2331603318452835, "tokens_per_second": 4442722.528854158, "eta_seconds": 6483.734041213988, "data_wait_seconds": 1.3530254364013672e-05}
116
+ {"step": 1150, "total_steps": 28611, "tokens": 1205862400, "loss": 2.9679278135299683, "ce_loss": 2.9574100375175476, "z_loss": 105.17729759216309, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.25381389260292053, "tokens_per_second": 2867369.904476779, "eta_seconds": 10042.28491449356, "data_wait_seconds": 1.9729137420654297e-05}
117
+ {"step": 1160, "total_steps": 28611, "tokens": 1216348160, "loss": 2.989925980567932, "ce_loss": 2.979156732559204, "z_loss": 107.69272994995117, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2653111517429352, "tokens_per_second": 4449130.0880358005, "eta_seconds": 6469.68265850544, "data_wait_seconds": 1.6748905181884766e-05}
118
+ {"step": 1170, "total_steps": 28611, "tokens": 1226833920, "loss": 3.009878635406494, "ce_loss": 2.9996926188468933, "z_loss": 101.8600025177002, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2624979317188263, "tokens_per_second": 2929610.1728631575, "eta_seconds": 9821.775703310966, "data_wait_seconds": 2.2351741790771484e-05}
119
+ {"step": 1180, "total_steps": 28611, "tokens": 1237319680, "loss": 2.969376802444458, "ce_loss": 2.959559381008148, "z_loss": 98.17385482788086, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2652915418148041, "tokens_per_second": 4313039.098299966, "eta_seconds": 6668.960702753067, "data_wait_seconds": 1.245737075805664e-05}
120
+ {"step": 1190, "total_steps": 28611, "tokens": 1247805440, "loss": 2.9502323865890503, "ce_loss": 2.939701735973358, "z_loss": 105.30671691894531, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.26739370822906494, "tokens_per_second": 2779617.887545611, "eta_seconds": 10344.228472852707, "data_wait_seconds": 1.817941665649414e-05}
121
+ {"step": 1200, "total_steps": 28611, "tokens": 1258291200, "loss": 2.942854583263397, "ce_loss": 2.9325884580612183, "z_loss": 102.66136932373047, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.27831903100013733, "tokens_per_second": 3769699.1147677815, "eta_seconds": 7624.618268179894, "data_wait_seconds": 1.8298625946044922e-05}
122
+ {"step": 1210, "total_steps": 28611, "tokens": 1268776960, "loss": 2.959108591079712, "ce_loss": 2.949122130870819, "z_loss": 99.86493682861328, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.24137961864471436, "tokens_per_second": 2819692.0330719836, "eta_seconds": 10189.776273083686, "data_wait_seconds": 1.9848346710205078e-05}
123
+ {"step": 1220, "total_steps": 28611, "tokens": 1279262720, "loss": 2.9814881682395935, "ce_loss": 2.9711174368858337, "z_loss": 103.70760536193848, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2526809871196747, "tokens_per_second": 2733898.4601215585, "eta_seconds": 10505.710301589965, "data_wait_seconds": 1.4781951904296875e-05}
124
+ {"step": 1230, "total_steps": 28611, "tokens": 1289748480, "loss": 2.941792130470276, "ce_loss": 2.9322060346603394, "z_loss": 95.86140060424805, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.28192755579948425, "tokens_per_second": 2999778.402822463, "eta_seconds": 9571.06012530327, "data_wait_seconds": 1.4185905456542969e-05}
125
+ {"step": 1240, "total_steps": 28611, "tokens": 1300234240, "loss": 2.9343279600143433, "ce_loss": 2.924325168132782, "z_loss": 100.02786254882812, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.27617835998535156, "tokens_per_second": 2796754.597810847, "eta_seconds": 10262.099405670166, "data_wait_seconds": 0.5918394327163696}
126
+ {"step": 1250, "total_steps": 28611, "tokens": 1310720000, "loss": 2.948109269142151, "ce_loss": 2.9382757544517517, "z_loss": 98.33492088317871, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.26368746161460876, "tokens_per_second": 4186756.0077161915, "eta_seconds": 6852.581780052185, "data_wait_seconds": 1.3530254364013672e-05}
127
+ {"step": 1260, "total_steps": 28611, "tokens": 1321205760, "loss": 2.971496045589447, "ce_loss": 2.9613267183303833, "z_loss": 101.69274139404297, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.24531595408916473, "tokens_per_second": 2803992.4854714177, "eta_seconds": 10228.130897140503, "data_wait_seconds": 1.7702579498291016e-05}
128
+ {"step": 1270, "total_steps": 28611, "tokens": 1331691520, "loss": 2.955128490924835, "ce_loss": 2.94521701335907, "z_loss": 99.11442375183105, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.25932687520980835, "tokens_per_second": 4300904.980933149, "eta_seconds": 6665.833479952813, "data_wait_seconds": 1.3768672943115234e-05}
129
+ {"step": 1280, "total_steps": 28611, "tokens": 1342177280, "loss": 2.946077525615692, "ce_loss": 2.93592631816864, "z_loss": 101.51190757751465, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2806369960308075, "tokens_per_second": 4485472.862670307, "eta_seconds": 6389.21057677269, "data_wait_seconds": 1.627206802368164e-05}
130
+ {"step": 1290, "total_steps": 28611, "tokens": 1352663040, "loss": 2.944730043411255, "ce_loss": 2.9347782731056213, "z_loss": 99.51744651794434, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.29145917296409607, "tokens_per_second": 2909570.0961173396, "eta_seconds": 9846.177940249443, "data_wait_seconds": 1.341104507446289e-05}
131
+ {"step": 1300, "total_steps": 28611, "tokens": 1363148800, "loss": 2.9730359315872192, "ce_loss": 2.9634987711906433, "z_loss": 95.37228393554688, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.22789190709590912, "tokens_per_second": 2794456.7352741533, "eta_seconds": 10248.023801732064, "data_wait_seconds": 1.800060272216797e-05}
132
+ {"step": 1310, "total_steps": 28611, "tokens": 1373634560, "loss": 2.895432710647583, "ce_loss": 2.885974407196045, "z_loss": 94.58375930786133, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2789342701435089, "tokens_per_second": 4088078.961650079, "eta_seconds": 7002.597954821586, "data_wait_seconds": 1.2755393981933594e-05}
133
+ {"step": 1320, "total_steps": 28611, "tokens": 1384120320, "loss": 2.9810404181480408, "ce_loss": 2.971461296081543, "z_loss": 95.79132270812988, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2708542048931122, "tokens_per_second": 2891210.01322593, "eta_seconds": 9897.823916316032, "data_wait_seconds": 1.5556812286376953e-05}
134
+ {"step": 1330, "total_steps": 28611, "tokens": 1394606080, "loss": 2.9147441387176514, "ce_loss": 2.905186712741852, "z_loss": 95.57498550415039, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.20258770883083344, "tokens_per_second": 2046465.3994213375, "eta_seconds": 13978.34620809555, "data_wait_seconds": 0.2998175024986267}
135
+ {"step": 1340, "total_steps": 28611, "tokens": 1405091840, "loss": 2.9466036558151245, "ce_loss": 2.937487483024597, "z_loss": 91.1613998413086, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.25795140862464905, "tokens_per_second": 4046311.5712559614, "eta_seconds": 7067.106818747521, "data_wait_seconds": 1.3053417205810547e-05}
136
+ {"step": 1350, "total_steps": 28611, "tokens": 1415577600, "loss": 2.9266106486320496, "ce_loss": 2.917757511138916, "z_loss": 88.53121566772461, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2825232148170471, "tokens_per_second": 2754422.5369153223, "eta_seconds": 10377.939460229874, "data_wait_seconds": 1.8298625946044922e-05}
137
+ {"step": 1360, "total_steps": 28611, "tokens": 1426063360, "loss": 2.9384437799453735, "ce_loss": 2.929522156715393, "z_loss": 89.2162857055664, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2702580988407135, "tokens_per_second": 4215303.816969955, "eta_seconds": 6778.810215520858, "data_wait_seconds": 1.3709068298339844e-05}
138
+ {"step": 1370, "total_steps": 28611, "tokens": 1436549120, "loss": 2.938688635826111, "ce_loss": 2.9296507239341736, "z_loss": 90.37870979309082, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.24388106167316437, "tokens_per_second": 2671172.3310675984, "eta_seconds": 10693.529011130333, "data_wait_seconds": 1.5079975128173828e-05}
139
+ {"step": 1380, "total_steps": 28611, "tokens": 1447034880, "loss": 2.9090296626091003, "ce_loss": 2.899482548236847, "z_loss": 95.47085952758789, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2389918565750122, "tokens_per_second": 3220292.247659675, "eta_seconds": 8866.826629400253, "data_wait_seconds": 0.18508422374725342}
140
+ {"step": 1390, "total_steps": 28611, "tokens": 1457520640, "loss": 2.940493643283844, "ce_loss": 2.9312936663627625, "z_loss": 92.00053024291992, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2677419185638428, "tokens_per_second": 2714049.414618192, "eta_seconds": 10516.863525867462, "data_wait_seconds": 1.7344951629638672e-05}
141
+ {"step": 1400, "total_steps": 28611, "tokens": 1468006400, "loss": 2.94255667924881, "ce_loss": 2.9335808753967285, "z_loss": 89.75720596313477, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2594571113586426, "tokens_per_second": 4431091.8214718085, "eta_seconds": 6439.225970840454, "data_wait_seconds": 1.6868114471435547e-05}
142
+ {"step": 1410, "total_steps": 28611, "tokens": 1478492160, "loss": 2.9317322969436646, "ce_loss": 2.922861933708191, "z_loss": 88.7037582397461, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2193724364042282, "tokens_per_second": 4327516.644826616, "eta_seconds": 6590.9199471473685, "data_wait_seconds": 1.2934207916259766e-05}
143
+ {"step": 1420, "total_steps": 28611, "tokens": 1488977920, "loss": 2.9296798706054688, "ce_loss": 2.9210609793663025, "z_loss": 86.18958473205566, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.26279255747795105, "tokens_per_second": 2835304.04291772, "eta_seconds": 10056.004430007935, "data_wait_seconds": 1.4007091522216797e-05}
144
+ {"step": 1430, "total_steps": 28611, "tokens": 1499463680, "loss": 2.9069865942001343, "ce_loss": 2.898414969444275, "z_loss": 85.71658134460449, "muon_lr": 0.03, "adamw_lr": 0.003, "grad_norm": 0.2581104636192322, "tokens_per_second": 2746164.2062098994, "eta_seconds": 10378.601611495018, "data_wait_seconds": 1.8656253814697266e-05}