Commit ·
e433240
1
Parent(s): 90ba9cc
manual checkpoint step2838528 continue400B clean from* retention (prefer over sdpa-native) (#401)
Browse files- 94d93c0a0c017f1d94309132f1afc3e81ab8ecd74f695ccbc3184e7330839580 (6ddc2b8386aee582d44cd4d2b4439257ba7024e1)
- fc210536c5898110012b0d541c2c2cfc3a986d282cc28bcd4ab35b5e9b919bf4 (605c3b19c94638944ceb85699668c2dd930a4c62)
- a278cca755d7ca47d605f683eae9ddd0ffcf24a885f9e249917bbc55f7ded06e (154fd87cf73ce64ab3208dc453e30caa2abc9442)
- afd7bfab91e552e74a5201623f2f6b1630ad90f979077e91125028a199c58f67 (8b94f38e30b4e7170c74116346df58d10e431b3e)
- 27ee63ceb7b4015bb810e26f223a456836f2600945281e69861a703192c902c9 (aca3ac54d589c348ac08f4697ee6bf9e152dd924)
checkpoints/step2838528_20260911/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.provenance.json
ADDED
|
@@ -0,0 +1,583 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"batch_size": 56,
|
| 3 |
+
"block_size": 2048,
|
| 4 |
+
"checkpoint_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt",
|
| 5 |
+
"checkpoint_type": "full",
|
| 6 |
+
"created_at_iso": "2026-09-11T01:38:20Z",
|
| 7 |
+
"created_at_unix": 1789090700.3356085,
|
| 8 |
+
"created_from_scratch": false,
|
| 9 |
+
"dataset_provenance": {
|
| 10 |
+
"chat": false,
|
| 11 |
+
"dataset_field_text": "text",
|
| 12 |
+
"dataset_hotload_config_path": "/workspace/agillm43_dataset_hotload.json",
|
| 13 |
+
"dataset_hotload_config_sha256": "602506ab099079c84fc7b0f6a4258b79973168a2181739ca889b94915ae6cea1",
|
| 14 |
+
"dataset_hotload_config_stamp": [
|
| 15 |
+
50,
|
| 16 |
+
1321526073,
|
| 17 |
+
1789032636329904720,
|
| 18 |
+
4751
|
| 19 |
+
],
|
| 20 |
+
"dataset_hotload_config_used": true,
|
| 21 |
+
"dataset_hotload_enabled": true,
|
| 22 |
+
"dataset_hotload_generation": 12,
|
| 23 |
+
"dataset_hotload_schema": "agillm.dataset_hotload.config.v2",
|
| 24 |
+
"dataset_revision_map_sha256": "8361995a504a2365136b20e1a879ba78944a06309f13f3f013bbd0cdfc72225b",
|
| 25 |
+
"dataset_revisions": {
|
| 26 |
+
"EleutherAI/proof-pile-2": "901a9273a770e9d4138c5ddd91802f9c5c6cdc4b",
|
| 27 |
+
"HuggingFaceFW/fineweb": "9bb295ddab0e05d785b879661af7260fed5140fc",
|
| 28 |
+
"HuggingFaceFW/fineweb-edu": "87f09149ef4734204d70ed1d046ddc9ca3f2b8f9",
|
| 29 |
+
"HuggingFaceTB/cosmopedia": "0ae6ec63f91742bd2d1eaef4f02232c55d719385",
|
| 30 |
+
"HuggingFaceTB/finemath": "e92b25a616738fe95dc186b64dfb19f9c8525594",
|
| 31 |
+
"HuggingFaceTB/smollm-corpus": "3ba9d605774198c5868892d7a8deda78031a781f",
|
| 32 |
+
"Skylion007/openwebtext": "79d93d786212f7344586290adb811d4ae6a1762c",
|
| 33 |
+
"allenai/c4": "1588ec454efa1a09f29cd18ddd04fe05fc8653a2",
|
| 34 |
+
"allenai/dolma": "7f48140530a023e9ea4c5cfb141160922727d4d3",
|
| 35 |
+
"allenai/dolmino-mix-1124": "a319f19eef1e257417b11ea8c30da266ae175557",
|
| 36 |
+
"codeparrot/codeparrot-clean": "35a59fb025bc0a102f7d96eac09d145b896d487b",
|
| 37 |
+
"mlfoundations/dclm-baseline-1.0": "a3b142c183aebe5af344955ae20836eb34dcf69b",
|
| 38 |
+
"wikimedia/wikipedia": "b04c8d1ceb2f5cd4588862100d08de323dccfbaa"
|
| 39 |
+
},
|
| 40 |
+
"dolma_payload_integrity": "external_url_manifest_pinned_payload_hashes_unavailable",
|
| 41 |
+
"dolma_url_list_sha256": "bc6ef505be920c94bb9e7147445638f606fdc3f5af622cd1d887dc48e21a2d5b",
|
| 42 |
+
"dolma_url_manifest_raw_sha256": "6fe8c40d51e5bebc9303453d7f533b232d69f7ae122123621a82621b1cd779b7",
|
| 43 |
+
"has_language_mix": true,
|
| 44 |
+
"has_numeracy": false,
|
| 45 |
+
"hot_config_expected_sha256": "",
|
| 46 |
+
"hot_config_has_append": true,
|
| 47 |
+
"hot_config_has_datasets": false,
|
| 48 |
+
"hot_config_mtime": 1789079605.4010906,
|
| 49 |
+
"hot_config_path": "/workspace/agillm43_production_v14/contracts/hot_config_owner_v25.json",
|
| 50 |
+
"hot_config_used": true,
|
| 51 |
+
"hot_reload_capable": true,
|
| 52 |
+
"phase": "pretrain",
|
| 53 |
+
"proof_pile_lfs_manifest_sha256": "d020dfb796b02a75564e4e11d9e03185a55b3f9c6639f87830cfbe10737ed329",
|
| 54 |
+
"proof_pile_paths_sha256": "8a30018017e3b3d9a9bcfa06a3c404ccd113297b0b1d0c48e2e2aadf113e52b0",
|
| 55 |
+
"schema": "agillm.dataset_provenance.v1",
|
| 56 |
+
"sequence_prefetch_depth": 256,
|
| 57 |
+
"source_count": 16,
|
| 58 |
+
"source_effective": "HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.96,wikimedia/wikipedia:20231101.en|1.63,allenai/c4:en|0.36,Skylion007/openwebtext|0.18,HuggingFaceTB/cosmopedia:web_samples_v2|0.64,HuggingFaceTB/smollm-corpus:cosmopedia-v2|0.65,EleutherAI/proof-pile-2:all|0.67,allenai/dolma:v1_6-sample|0.5,codeparrot/codeparrot-clean|0.24,HuggingFaceTB/finemath:finemath-4plus|0.32,HuggingFaceTB/finemath:infiwebmath-4plus|0.23,HuggingFaceTB/smollm-corpus:fineweb-edu-dedup|1.05,allenai/dolmino-mix-1124|0.78,mlfoundations/dclm-baseline-1.0|1.05,HuggingFaceFW/fineweb-edu:sample-100BT|2.4,json:/workspace/agillm_wikitext103_raw_docs/train.jsonl|0.03",
|
| 59 |
+
"source_requested": "HuggingFaceFW/fineweb-edu:sample-10BT|1.60,HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.50,wikimedia/wikipedia:20231101.en|1.10,allenai/c4:en|0.40,Skylion007/openwebtext|0.35,HuggingFaceTB/cosmopedia:web_samples_v2|1.30,HuggingFaceTB/smollm-corpus:cosmopedia-v2|1.30,EleutherAI/proof-pile-2:all|1.50,allenai/dolma:v1_6-sample|0.50,codeparrot/codeparrot-clean|1.25",
|
| 60 |
+
"source_specs": [
|
| 61 |
+
"HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.96",
|
| 62 |
+
"wikimedia/wikipedia:20231101.en|1.63",
|
| 63 |
+
"allenai/c4:en|0.36",
|
| 64 |
+
"Skylion007/openwebtext|0.18",
|
| 65 |
+
"HuggingFaceTB/cosmopedia:web_samples_v2|0.64",
|
| 66 |
+
"HuggingFaceTB/smollm-corpus:cosmopedia-v2|0.65",
|
| 67 |
+
"EleutherAI/proof-pile-2:all|0.67",
|
| 68 |
+
"allenai/dolma:v1_6-sample|0.5",
|
| 69 |
+
"codeparrot/codeparrot-clean|0.24",
|
| 70 |
+
"HuggingFaceTB/finemath:finemath-4plus|0.32",
|
| 71 |
+
"HuggingFaceTB/finemath:infiwebmath-4plus|0.23",
|
| 72 |
+
"HuggingFaceTB/smollm-corpus:fineweb-edu-dedup|1.05",
|
| 73 |
+
"allenai/dolmino-mix-1124|0.78",
|
| 74 |
+
"mlfoundations/dclm-baseline-1.0|1.05",
|
| 75 |
+
"HuggingFaceFW/fineweb-edu:sample-100BT|2.4",
|
| 76 |
+
"json:/workspace/agillm_wikitext103_raw_docs/train.jsonl|0.03"
|
| 77 |
+
],
|
| 78 |
+
"val_source_effective": "HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.96,wikimedia/wikipedia:20231101.en|1.63,allenai/c4:en|0.36,Skylion007/openwebtext|0.18,HuggingFaceTB/cosmopedia:web_samples_v2|0.64,HuggingFaceTB/smollm-corpus:cosmopedia-v2|0.65,EleutherAI/proof-pile-2:all|0.67,allenai/dolma:v1_6-sample|0.5,codeparrot/codeparrot-clean|0.24,HuggingFaceTB/finemath:finemath-4plus|0.32,HuggingFaceTB/finemath:infiwebmath-4plus|0.23,HuggingFaceTB/smollm-corpus:fineweb-edu-dedup|1.05,allenai/dolmino-mix-1124|0.78,mlfoundations/dclm-baseline-1.0|1.05,HuggingFaceFW/fineweb-edu:sample-100BT|2.4,json:/workspace/agillm_wikitext103_raw_docs/train.jsonl|0.03",
|
| 79 |
+
"val_source_requested": ""
|
| 80 |
+
},
|
| 81 |
+
"effective_global_step": 5676611,
|
| 82 |
+
"effective_seen_tok": 282430494720,
|
| 83 |
+
"global_origin_seen_tok": 0,
|
| 84 |
+
"global_origin_step": 2838083,
|
| 85 |
+
"gpu": {
|
| 86 |
+
"allocated_gb": 7.13,
|
| 87 |
+
"peak_allocated_gb": 11.81,
|
| 88 |
+
"reserved_gb": 13.34
|
| 89 |
+
},
|
| 90 |
+
"hostname": "ead6505f92c4",
|
| 91 |
+
"lane": "pretrain",
|
| 92 |
+
"launch": {
|
| 93 |
+
"argv": [
|
| 94 |
+
"/workspace/agillm44_throughput_80k_20260910/runtime_retain_activations_v2.py",
|
| 95 |
+
"train",
|
| 96 |
+
"--preset",
|
| 97 |
+
"agillm4_floor",
|
| 98 |
+
"--block",
|
| 99 |
+
"2048",
|
| 100 |
+
"--source",
|
| 101 |
+
"HuggingFaceFW/fineweb-edu:sample-10BT|1.60,HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.50,wikimedia/wikipedia:20231101.en|1.10,allenai/c4:en|0.40,Skylion007/openwebtext|0.35,HuggingFaceTB/cosmopedia:web_samples_v2|1.30,HuggingFaceTB/smollm-corpus:cosmopedia-v2|1.30,EleutherAI/proof-pile-2:all|1.50,allenai/dolma:v1_6-sample|0.50,codeparrot/codeparrot-clean|1.25",
|
| 102 |
+
"--nat_mask_token_id",
|
| 103 |
+
"2",
|
| 104 |
+
"--tie_weights",
|
| 105 |
+
"--tie_kv",
|
| 106 |
+
"--moe_ffn",
|
| 107 |
+
"--moe_experts",
|
| 108 |
+
"2",
|
| 109 |
+
"--moe_top_k",
|
| 110 |
+
"1",
|
| 111 |
+
"--moe_mlp_mult",
|
| 112 |
+
"4",
|
| 113 |
+
"--moe_shared_experts",
|
| 114 |
+
"1",
|
| 115 |
+
"--moe_shared_mlp_mult",
|
| 116 |
+
"2",
|
| 117 |
+
"--moe_aux_coef",
|
| 118 |
+
"0.01",
|
| 119 |
+
"--moe_z_coef",
|
| 120 |
+
"0.001",
|
| 121 |
+
"--amp",
|
| 122 |
+
"--grad_checkpoint",
|
| 123 |
+
"--attn_backend",
|
| 124 |
+
"sdpa",
|
| 125 |
+
"--sublinear_window",
|
| 126 |
+
"128",
|
| 127 |
+
"--sublinear_stride",
|
| 128 |
+
"128",
|
| 129 |
+
"--sublinear_max_anchors",
|
| 130 |
+
"128",
|
| 131 |
+
"--sublinear_chunk",
|
| 132 |
+
"128",
|
| 133 |
+
"--sublinear_sinks",
|
| 134 |
+
"4",
|
| 135 |
+
"--sublinear_recent_anchors",
|
| 136 |
+
"64",
|
| 137 |
+
"--no-sublinear_pooled_landmarks",
|
| 138 |
+
"--anchor_stride",
|
| 139 |
+
"256",
|
| 140 |
+
"--anchor_max",
|
| 141 |
+
"2048",
|
| 142 |
+
"--anchor_position",
|
| 143 |
+
"-1",
|
| 144 |
+
"--alibi_mode",
|
| 145 |
+
"corrected",
|
| 146 |
+
"--alibi_scale",
|
| 147 |
+
"0.0",
|
| 148 |
+
"--optimizer",
|
| 149 |
+
"adamw8bit",
|
| 150 |
+
"--weight_decay",
|
| 151 |
+
"0.0",
|
| 152 |
+
"--lr_decay",
|
| 153 |
+
"cosine",
|
| 154 |
+
"--lr_decay_tokens",
|
| 155 |
+
"49000954880",
|
| 156 |
+
"--lr_min_mult",
|
| 157 |
+
"0.2",
|
| 158 |
+
"--lr_schedule_reset_on_resume",
|
| 159 |
+
"--lr_warmup_tokens",
|
| 160 |
+
"10000000",
|
| 161 |
+
"--lr_warmup_min_mult",
|
| 162 |
+
"0.1",
|
| 163 |
+
"--dblock",
|
| 164 |
+
"--dblock_schedule",
|
| 165 |
+
"roundrobin",
|
| 166 |
+
"--dblock_router_ramp_steps",
|
| 167 |
+
"64",
|
| 168 |
+
"--dblock_warmup_steps",
|
| 169 |
+
"14",
|
| 170 |
+
"--dblock_loop_layers",
|
| 171 |
+
"0",
|
| 172 |
+
"--dblock_loop_start",
|
| 173 |
+
"0",
|
| 174 |
+
"--dblock_loop_cond_scale",
|
| 175 |
+
"1.0",
|
| 176 |
+
"--dblock_sigma_curriculum_steps",
|
| 177 |
+
"2000",
|
| 178 |
+
"--dblock_sigma_sampling",
|
| 179 |
+
"lognormal",
|
| 180 |
+
"--dblock_sigma_stratified",
|
| 181 |
+
"--dblock_sigma_min",
|
| 182 |
+
"0.002",
|
| 183 |
+
"--dblock_sigma_max",
|
| 184 |
+
"2.0",
|
| 185 |
+
"--dblock_sigma_pmean",
|
| 186 |
+
"-1.2",
|
| 187 |
+
"--dblock_sigma_pstd",
|
| 188 |
+
"1.2",
|
| 189 |
+
"--dblock_edm_wmax",
|
| 190 |
+
"5.0",
|
| 191 |
+
"--dblock_ar_prob",
|
| 192 |
+
"0.8",
|
| 193 |
+
"--dblock_sat_prob",
|
| 194 |
+
"0.1",
|
| 195 |
+
"--dblock_nat_prob",
|
| 196 |
+
"0.1",
|
| 197 |
+
"--dblock_ar_loss_tokens",
|
| 198 |
+
"8192",
|
| 199 |
+
"--dblock_sat_loss_tokens",
|
| 200 |
+
"4096",
|
| 201 |
+
"--dblock_nat_loss_tokens",
|
| 202 |
+
"4096",
|
| 203 |
+
"--dblock_nat_embed_noise_mode",
|
| 204 |
+
"off",
|
| 205 |
+
"--dblock_nat_embed_noise_scale",
|
| 206 |
+
"1.0",
|
| 207 |
+
"--nat_loss_weight",
|
| 208 |
+
"1.0",
|
| 209 |
+
"--nat_mask_ratio",
|
| 210 |
+
"0.5",
|
| 211 |
+
"--nat_max_tokens",
|
| 212 |
+
"0",
|
| 213 |
+
"--nat_span_mask_prob",
|
| 214 |
+
"0.35",
|
| 215 |
+
"--nat_suffix_mask_prob",
|
| 216 |
+
"0.20",
|
| 217 |
+
"--nat_span_max_tokens",
|
| 218 |
+
"0",
|
| 219 |
+
"--loss_spike_skip",
|
| 220 |
+
"0.0",
|
| 221 |
+
"--sat_every",
|
| 222 |
+
"1",
|
| 223 |
+
"--nat_every",
|
| 224 |
+
"1",
|
| 225 |
+
"--dblock_fullstack_anchor_spike_skip",
|
| 226 |
+
"0.0",
|
| 227 |
+
"--dblock_fullstack_anchor_max_ce",
|
| 228 |
+
"0.0",
|
| 229 |
+
"--dblock_anchor_softcap_mult",
|
| 230 |
+
"4.0",
|
| 231 |
+
"--dblock_anchor_softcap_min_ce",
|
| 232 |
+
"25.0",
|
| 233 |
+
"--dblock_anchor_softcap_max_ce",
|
| 234 |
+
"50.0",
|
| 235 |
+
"--dblock_fullstack_ar_every",
|
| 236 |
+
"3",
|
| 237 |
+
"--dblock_fullstack_ar_offset",
|
| 238 |
+
"0",
|
| 239 |
+
"--dblock_fullstack_ar_tokens",
|
| 240 |
+
"256",
|
| 241 |
+
"--dblock_fullstack_sat_every",
|
| 242 |
+
"3",
|
| 243 |
+
"--dblock_fullstack_sat_offset",
|
| 244 |
+
"1",
|
| 245 |
+
"--dblock_fullstack_sat_tokens",
|
| 246 |
+
"128",
|
| 247 |
+
"--dblock_fullstack_nat_every",
|
| 248 |
+
"3",
|
| 249 |
+
"--dblock_fullstack_nat_offset",
|
| 250 |
+
"2",
|
| 251 |
+
"--dblock_fullstack_nat_tokens",
|
| 252 |
+
"128",
|
| 253 |
+
"--dblock_fullstack_nat_mask_id",
|
| 254 |
+
"-1",
|
| 255 |
+
"--ckpt_codec",
|
| 256 |
+
"block-sharded-zstd",
|
| 257 |
+
"--async_update_every_steps",
|
| 258 |
+
"0",
|
| 259 |
+
"--dblock_router_hidden",
|
| 260 |
+
"64",
|
| 261 |
+
"--dblock_router_heads",
|
| 262 |
+
"4",
|
| 263 |
+
"--dblock_router_layers",
|
| 264 |
+
"2",
|
| 265 |
+
"--dblock_router_lr",
|
| 266 |
+
"0.0005",
|
| 267 |
+
"--dblock_explore",
|
| 268 |
+
"0.08",
|
| 269 |
+
"--dblock_max_stale_steps",
|
| 270 |
+
"64",
|
| 271 |
+
"--dblock_max_count_skew",
|
| 272 |
+
"1.35",
|
| 273 |
+
"--dblock_stale_bonus",
|
| 274 |
+
"0.35",
|
| 275 |
+
"--dblock_undertrain_bonus",
|
| 276 |
+
"0.25",
|
| 277 |
+
"--dblock_activation_offload_min_mb",
|
| 278 |
+
"1.0",
|
| 279 |
+
"--nat_document_boundary_aware",
|
| 280 |
+
"--dblock_fullstack_anchor_deterministic_eval",
|
| 281 |
+
"--dblock_fullstack_anchor_spike_retry_limit",
|
| 282 |
+
"3",
|
| 283 |
+
"--dblock_local_spike_retry_limit",
|
| 284 |
+
"3",
|
| 285 |
+
"--dblock_fullstack_nat_no_valid_crop_limit",
|
| 286 |
+
"8",
|
| 287 |
+
"--target_tokens",
|
| 288 |
+
"400000000000",
|
| 289 |
+
"--lr_schedule_reanchor_on_resume",
|
| 290 |
+
"--save_every_sec",
|
| 291 |
+
"1800",
|
| 292 |
+
"--delta_max_keep",
|
| 293 |
+
"0",
|
| 294 |
+
"--max_ckpts",
|
| 295 |
+
"2",
|
| 296 |
+
"--disk_free_floor_gb",
|
| 297 |
+
"12.0",
|
| 298 |
+
"--heartbeat_every_sec",
|
| 299 |
+
"30",
|
| 300 |
+
"--dblock_fullstack_ar_weight",
|
| 301 |
+
"1.0",
|
| 302 |
+
"--dblock_fullstack_sat_weight",
|
| 303 |
+
"0.1",
|
| 304 |
+
"--dblock_fullstack_nat_weight",
|
| 305 |
+
"0.1",
|
| 306 |
+
"--dblock_local_aux_warn_ce",
|
| 307 |
+
"0.0",
|
| 308 |
+
"--dblock_local_aux_max_ce",
|
| 309 |
+
"0.0",
|
| 310 |
+
"--lr_schedule_origin_tokens",
|
| 311 |
+
"85002387456",
|
| 312 |
+
"--dblock_stop_after_commits",
|
| 313 |
+
"0",
|
| 314 |
+
"--no-oom_auto_backoff",
|
| 315 |
+
"--require_exact_resume_state",
|
| 316 |
+
"--dblock_objective_mode",
|
| 317 |
+
"committed_9_1_1",
|
| 318 |
+
"--dblock_ar_weight",
|
| 319 |
+
"0.25",
|
| 320 |
+
"--dblock_sat_weight",
|
| 321 |
+
"0.25",
|
| 322 |
+
"--dblock_nat_weight",
|
| 323 |
+
"0.25",
|
| 324 |
+
"--dblock_aux_grad_ratio",
|
| 325 |
+
"0.5",
|
| 326 |
+
"--dblock_min_ar_share",
|
| 327 |
+
"0.8",
|
| 328 |
+
"--lr_core",
|
| 329 |
+
"2e-7",
|
| 330 |
+
"--lr_head",
|
| 331 |
+
"2e-7",
|
| 332 |
+
"--dblock_checkpoint_stride",
|
| 333 |
+
"2",
|
| 334 |
+
"--empty_cache_every_steps",
|
| 335 |
+
"0",
|
| 336 |
+
"--dblock_log_every",
|
| 337 |
+
"25",
|
| 338 |
+
"--cuda_max_reserved_mib",
|
| 339 |
+
"0",
|
| 340 |
+
"--cuda_min_free_mib",
|
| 341 |
+
"0",
|
| 342 |
+
"--dblock_clean_context",
|
| 343 |
+
"e2e",
|
| 344 |
+
"--dblock_local_rows",
|
| 345 |
+
"2",
|
| 346 |
+
"--dblock_clean_context_noise_scale",
|
| 347 |
+
"0.0",
|
| 348 |
+
"--dblock_depth_anchor_every",
|
| 349 |
+
"50",
|
| 350 |
+
"--dblock_fullstack_anchor_cap",
|
| 351 |
+
"256",
|
| 352 |
+
"--dblock_fullstack_ar_rows",
|
| 353 |
+
"1",
|
| 354 |
+
"--detachable_50m_upgrade_multimode",
|
| 355 |
+
"--detachable_50m_upgrade_logit_distill",
|
| 356 |
+
"--detachable_50m_upgrade_sat_gate_v2713",
|
| 357 |
+
"--detachable_50m_create_from_legacy",
|
| 358 |
+
"--detachable_50m_seed_from_mature",
|
| 359 |
+
"--detachable_50m_lr",
|
| 360 |
+
"0.0001",
|
| 361 |
+
"--detachable_50m_weight_decay",
|
| 362 |
+
"0.0",
|
| 363 |
+
"--detachable_50m_update_mode",
|
| 364 |
+
"sublayer",
|
| 365 |
+
"--detachable_50m_batch",
|
| 366 |
+
"1",
|
| 367 |
+
"--detachable_50m_tokens",
|
| 368 |
+
"64",
|
| 369 |
+
"--detachable_50m_loss_tokens",
|
| 370 |
+
"64",
|
| 371 |
+
"--detachable_50m_every",
|
| 372 |
+
"1",
|
| 373 |
+
"--detachable_50m_teacher_every",
|
| 374 |
+
"1",
|
| 375 |
+
"--detachable_50m_distill_weight",
|
| 376 |
+
"1.0",
|
| 377 |
+
"--detachable_50m_logit_distill_alpha",
|
| 378 |
+
"0.30",
|
| 379 |
+
"--detachable_50m_logit_distill_temperature",
|
| 380 |
+
"1.0",
|
| 381 |
+
"--detachable_50m_logit_distill_tokens",
|
| 382 |
+
"64",
|
| 383 |
+
"--detachable_50m_hidden_distill_weight",
|
| 384 |
+
"0.0",
|
| 385 |
+
"--detachable_50m_anchor_every",
|
| 386 |
+
"128",
|
| 387 |
+
"--detachable_50m_anchor_tokens",
|
| 388 |
+
"32",
|
| 389 |
+
"--detachable_50m_seed",
|
| 390 |
+
"20260830",
|
| 391 |
+
"--detachable_50m_log_every",
|
| 392 |
+
"10",
|
| 393 |
+
"--detachable_50m_telemetry_keep",
|
| 394 |
+
"128",
|
| 395 |
+
"--weight_accel_mode",
|
| 396 |
+
"shadow",
|
| 397 |
+
"--weight_accel_dir",
|
| 398 |
+
"/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/weight_trajectory_accel",
|
| 399 |
+
"--weight_accel_every",
|
| 400 |
+
"1",
|
| 401 |
+
"--weight_accel_train_every",
|
| 402 |
+
"4",
|
| 403 |
+
"--weight_accel_record_every",
|
| 404 |
+
"14",
|
| 405 |
+
"--weight_accel_state_every",
|
| 406 |
+
"256",
|
| 407 |
+
"--detachable_50m",
|
| 408 |
+
"--detachable_50m_optimizer",
|
| 409 |
+
"paged_adamw8bit",
|
| 410 |
+
"--delta_every_steps",
|
| 411 |
+
"0",
|
| 412 |
+
"--delta_every_sec",
|
| 413 |
+
"0",
|
| 414 |
+
"--resume",
|
| 415 |
+
"/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838083_from02835222_20260911T010911_1789088951909741831_3003552_000003.pt",
|
| 416 |
+
"--save_dir",
|
| 417 |
+
"/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B",
|
| 418 |
+
"--data_seed",
|
| 419 |
+
"2838125",
|
| 420 |
+
"--batch_size",
|
| 421 |
+
"56",
|
| 422 |
+
"--dblock_blocks",
|
| 423 |
+
"14",
|
| 424 |
+
"--dblock_checkpoint_skip_tail",
|
| 425 |
+
"2",
|
| 426 |
+
"--dblock_direct56_enabled",
|
| 427 |
+
"1",
|
| 428 |
+
"--dblock_layers_per_update",
|
| 429 |
+
"0",
|
| 430 |
+
"--dblock_router",
|
| 431 |
+
"shadow",
|
| 432 |
+
"--dblock_router_blend",
|
| 433 |
+
"0.0",
|
| 434 |
+
"--dblock_router_full_anchor_every",
|
| 435 |
+
"128",
|
| 436 |
+
"--dblock_router_train_layer_policy",
|
| 437 |
+
"cyclic",
|
| 438 |
+
"--dblock_router_train_layers_per_update",
|
| 439 |
+
"1",
|
| 440 |
+
"--dblock_sublayer_mode",
|
| 441 |
+
"off",
|
| 442 |
+
"--dblock_target_route_activate_from_checkpoint",
|
| 443 |
+
"1",
|
| 444 |
+
"--dblock_target_route_activation_checkpoint_step",
|
| 445 |
+
"2717509",
|
| 446 |
+
"--dblock_target_route_activation_parent_pointer_sha256",
|
| 447 |
+
"04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827",
|
| 448 |
+
"--dblock_target_route_offset",
|
| 449 |
+
"1",
|
| 450 |
+
"--dblock_target_route_policy",
|
| 451 |
+
"xor_fold56_v1",
|
| 452 |
+
"--dblock_train_layer_policy",
|
| 453 |
+
"cyclic",
|
| 454 |
+
"--dblock_train_layers_per_update",
|
| 455 |
+
"1",
|
| 456 |
+
"--dblock_train_sublayer_policy",
|
| 457 |
+
"cyclic",
|
| 458 |
+
"--dblock_train_sublayers_per_update",
|
| 459 |
+
"1",
|
| 460 |
+
"--granularity_label",
|
| 461 |
+
"one-sublayer-local-grad-direct56-teacher-v27.16",
|
| 462 |
+
"--oom_known_safe_batch",
|
| 463 |
+
"56",
|
| 464 |
+
"--val_every_sec",
|
| 465 |
+
"600",
|
| 466 |
+
"--val_file",
|
| 467 |
+
"/workspace/agillm43_heldout_v3/token_ids.json",
|
| 468 |
+
"--val_sha256",
|
| 469 |
+
"513b30203ffd18167655deb2d9cd610541df87d1ea0b8a0f9241c58e68703d21"
|
| 470 |
+
],
|
| 471 |
+
"argv_string": "/workspace/agillm44_throughput_80k_20260910/runtime_retain_activations_v2.py train --preset agillm4_floor --block 2048 --source HuggingFaceFW/fineweb-edu:sample-10BT|1.60,HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.50,wikimedia/wikipedia:20231101.en|1.10,allenai/c4:en|0.40,Skylion007/openwebtext|0.35,HuggingFaceTB/cosmopedia:web_samples_v2|1.30,HuggingFaceTB/smollm-corpus:cosmopedia-v2|1.30,EleutherAI/proof-pile-2:all|1.50,allenai/dolma:v1_6-sample|0.50,codeparrot/codeparrot-clean|1.25 --nat_mask_token_id 2 --tie_weights --tie_kv --moe_ffn --moe_experts 2 --moe_top_k 1 --moe_mlp_mult 4 --moe_shared_experts 1 --moe_shared_mlp_mult 2 --moe_aux_coef 0.01 --moe_z_coef 0.001 --amp --grad_checkpoint --attn_backend sdpa --sublinear_window 128 --sublinear_stride 128 --sublinear_max_anchors 128 --sublinear_chunk 128 --sublinear_sinks 4 --sublinear_recent_anchors 64 --no-sublinear_pooled_landmarks --anchor_stride 256 --anchor_max 2048 --anchor_position -1 --alibi_mode corrected --alibi_scale 0.0 --optimizer adamw8bit --weight_decay 0.0 --lr_decay cosine --lr_decay_tokens 49000954880 --lr_min_mult 0.2 --lr_schedule_reset_on_resume --lr_warmup_tokens 10000000 --lr_warmup_min_mult 0.1 --dblock --dblock_schedule roundrobin --dblock_router_ramp_steps 64 --dblock_warmup_steps 14 --dblock_loop_layers 0 --dblock_loop_start 0 --dblock_loop_cond_scale 1.0 --dblock_sigma_curriculum_steps 2000 --dblock_sigma_sampling lognormal --dblock_sigma_stratified --dblock_sigma_min 0.002 --dblock_sigma_max 2.0 --dblock_sigma_pmean -1.2 --dblock_sigma_pstd 1.2 --dblock_edm_wmax 5.0 --dblock_ar_prob 0.8 --dblock_sat_prob 0.1 --dblock_nat_prob 0.1 --dblock_ar_loss_tokens 8192 --dblock_sat_loss_tokens 4096 --dblock_nat_loss_tokens 4096 --dblock_nat_embed_noise_mode off --dblock_nat_embed_noise_scale 1.0 --nat_loss_weight 1.0 --nat_mask_ratio 0.5 --nat_max_tokens 0 --nat_span_mask_prob 0.35 --nat_suffix_mask_prob 0.20 --nat_span_max_tokens 0 --loss_spike_skip 0.0 --sat_every 1 --nat_every 1 --dblock_fullstack_anchor_spike_skip 0.0 --dblock_fullstack_anchor_max_ce 0.0 --dblock_anchor_softcap_mult 4.0 --dblock_anchor_softcap_min_ce 25.0 --dblock_anchor_softcap_max_ce 50.0 --dblock_fullstack_ar_every 3 --dblock_fullstack_ar_offset 0 --dblock_fullstack_ar_tokens 256 --dblock_fullstack_sat_every 3 --dblock_fullstack_sat_offset 1 --dblock_fullstack_sat_tokens 128 --dblock_fullstack_nat_every 3 --dblock_fullstack_nat_offset 2 --dblock_fullstack_nat_tokens 128 --dblock_fullstack_nat_mask_id -1 --ckpt_codec block-sharded-zstd --async_update_every_steps 0 --dblock_router_hidden 64 --dblock_router_heads 4 --dblock_router_layers 2 --dblock_router_lr 0.0005 --dblock_explore 0.08 --dblock_max_stale_steps 64 --dblock_max_count_skew 1.35 --dblock_stale_bonus 0.35 --dblock_undertrain_bonus 0.25 --dblock_activation_offload_min_mb 1.0 --nat_document_boundary_aware --dblock_fullstack_anchor_deterministic_eval --dblock_fullstack_anchor_spike_retry_limit 3 --dblock_local_spike_retry_limit 3 --dblock_fullstack_nat_no_valid_crop_limit 8 --target_tokens 400000000000 --lr_schedule_reanchor_on_resume --save_every_sec 1800 --delta_max_keep 0 --max_ckpts 2 --disk_free_floor_gb 12.0 --heartbeat_every_sec 30 --dblock_fullstack_ar_weight 1.0 --dblock_fullstack_sat_weight 0.1 --dblock_fullstack_nat_weight 0.1 --dblock_local_aux_warn_ce 0.0 --dblock_local_aux_max_ce 0.0 --lr_schedule_origin_tokens 85002387456 --dblock_stop_after_commits 0 --no-oom_auto_backoff --require_exact_resume_state --dblock_objective_mode committed_9_1_1 --dblock_ar_weight 0.25 --dblock_sat_weight 0.25 --dblock_nat_weight 0.25 --dblock_aux_grad_ratio 0.5 --dblock_min_ar_share 0.8 --lr_core 2e-7 --lr_head 2e-7 --dblock_checkpoint_stride 2 --empty_cache_every_steps 0 --dblock_log_every 25 --cuda_max_reserved_mib 0 --cuda_min_free_mib 0 --dblock_clean_context e2e --dblock_local_rows 2 --dblock_clean_context_noise_scale 0.0 --dblock_depth_anchor_every 50 --dblock_fullstack_anchor_cap 256 --dblock_fullstack_ar_rows 1 --detachable_50m_upgrade_multimode --detachable_50m_upgrade_logit_distill --detachable_50m_upgrade_sat_gate_v2713 --detachable_50m_create_from_legacy --detachable_50m_seed_from_mature --detachable_50m_lr 0.0001 --detachable_50m_weight_decay 0.0 --detachable_50m_update_mode sublayer --detachable_50m_batch 1 --detachable_50m_tokens 64 --detachable_50m_loss_tokens 64 --detachable_50m_every 1 --detachable_50m_teacher_every 1 --detachable_50m_distill_weight 1.0 --detachable_50m_logit_distill_alpha 0.30 --detachable_50m_logit_distill_temperature 1.0 --detachable_50m_logit_distill_tokens 64 --detachable_50m_hidden_distill_weight 0.0 --detachable_50m_anchor_every 128 --detachable_50m_anchor_tokens 32 --detachable_50m_seed 20260830 --detachable_50m_log_every 10 --detachable_50m_telemetry_keep 128 --weight_accel_mode shadow --weight_accel_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/weight_trajectory_accel --weight_accel_every 1 --weight_accel_train_every 4 --weight_accel_record_every 14 --weight_accel_state_every 256 --detachable_50m --detachable_50m_optimizer paged_adamw8bit --delta_every_steps 0 --delta_every_sec 0 --resume /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838083_from02835222_20260911T010911_1789088951909741831_3003552_000003.pt --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --data_seed 2838125 --batch_size 56 --dblock_blocks 14 --dblock_checkpoint_skip_tail 2 --dblock_direct56_enabled 1 --dblock_layers_per_update 0 --dblock_router shadow --dblock_router_blend 0.0 --dblock_router_full_anchor_every 128 --dblock_router_train_layer_policy cyclic --dblock_router_train_layers_per_update 1 --dblock_sublayer_mode off --dblock_target_route_activate_from_checkpoint 1 --dblock_target_route_activation_checkpoint_step 2717509 --dblock_target_route_activation_parent_pointer_sha256 04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827 --dblock_target_route_offset 1 --dblock_target_route_policy xor_fold56_v1 --dblock_train_layer_policy cyclic --dblock_train_layers_per_update 1 --dblock_train_sublayer_policy cyclic --dblock_train_sublayers_per_update 1 --granularity_label one-sublayer-local-grad-direct56-teacher-v27.16 --oom_known_safe_batch 56 --val_every_sec 600 --val_file /workspace/agillm43_heldout_v3/token_ids.json --val_sha256 513b30203ffd18167655deb2d9cd610541df87d1ea0b8a0f9241c58e68703d21",
|
| 472 |
+
"cwd": "/workspace/agillm44_throughput_80k_20260910",
|
| 473 |
+
"env": {
|
| 474 |
+
"AGILLM43_DBLOCK_CC_NOISE_SCALE": "0.0",
|
| 475 |
+
"AGILLM43_DBLOCK_CLEAN_CONTEXT": "e2e",
|
| 476 |
+
"AGILLM43_DBLOCK_CLEAN_CONTEXT_NOISE_SCALE": "0.0",
|
| 477 |
+
"AGILLM43_DBLOCK_DEPTH_ANCHOR_EVERY": "50",
|
| 478 |
+
"AGILLM43_DBLOCK_FULLSTACK_ANCHOR_CAP": "256",
|
| 479 |
+
"AGILLM43_DBLOCK_FULLSTACK_AR_ROWS": "1",
|
| 480 |
+
"AGILLM43_DBLOCK_LOCAL_ROWS": "2",
|
| 481 |
+
"AGILLM43_DBLOCK_TARGET_ROUTE_ACTIVATE_FROM_CHECKPOINT": "1",
|
| 482 |
+
"AGILLM43_DBLOCK_TARGET_ROUTE_ACTIVATION_CHECKPOINT_STEP": "2717509",
|
| 483 |
+
"AGILLM43_DBLOCK_TARGET_ROUTE_ACTIVATION_PARENT_POINTER_SHA256": "04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827",
|
| 484 |
+
"AGILLM43_DBLOCK_TARGET_ROUTE_OFFSET": "1",
|
| 485 |
+
"AGILLM43_DBLOCK_TARGET_ROUTE_POLICY": "xor_fold56_v1",
|
| 486 |
+
"AGILLM43_DECOMPRESS_CACHE": "0",
|
| 487 |
+
"AGILLM43_DETACHABLE50M_ANCHOR_EVERY": "128",
|
| 488 |
+
"AGILLM43_DETACHABLE50M_ANCHOR_TOKENS": "<redacted>",
|
| 489 |
+
"AGILLM43_DETACHABLE50M_BATCH": "1",
|
| 490 |
+
"AGILLM43_DETACHABLE50M_CREATE_FROM_LEGACY": "1",
|
| 491 |
+
"AGILLM43_DETACHABLE50M_DISTILL_WEIGHT": "1.0",
|
| 492 |
+
"AGILLM43_DETACHABLE50M_ENABLED": "1",
|
| 493 |
+
"AGILLM43_DETACHABLE50M_EVERY": "1",
|
| 494 |
+
"AGILLM43_DETACHABLE50M_GATE_V2713_START_CHECKPOINT_SHA256": "39f442fc0e6f9fad6f875e067f63db67f33a41a53e8cbc49ce1591010bb4aec3",
|
| 495 |
+
"AGILLM43_DETACHABLE50M_HIDDEN_DISTILL_WEIGHT": "0.0",
|
| 496 |
+
"AGILLM43_DETACHABLE50M_LOGIT_DISTILL_ALPHA": "0.30",
|
| 497 |
+
"AGILLM43_DETACHABLE50M_LOGIT_DISTILL_TEMPERATURE": "1.0",
|
| 498 |
+
"AGILLM43_DETACHABLE50M_LOGIT_DISTILL_TOKENS": "<redacted>",
|
| 499 |
+
"AGILLM43_DETACHABLE50M_LOG_EVERY": "10",
|
| 500 |
+
"AGILLM43_DETACHABLE50M_LOSS_TOKENS": "<redacted>",
|
| 501 |
+
"AGILLM43_DETACHABLE50M_LR": "0.0001",
|
| 502 |
+
"AGILLM43_DETACHABLE50M_OPTIMIZER": "paged_adamw8bit",
|
| 503 |
+
"AGILLM43_DETACHABLE50M_SEED": "20260830",
|
| 504 |
+
"AGILLM43_DETACHABLE50M_SEED_FROM_MATURE": "1",
|
| 505 |
+
"AGILLM43_DETACHABLE50M_TEACHER_EVERY": "1",
|
| 506 |
+
"AGILLM43_DETACHABLE50M_TELEMETRY_KEEP": "128",
|
| 507 |
+
"AGILLM43_DETACHABLE50M_TOKENS": "<redacted>",
|
| 508 |
+
"AGILLM43_DETACHABLE50M_UPDATE_MODE": "sublayer",
|
| 509 |
+
"AGILLM43_DETACHABLE50M_UPGRADE_LOGIT_DISTILL": "1",
|
| 510 |
+
"AGILLM43_DETACHABLE50M_UPGRADE_MULTIMODE": "1",
|
| 511 |
+
"AGILLM43_DETACHABLE50M_UPGRADE_SAT_GATE_V2713": "1",
|
| 512 |
+
"AGILLM43_DETACHABLE50M_WEIGHT_DECAY": "0.0",
|
| 513 |
+
"AGILLM43_SEQUENCE_PREFETCH_DEPTH": "256",
|
| 514 |
+
"AGILLM43_SINGLEFILE_PROFILE": "intelligence-v25-satvar-pretrain",
|
| 515 |
+
"AGILLM43_SUPERVISE_SET_ARGS_JSON": "{\"--batch_size\":\"56\",\"--dblock_blocks\":\"14\",\"--dblock_checkpoint_skip_tail\":\"2\",\"--dblock_direct56_enabled\":\"1\",\"--dblock_layers_per_update\":\"0\",\"--dblock_router\":\"shadow\",\"--dblock_router_blend\":\"0.0\",\"--dblock_router_full_anchor_every\":\"128\",\"--dblock_router_train_layer_policy\":\"cyclic\",\"--dblock_router_train_layers_per_update\":\"1\",\"--dblock_sublayer_mode\":\"off\",\"--dblock_target_route_activate_from_checkpoint\":\"1\",\"--dblock_target_route_activation_checkpoint_step\":\"2717509\",\"--dblock_target_route_activation_parent_pointer_sha256\":\"04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827\",\"--dblock_target_route_offset\":\"1\",\"--dblock_target_route_policy\":\"xor_fold56_v1\",\"--dblock_train_layer_policy\":\"cyclic\",\"--dblock_train_layers_per_update\":\"1\",\"--dblock_train_sublayer_policy\":\"cyclic\",\"--dblock_train_sublayers_per_update\":\"1\",\"--granularity_label\":\"one-sublayer-local-grad-direct56-teacher-v27.16\",\"--oom_known_safe_batch\":\"56\",\"--val_every_sec\":\"600\",\"--val_file\":\"/workspace/agillm43_heldout_v3/token_ids.json\",\"--val_sha256\":\"513b30203ffd18167655deb2d9cd610541df87d1ea0b8a0f9241c58e68703d21\"}",
|
| 516 |
+
"AGILLM43_WEIGHT_ACCEL_MODE": "shadow",
|
| 517 |
+
"AGILLM_DATASETS_STREAMING_READ_MAX_RETRIES": "2",
|
| 518 |
+
"AGILLM_DATASETS_STREAMING_READ_RETRY_INTERVAL": "1",
|
| 519 |
+
"AGILLM_DATASET_AGENT_ROUTER": "0",
|
| 520 |
+
"AGILLM_DATASET_HOTLOAD_CONFIG": "/workspace/agillm43_dataset_hotload.json",
|
| 521 |
+
"AGILLM_DATASET_HOT_RELOAD_CHECK_TOKENS": "<redacted>",
|
| 522 |
+
"AGILLM_DATASET_HOT_RELOAD_SEC": "5",
|
| 523 |
+
"AGILLM_DATASET_NN_ROUTER": "1",
|
| 524 |
+
"AGILLM_DATASET_ROUTER_BLEND": "0.25",
|
| 525 |
+
"AGILLM_DATASET_ROUTER_EXPLORE": "0.03",
|
| 526 |
+
"AGILLM_DATASET_ROUTER_LR": "0.005",
|
| 527 |
+
"AGILLM_DATASET_ROUTER_MIN_SCORE": "0.15",
|
| 528 |
+
"AGILLM_DATASET_ROUTER_SHARPNESS": "2.0",
|
| 529 |
+
"AGILLM_ENABLE_CUPY": "0",
|
| 530 |
+
"AGILLM_HF_HUB_DOWNLOAD_TIMEOUT": "30",
|
| 531 |
+
"AGILLM_HF_TRANSPORT": "geth_hf_relay",
|
| 532 |
+
"AGILLM_HOT_CONFIG": "/workspace/agillm43_production_v14/contracts/hot_config_owner_v25.json",
|
| 533 |
+
"AGILLM_HOT_CONFIG_SHA256": "",
|
| 534 |
+
"AGILLM_MAX_EXAMPLE_CHARS": "32768",
|
| 535 |
+
"AGILLM_MAX_EXAMPLE_TOKENS": "<redacted>",
|
| 536 |
+
"AGILLM_PARENT_POINTER_SHA256": "e61a545c9899a483db818a2a31b3d3ddedd9ec79d21b900663222f225d6b5b64",
|
| 537 |
+
"AGILLM_REMOTE_SHARDS_PER_ITER": "4",
|
| 538 |
+
"AGILLM_RUN_ID": "singlefile_intelligence_v25_20260911T011303Z_from_step000002838083",
|
| 539 |
+
"AGILLM_STREAM_NEXT_TIMEOUT_SEC": "60",
|
| 540 |
+
"AGILLM_STREAM_OPEN_TIMEOUT_SEC": "120",
|
| 541 |
+
"AGILLM_STREAM_SOCKET_TIMEOUT_SEC": "30",
|
| 542 |
+
"AGILLM_STREAM_SOURCE_FATAL_COOLDOWN_SEC": "600",
|
| 543 |
+
"AGILLM_STREAM_SOURCE_MAX_COOLDOWN_SEC": "120",
|
| 544 |
+
"AGILLM_SYNTHETIC_TOKENIZER": "<redacted>",
|
| 545 |
+
"CUDA_MODULE_LOADING": "LAZY",
|
| 546 |
+
"CUDA_VERSION": "12.1.1",
|
| 547 |
+
"CUDA_VISIBLE_DEVICES": "0",
|
| 548 |
+
"HF_HUB_DOWNLOAD_TIMEOUT": "30",
|
| 549 |
+
"HF_HUB_ETAG_TIMEOUT": "10",
|
| 550 |
+
"PYTORCH_CUDA_ALLOC_CONF": "backend:cudaMallocAsync",
|
| 551 |
+
"PYTORCH_VERSION": "2.2.1"
|
| 552 |
+
},
|
| 553 |
+
"parent_proc_cmdline": "/opt/conda/bin/python -u /workspace/agillm44_throughput_80k_20260910/runtime_retain_activations_v2.py supervise --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --log /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/logs/singlefile_intelligence_v25_satvar_pretrain.log --sleep_sec 15",
|
| 554 |
+
"pid": 3427606,
|
| 555 |
+
"ppid": 3427520,
|
| 556 |
+
"proc_cmdline": "/opt/conda/bin/python -u /workspace/agillm44_throughput_80k_20260910/runtime_retain_activations_v2.py train --preset agillm4_floor --block 2048 --source HuggingFaceFW/fineweb-edu:sample-10BT|1.60,HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.50,wikimedia/wikipedia:20231101.en|1.10,allenai/c4:en|0.40,Skylion007/openwebtext|0.35,HuggingFaceTB/cosmopedia:web_samples_v2|1.30,HuggingFaceTB/smollm-corpus:cosmopedia-v2|1.30,EleutherAI/proof-pile-2:all|1.50,allenai/dolma:v1_6-sample|0.50,codeparrot/codeparrot-clean|1.25 --nat_mask_token_id 2 --tie_weights --tie_kv --moe_ffn --moe_experts 2 --moe_top_k 1 --moe_mlp_mult 4 --moe_shared_experts 1 --moe_shared_mlp_mult 2 --moe_aux_coef 0.01 --moe_z_coef 0.001 --amp --grad_checkpoint --attn_backend sdpa --sublinear_window 128 --sublinear_stride 128 --sublinear_max_anchors 128 --sublinear_chunk 128 --sublinear_sinks 4 --sublinear_recent_anchors 64 --no-sublinear_pooled_landmarks --anchor_stride 256 --anchor_max 2048 --anchor_position -1 --alibi_mode corrected --alibi_scale 0.0 --optimizer adamw8bit --weight_decay 0.0 --lr_decay cosine --lr_decay_tokens 49000954880 --lr_min_mult 0.2 --lr_schedule_reset_on_resume --lr_warmup_tokens 10000000 --lr_warmup_min_mult 0.1 --dblock --dblock_schedule roundrobin --dblock_router_ramp_steps 64 --dblock_warmup_steps 14 --dblock_loop_layers 0 --dblock_loop_start 0 --dblock_loop_cond_scale 1.0 --dblock_sigma_curriculum_steps 2000 --dblock_sigma_sampling lognormal --dblock_sigma_stratified --dblock_sigma_min 0.002 --dblock_sigma_max 2.0 --dblock_sigma_pmean -1.2 --dblock_sigma_pstd 1.2 --dblock_edm_wmax 5.0 --dblock_ar_prob 0.8 --dblock_sat_prob 0.1 --dblock_nat_prob 0.1 --dblock_ar_loss_tokens 8192 --dblock_sat_loss_tokens 4096 --dblock_nat_loss_tokens 4096 --dblock_nat_embed_noise_mode off --dblock_nat_embed_noise_scale 1.0 --nat_loss_weight 1.0 --nat_mask_ratio 0.5 --nat_max_tokens 0 --nat_span_mask_prob 0.35 --nat_suffix_mask_prob 0.20 --nat_span_max_tokens 0 --loss_spike_skip 0.0 --sat_every 1 --nat_every 1 --dblock_fullstack_anchor_spike_skip 0.0 --dblock_fullstack_anchor_max_ce 0.0 --dblock_anchor_softcap_mult 4.0 --dblock_anchor_softcap_min_ce 25.0 --dblock_anchor_softcap_max_ce 50.0 --dblock_fullstack_ar_every 3 --dblock_fullstack_ar_offset 0 --dblock_fullstack_ar_tokens 256 --dblock_fullstack_sat_every 3 --dblock_fullstack_sat_offset 1 --dblock_fullstack_sat_tokens 128 --dblock_fullstack_nat_every 3 --dblock_fullstack_nat_offset 2 --dblock_fullstack_nat_tokens 128 --dblock_fullstack_nat_mask_id -1 --ckpt_codec block-sharded-zstd --async_update_every_steps 0 --dblock_router_hidden 64 --dblock_router_heads 4 --dblock_router_layers 2 --dblock_router_lr 0.0005 --dblock_explore 0.08 --dblock_max_stale_steps 64 --dblock_max_count_skew 1.35 --dblock_stale_bonus 0.35 --dblock_undertrain_bonus 0.25 --dblock_activation_offload_min_mb 1.0 --nat_document_boundary_aware --dblock_fullstack_anchor_deterministic_eval --dblock_fullstack_anchor_spike_retry_limit 3 --dblock_local_spike_retry_limit 3 --dblock_fullstack_nat_no_valid_crop_limit 8 --target_tokens 400000000000 --lr_schedule_reanchor_on_resume --save_every_sec 1800 --delta_max_keep 0 --max_ckpts 2 --disk_free_floor_gb 12.0 --heartbeat_every_sec 30 --dblock_fullstack_ar_weight 1.0 --dblock_fullstack_sat_weight 0.1 --dblock_fullstack_nat_weight 0.1 --dblock_local_aux_warn_ce 0.0 --dblock_local_aux_max_ce 0.0 --lr_schedule_origin_tokens 85002387456 --dblock_stop_after_commits 0 --no-oom_auto_backoff --require_exact_resume_state --dblock_objective_mode committed_9_1_1 --dblock_ar_weight 0.25 --dblock_sat_weight 0.25 --dblock_nat_weight 0.25 --dblock_aux_grad_ratio 0.5 --dblock_min_ar_share 0.8 --lr_core 2e-7 --lr_head 2e-7 --dblock_checkpoint_stride 2 --empty_cache_every_steps 0 --dblock_log_every 25 --cuda_max_reserved_mib 0 --cuda_min_free_mib 0 --dblock_clean_context e2e --dblock_local_rows 2 --dblock_clean_context_noise_scale 0.0 --dblock_depth_anchor_every 50 --dblock_fullstack_anchor_cap 256 --dblock_fullstack_ar_rows 1 --detachable_50m_upgrade_multimode --detachable_50m_upgrade_logit_distill --detachable_50m_upgrade_sat_gate_v2713 --detachable_50m_create_from_legacy --detachable_50m_seed_from_mature --detachable_50m_lr 0.0001 --detachable_50m_weight_decay 0.0 --detachable_50m_update_mode sublayer --detachable_50m_batch 1 --detachable_50m_tokens 64 --detachable_50m_loss_tokens 64 --detachable_50m_every 1 --detachable_50m_teacher_every 1 --detachable_50m_distill_weight 1.0 --detachable_50m_logit_distill_alpha 0.30 --detachable_50m_logit_distill_temperature 1.0 --detachable_50m_logit_distill_tokens 64 --detachable_50m_hidden_distill_weight 0.0 --detachable_50m_anchor_every 128 --detachable_50m_anchor_tokens 32 --detachable_50m_seed 20260830 --detachable_50m_log_every 10 --detachable_50m_telemetry_keep 128 --weight_accel_mode shadow --weight_accel_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/weight_trajectory_accel --weight_accel_every 1 --weight_accel_train_every 4 --weight_accel_record_every 14 --weight_accel_state_every 256 --detachable_50m --detachable_50m_optimizer paged_adamw8bit --delta_every_steps 0 --delta_every_sec 0 --resume /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838083_from02835222_20260911T010911_1789088951909741831_3003552_000003.pt --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --data_seed 2838125 --batch_size 56 --dblock_blocks 14 --dblock_checkpoint_skip_tail 2 --dblock_direct56_enabled 1 --dblock_layers_per_update 0 --dblock_router shadow --dblock_router_blend 0.0 --dblock_router_full_anchor_every 128 --dblock_router_train_layer_policy cyclic --dblock_router_train_layers_per_update 1 --dblock_sublayer_mode off --dblock_target_route_activate_from_checkpoint 1 --dblock_target_route_activation_checkpoint_step 2717509 --dblock_target_route_activation_parent_pointer_sha256 04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827 --dblock_target_route_offset 1 --dblock_target_route_policy xor_fold56_v1 --dblock_train_layer_policy cyclic --dblock_train_layers_per_update 1 --dblock_train_sublayer_policy cyclic --dblock_train_sublayers_per_update 1 --granularity_label one-sublayer-local-grad-direct56-teacher-v27.16 --oom_known_safe_batch 56 --val_every_sec 600 --val_file /workspace/agillm43_heldout_v3/token_ids.json --val_sha256 513b30203ffd18167655deb2d9cd610541df87d1ea0b8a0f9241c58e68703d21",
|
| 557 |
+
"schema": "agillm.launch.v1"
|
| 558 |
+
},
|
| 559 |
+
"local_step": 2838528,
|
| 560 |
+
"loss": 5.946066379547119,
|
| 561 |
+
"nat_mask_contract": "agillm.nat-mask.v1",
|
| 562 |
+
"nat_mask_contract_source": "explicit-legacy-recovery",
|
| 563 |
+
"nat_mask_eos_id": 1,
|
| 564 |
+
"nat_mask_legacy_blank_id": 1,
|
| 565 |
+
"nat_mask_migrated_from_id": 1,
|
| 566 |
+
"nat_mask_schema_version": 1,
|
| 567 |
+
"nat_mask_token_id": 2,
|
| 568 |
+
"pid": 3427606,
|
| 569 |
+
"schema_version": 1,
|
| 570 |
+
"seen_tok": 282430494720,
|
| 571 |
+
"source_is_lane_checkpoint": true,
|
| 572 |
+
"source_is_master_checkpoint": false,
|
| 573 |
+
"source_is_non_master_checkpoint": false,
|
| 574 |
+
"step": 2838528,
|
| 575 |
+
"train_argv": "/workspace/agillm44_throughput_80k_20260910/runtime_retain_activations_v2.py train --preset agillm4_floor --block 2048 --source HuggingFaceFW/fineweb-edu:sample-10BT|1.60,HuggingFaceFW/fineweb:CC-MAIN-2024-10|0.50,wikimedia/wikipedia:20231101.en|1.10,allenai/c4:en|0.40,Skylion007/openwebtext|0.35,HuggingFaceTB/cosmopedia:web_samples_v2|1.30,HuggingFaceTB/smollm-corpus:cosmopedia-v2|1.30,EleutherAI/proof-pile-2:all|1.50,allenai/dolma:v1_6-sample|0.50,codeparrot/codeparrot-clean|1.25 --nat_mask_token_id 2 --tie_weights --tie_kv --moe_ffn --moe_experts 2 --moe_top_k 1 --moe_mlp_mult 4 --moe_shared_experts 1 --moe_shared_mlp_mult 2 --moe_aux_coef 0.01 --moe_z_coef 0.001 --amp --grad_checkpoint --attn_backend sdpa --sublinear_window 128 --sublinear_stride 128 --sublinear_max_anchors 128 --sublinear_chunk 128 --sublinear_sinks 4 --sublinear_recent_anchors 64 --no-sublinear_pooled_landmarks --anchor_stride 256 --anchor_max 2048 --anchor_position -1 --alibi_mode corrected --alibi_scale 0.0 --optimizer adamw8bit --weight_decay 0.0 --lr_decay cosine --lr_decay_tokens 49000954880 --lr_min_mult 0.2 --lr_schedule_reset_on_resume --lr_warmup_tokens 10000000 --lr_warmup_min_mult 0.1 --dblock --dblock_schedule roundrobin --dblock_router_ramp_steps 64 --dblock_warmup_steps 14 --dblock_loop_layers 0 --dblock_loop_start 0 --dblock_loop_cond_scale 1.0 --dblock_sigma_curriculum_steps 2000 --dblock_sigma_sampling lognormal --dblock_sigma_stratified --dblock_sigma_min 0.002 --dblock_sigma_max 2.0 --dblock_sigma_pmean -1.2 --dblock_sigma_pstd 1.2 --dblock_edm_wmax 5.0 --dblock_ar_prob 0.8 --dblock_sat_prob 0.1 --dblock_nat_prob 0.1 --dblock_ar_loss_tokens 8192 --dblock_sat_loss_tokens 4096 --dblock_nat_loss_tokens 4096 --dblock_nat_embed_noise_mode off --dblock_nat_embed_noise_scale 1.0 --nat_loss_weight 1.0 --nat_mask_ratio 0.5 --nat_max_tokens 0 --nat_span_mask_prob 0.35 --nat_suffix_mask_prob 0.20 --nat_span_max_tokens 0 --loss_spike_skip 0.0 --sat_every 1 --nat_every 1 --dblock_fullstack_anchor_spike_skip 0.0 --dblock_fullstack_anchor_max_ce 0.0 --dblock_anchor_softcap_mult 4.0 --dblock_anchor_softcap_min_ce 25.0 --dblock_anchor_softcap_max_ce 50.0 --dblock_fullstack_ar_every 3 --dblock_fullstack_ar_offset 0 --dblock_fullstack_ar_tokens 256 --dblock_fullstack_sat_every 3 --dblock_fullstack_sat_offset 1 --dblock_fullstack_sat_tokens 128 --dblock_fullstack_nat_every 3 --dblock_fullstack_nat_offset 2 --dblock_fullstack_nat_tokens 128 --dblock_fullstack_nat_mask_id -1 --ckpt_codec block-sharded-zstd --async_update_every_steps 0 --dblock_router_hidden 64 --dblock_router_heads 4 --dblock_router_layers 2 --dblock_router_lr 0.0005 --dblock_explore 0.08 --dblock_max_stale_steps 64 --dblock_max_count_skew 1.35 --dblock_stale_bonus 0.35 --dblock_undertrain_bonus 0.25 --dblock_activation_offload_min_mb 1.0 --nat_document_boundary_aware --dblock_fullstack_anchor_deterministic_eval --dblock_fullstack_anchor_spike_retry_limit 3 --dblock_local_spike_retry_limit 3 --dblock_fullstack_nat_no_valid_crop_limit 8 --target_tokens 400000000000 --lr_schedule_reanchor_on_resume --save_every_sec 1800 --delta_max_keep 0 --max_ckpts 2 --disk_free_floor_gb 12.0 --heartbeat_every_sec 30 --dblock_fullstack_ar_weight 1.0 --dblock_fullstack_sat_weight 0.1 --dblock_fullstack_nat_weight 0.1 --dblock_local_aux_warn_ce 0.0 --dblock_local_aux_max_ce 0.0 --lr_schedule_origin_tokens 85002387456 --dblock_stop_after_commits 0 --no-oom_auto_backoff --require_exact_resume_state --dblock_objective_mode committed_9_1_1 --dblock_ar_weight 0.25 --dblock_sat_weight 0.25 --dblock_nat_weight 0.25 --dblock_aux_grad_ratio 0.5 --dblock_min_ar_share 0.8 --lr_core 2e-7 --lr_head 2e-7 --dblock_checkpoint_stride 2 --empty_cache_every_steps 0 --dblock_log_every 25 --cuda_max_reserved_mib 0 --cuda_min_free_mib 0 --dblock_clean_context e2e --dblock_local_rows 2 --dblock_clean_context_noise_scale 0.0 --dblock_depth_anchor_every 50 --dblock_fullstack_anchor_cap 256 --dblock_fullstack_ar_rows 1 --detachable_50m_upgrade_multimode --detachable_50m_upgrade_logit_distill --detachable_50m_upgrade_sat_gate_v2713 --detachable_50m_create_from_legacy --detachable_50m_seed_from_mature --detachable_50m_lr 0.0001 --detachable_50m_weight_decay 0.0 --detachable_50m_update_mode sublayer --detachable_50m_batch 1 --detachable_50m_tokens 64 --detachable_50m_loss_tokens 64 --detachable_50m_every 1 --detachable_50m_teacher_every 1 --detachable_50m_distill_weight 1.0 --detachable_50m_logit_distill_alpha 0.30 --detachable_50m_logit_distill_temperature 1.0 --detachable_50m_logit_distill_tokens 64 --detachable_50m_hidden_distill_weight 0.0 --detachable_50m_anchor_every 128 --detachable_50m_anchor_tokens 32 --detachable_50m_seed 20260830 --detachable_50m_log_every 10 --detachable_50m_telemetry_keep 128 --weight_accel_mode shadow --weight_accel_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/weight_trajectory_accel --weight_accel_every 1 --weight_accel_train_every 4 --weight_accel_record_every 14 --weight_accel_state_every 256 --detachable_50m --detachable_50m_optimizer paged_adamw8bit --delta_every_steps 0 --delta_every_sec 0 --resume /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838083_from02835222_20260911T010911_1789088951909741831_3003552_000003.pt --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --data_seed 2838125 --batch_size 56 --dblock_blocks 14 --dblock_checkpoint_skip_tail 2 --dblock_direct56_enabled 1 --dblock_layers_per_update 0 --dblock_router shadow --dblock_router_blend 0.0 --dblock_router_full_anchor_every 128 --dblock_router_train_layer_policy cyclic --dblock_router_train_layers_per_update 1 --dblock_sublayer_mode off --dblock_target_route_activate_from_checkpoint 1 --dblock_target_route_activation_checkpoint_step 2717509 --dblock_target_route_activation_parent_pointer_sha256 04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827 --dblock_target_route_offset 1 --dblock_target_route_policy xor_fold56_v1 --dblock_train_layer_policy cyclic --dblock_train_layers_per_update 1 --dblock_train_sublayer_policy cyclic --dblock_train_sublayers_per_update 1 --granularity_label one-sublayer-local-grad-direct56-teacher-v27.16 --oom_known_safe_batch 56 --val_every_sec 600 --val_file /workspace/agillm43_heldout_v3/token_ids.json --val_sha256 513b30203ffd18167655deb2d9cd610541df87d1ea0b8a0f9241c58e68703d21",
|
| 576 |
+
"train_script": "runtime_retain_activations_v2.py",
|
| 577 |
+
"train_script_sha256": "714ee71c440ba9e9699034d7ea5f1de6f73d65cc80caf9f9350cf30f11a33e28",
|
| 578 |
+
"warmstart_base_seen_tok": 0,
|
| 579 |
+
"warmstart_base_step": 2838083,
|
| 580 |
+
"warmstart_kind": "warmstarted_from_lane_checkpoint",
|
| 581 |
+
"warmstart_source_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838083_from02835222_20260911T010911_1789088951909741831_3003552_000003.pt",
|
| 582 |
+
"warmstart_source_step": 2838083
|
| 583 |
+
}
|
checkpoints/step2838528_20260911/training_latest.json
ADDED
|
@@ -0,0 +1,455 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"checkpoint_main_nbytes": 18513489,
|
| 3 |
+
"checkpoint_main_sha256": "c0c72fb1662ef4f6854c6db0a1d2f1f7e55491e1be230620ec8dc2d6635b0d4c",
|
| 4 |
+
"checkpoint_name": "pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt",
|
| 5 |
+
"checkpoint_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt",
|
| 6 |
+
"continuation_state_schema": "agillm43.continuation.state.v2",
|
| 7 |
+
"continuation_state_sha256": "6790e6a53b2f1100062371f0e416a3352cae1729e168d509c8e51456e7fbf8b2",
|
| 8 |
+
"dblock_committed_step": 1108177,
|
| 9 |
+
"dblock_target_route_clock": 121019,
|
| 10 |
+
"dblock_target_route_controller_source_sha256": "714ee71c440ba9e9699034d7ea5f1de6f73d65cc80caf9f9350cf30f11a33e28",
|
| 11 |
+
"dblock_target_route_last_committed_target": {
|
| 12 |
+
"attempts": 1,
|
| 13 |
+
"block": 6,
|
| 14 |
+
"canonical_index": 3,
|
| 15 |
+
"committed_step_after": 1108177,
|
| 16 |
+
"first_attempt_step": 121018,
|
| 17 |
+
"k": 3,
|
| 18 |
+
"last_attempt_step": 121018,
|
| 19 |
+
"layer_offset": 1,
|
| 20 |
+
"phase": 3,
|
| 21 |
+
"physical_layer": 13,
|
| 22 |
+
"q": 0,
|
| 23 |
+
"route_commit_clock": 121018,
|
| 24 |
+
"route_offset": 1,
|
| 25 |
+
"selection": {
|
| 26 |
+
"authority": "learned",
|
| 27 |
+
"authority_granted": true,
|
| 28 |
+
"baseline_prediction": 6.236617569394737,
|
| 29 |
+
"candidate_hash": "7366174b2e1fbeb670854a846f1be186bf4933755de5e3bcaf7aa6b0bd537baa",
|
| 30 |
+
"candidate_persisted": true,
|
| 31 |
+
"clock": 121018,
|
| 32 |
+
"fault_count": 0,
|
| 33 |
+
"guard_override": false,
|
| 34 |
+
"guard_reason": null,
|
| 35 |
+
"nominated": true,
|
| 36 |
+
"override": false,
|
| 37 |
+
"phase": 3,
|
| 38 |
+
"raw_prediction": 27,
|
| 39 |
+
"reason": "candidate_scheduled",
|
| 40 |
+
"schema": "agillm44.direct56.controller.v1",
|
| 41 |
+
"selected": 27,
|
| 42 |
+
"selected_target": 27,
|
| 43 |
+
"shadow_override": false,
|
| 44 |
+
"stage": "learned_full",
|
| 45 |
+
"teacher_ce": 0.000266278104390949,
|
| 46 |
+
"teacher_target": 27,
|
| 47 |
+
"teacher_training_receipts": 40322,
|
| 48 |
+
"teacher_value_prediction": 5.472392559051514,
|
| 49 |
+
"validation_samples": 1458,
|
| 50 |
+
"value_prediction": 5.472392559051514
|
| 51 |
+
},
|
| 52 |
+
"sublayer": "ffn",
|
| 53 |
+
"target_id": 27,
|
| 54 |
+
"target_sha256": "d01da994f8995230a993b5e4ea9f89ceb914351e97c1a85c37c575c0aaaadc5c",
|
| 55 |
+
"teacher_target_id": 27
|
| 56 |
+
},
|
| 57 |
+
"dblock_target_route_offset": 1,
|
| 58 |
+
"dblock_target_route_payload_sha256": "4d41324260f03d8f10654853943dd2db489cd53b174fedebacac3deb876fa28f",
|
| 59 |
+
"dblock_target_route_pending_target": null,
|
| 60 |
+
"dblock_target_route_policy_id": "guarded-direct56-route-equiv56-v2",
|
| 61 |
+
"dblock_target_route_policy_sha256": "d4241a9c3a0c319cab6b268ea3792277c2cf7088b8d11492e459d2e58e2bde47",
|
| 62 |
+
"dblock_target_route_schema": "agillm43.dblock.target-route.direct56.v1",
|
| 63 |
+
"dblock_target_route_target_map_sha256": "31e64396588ebf72dd6aa20de6cbea14c20b6b0823269fade6f7471d37cc0836",
|
| 64 |
+
"dblock_target_route_topology_sha256": "c3b1679046220655820501bb30dc2733efc0628f661273ac3f2217355bdc2102",
|
| 65 |
+
"direct56_checkpointed": true,
|
| 66 |
+
"direct56_payload_digest": "cb2621feb2108b7347405ef188dd50a16eb3edff591040ecf354681e083fc0e2",
|
| 67 |
+
"direct56_schema": "agillm44.direct56.controller.v1",
|
| 68 |
+
"package_format": "block-sharded",
|
| 69 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt",
|
| 70 |
+
"quality_gate_used_for_training": false,
|
| 71 |
+
"record_sha256": "6f321fffe167db829c53809fd898d366c13c8c23356fcaf4f7cfa375b2253729",
|
| 72 |
+
"role": "training",
|
| 73 |
+
"saved_at": "2026-09-11T01:39:17Z",
|
| 74 |
+
"schema": "agillm43.training.resume.v2",
|
| 75 |
+
"seen_tok": 282430494720,
|
| 76 |
+
"shard_dir": "pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards",
|
| 77 |
+
"shards": [
|
| 78 |
+
{
|
| 79 |
+
"name": "core_base.pt",
|
| 80 |
+
"nbytes": 608644949,
|
| 81 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_base.pt",
|
| 82 |
+
"sha256": "3f3cfb5005dffbb9c8ed1b5997d107c5e56c226f3dd212a4f108de39027ceb53"
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"name": "core_block_0000.pt",
|
| 86 |
+
"nbytes": 139938036,
|
| 87 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0000.pt",
|
| 88 |
+
"sha256": "51091fb47db3d9a3848f771147fbae9f6076b767a7cf1ef53cbb2d51cca3ea89"
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"name": "core_block_0001.pt",
|
| 92 |
+
"nbytes": 140209292,
|
| 93 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0001.pt",
|
| 94 |
+
"sha256": "9a8e38312713abff1b2bdce5d34bb9245ed272ce216fc54e259dd58963fb7535"
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"name": "core_block_0002.pt",
|
| 98 |
+
"nbytes": 140184371,
|
| 99 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0002.pt",
|
| 100 |
+
"sha256": "f295f5456756c5156c50506cb44d4efe1fbad709c5a37b9e8186e7d1d828ad54"
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"name": "core_block_0003.pt",
|
| 104 |
+
"nbytes": 140173486,
|
| 105 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0003.pt",
|
| 106 |
+
"sha256": "ede6c0d0a9a5c81033f6c2daf3756d5c5993a9a61d4f278814610d3313d1ee35"
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"name": "core_block_0004.pt",
|
| 110 |
+
"nbytes": 140327374,
|
| 111 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0004.pt",
|
| 112 |
+
"sha256": "7697b65435757595cb76cd1cacbbc3b7e85940832c283e9ca6f128a411b7d4a5"
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"name": "core_block_0005.pt",
|
| 116 |
+
"nbytes": 140267592,
|
| 117 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0005.pt",
|
| 118 |
+
"sha256": "11613518dae4b65bf65f29a0e693d258c01d24d9aa10c57a64ca338c45399699"
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"name": "core_block_0006.pt",
|
| 122 |
+
"nbytes": 140626095,
|
| 123 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0006.pt",
|
| 124 |
+
"sha256": "dfd1f78f66a0bc74d028757898d31a0557ac69a375280ecdfec504f643ca50a5"
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"name": "core_block_0007.pt",
|
| 128 |
+
"nbytes": 140778054,
|
| 129 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0007.pt",
|
| 130 |
+
"sha256": "6433dbe5ab252c8e3c76e2a6b8587e8a406a67416bc3d831b0759304b46c7639"
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"name": "core_block_0008.pt",
|
| 134 |
+
"nbytes": 140653394,
|
| 135 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0008.pt",
|
| 136 |
+
"sha256": "221b12c59f35daa9aabdc1b106da50637811fe9d60bd982734859d2f61da2162"
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"name": "core_block_0009.pt",
|
| 140 |
+
"nbytes": 140486576,
|
| 141 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0009.pt",
|
| 142 |
+
"sha256": "496c16999c08c96d4f9ed48e53b0e40aab3c7f5b726b17025ac8e87117f89cc1"
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"name": "core_block_0010.pt",
|
| 146 |
+
"nbytes": 140515457,
|
| 147 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0010.pt",
|
| 148 |
+
"sha256": "320045bdac281c70de8b1fe00f398deed9f9d0f6664c95632d8586d9aed37b6c"
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"name": "core_block_0011.pt",
|
| 152 |
+
"nbytes": 140591816,
|
| 153 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0011.pt",
|
| 154 |
+
"sha256": "8515ba1f803365c90b2eab852f45bd06305b85c7e55724f97ea9a632011adcf1"
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"name": "core_block_0012.pt",
|
| 158 |
+
"nbytes": 140607662,
|
| 159 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0012.pt",
|
| 160 |
+
"sha256": "1d86355d0a5de97b30cb844a2c17c627bc232c96eb787e54fff9134d51b19adc"
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"name": "core_block_0013.pt",
|
| 164 |
+
"nbytes": 140410614,
|
| 165 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0013.pt",
|
| 166 |
+
"sha256": "88e3cf6095af6925a1bfdda21a43f3e22c50e65461d5b7c08ee01dad2098c774"
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"name": "core_block_0014.pt",
|
| 170 |
+
"nbytes": 140495439,
|
| 171 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0014.pt",
|
| 172 |
+
"sha256": "e0804e125adfda57b9104925bec383c17a038cac1d0478a47a20820a5cf4d4ca"
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"name": "core_block_0015.pt",
|
| 176 |
+
"nbytes": 140219923,
|
| 177 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0015.pt",
|
| 178 |
+
"sha256": "823cf3eb84423697da64126f22867443e0ad429fe521179df4b4d6c3dc3eac76"
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"name": "core_block_0016.pt",
|
| 182 |
+
"nbytes": 140493183,
|
| 183 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0016.pt",
|
| 184 |
+
"sha256": "59160f09c6302be66557fc853cab028d61264b1dc364d0e82db6f2eecdaf6b48"
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"name": "core_block_0017.pt",
|
| 188 |
+
"nbytes": 140096473,
|
| 189 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0017.pt",
|
| 190 |
+
"sha256": "98a5dc70e4f7fb3031eaab2c8143b19ca9fa61f046d5a644dee5fc271a74457a"
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"name": "core_block_0018.pt",
|
| 194 |
+
"nbytes": 140430102,
|
| 195 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0018.pt",
|
| 196 |
+
"sha256": "8e4cf7e1d7357d03f8bfb86191aba265c7563531c5f1abdee8c8b6c5700757d6"
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"name": "core_block_0019.pt",
|
| 200 |
+
"nbytes": 140273565,
|
| 201 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0019.pt",
|
| 202 |
+
"sha256": "213c236c801e065b09124b06e96608b5884210ac9ef4d2e591766fbd48717ed4"
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"name": "core_block_0020.pt",
|
| 206 |
+
"nbytes": 140355298,
|
| 207 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0020.pt",
|
| 208 |
+
"sha256": "39f5f277259772e930e0bd2259c39b4b2926a20678a5891e79df3b18a1a00e2a"
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"name": "core_block_0021.pt",
|
| 212 |
+
"nbytes": 140165080,
|
| 213 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0021.pt",
|
| 214 |
+
"sha256": "82935898a9f363a4adfc390065bcd77cf700fb44f71a49d26fee1e836c642572"
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"name": "core_block_0022.pt",
|
| 218 |
+
"nbytes": 140517129,
|
| 219 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0022.pt",
|
| 220 |
+
"sha256": "ce54d804873842e0ee2e2f71990decd961cadb1e1cc07f0420689b023072dee8"
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"name": "core_block_0023.pt",
|
| 224 |
+
"nbytes": 140773442,
|
| 225 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0023.pt",
|
| 226 |
+
"sha256": "1316addeeb8e11423fee07eb87442ff30f27897501fa96f1bac16638c84c64d1"
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"name": "core_block_0024.pt",
|
| 230 |
+
"nbytes": 139689242,
|
| 231 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0024.pt",
|
| 232 |
+
"sha256": "75d11107a7be7681e7d428817bad7e4ec2c543a2dbe3c9503fe0270ddfb56498"
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"name": "core_block_0025.pt",
|
| 236 |
+
"nbytes": 139713526,
|
| 237 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0025.pt",
|
| 238 |
+
"sha256": "0c7ea136440a15454c63640efcc173049f1c1177aed31a5c77d7073efc116fdc"
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"name": "core_block_0026.pt",
|
| 242 |
+
"nbytes": 139653476,
|
| 243 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0026.pt",
|
| 244 |
+
"sha256": "726747b47c0b2910fa7b5a4a28de5acf6e3e6c38e4c16b1e217d8bb841f6b312"
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"name": "core_block_0027.pt",
|
| 248 |
+
"nbytes": 139830606,
|
| 249 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/core_block_0027.pt",
|
| 250 |
+
"sha256": "38173b9baeeabf3c274b406691d2e5268bc919cadb44b15221fd8ec2b76350e5"
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"name": "detachable_50m_base.pt",
|
| 254 |
+
"nbytes": 122165852,
|
| 255 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_base.pt",
|
| 256 |
+
"sha256": "f5e679fbb833072d308f7de182e6568435b3a276108ca7a70cb02d3079a74f7b"
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"name": "detachable_50m_block_0000.pt",
|
| 260 |
+
"nbytes": 2721411,
|
| 261 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0000.pt",
|
| 262 |
+
"sha256": "4c9d646968ca23fb7110ab5c6662cf1cbbab23817b55b1566cdba33b62347d43"
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"name": "detachable_50m_block_0001.pt",
|
| 266 |
+
"nbytes": 2699796,
|
| 267 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0001.pt",
|
| 268 |
+
"sha256": "c1696846000b73e196f6abd5afddb15d35179b2cf1653174cd1c37005deef450"
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"name": "detachable_50m_block_0002.pt",
|
| 272 |
+
"nbytes": 2703658,
|
| 273 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0002.pt",
|
| 274 |
+
"sha256": "9e3909e35a7d24e481b517e0ed9555a6209bfccc69df53cfc0ad36ccb3bf26b6"
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"name": "detachable_50m_block_0003.pt",
|
| 278 |
+
"nbytes": 2703453,
|
| 279 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0003.pt",
|
| 280 |
+
"sha256": "0a193ee86658cf8e06879fd97f7e5cbe651db635369eff1a1733a04b50245b8b"
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"name": "detachable_50m_block_0004.pt",
|
| 284 |
+
"nbytes": 2704985,
|
| 285 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0004.pt",
|
| 286 |
+
"sha256": "62dfc1607b12dee36c49e384c5a8f6c83dbfbff0c18ac1fd14ef0803459d165d"
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"name": "detachable_50m_block_0005.pt",
|
| 290 |
+
"nbytes": 2710948,
|
| 291 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0005.pt",
|
| 292 |
+
"sha256": "d4c188d200f7ff03ae6acb5fd9155d2344107febf0b097b2f2baa97c09adda05"
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"name": "detachable_50m_block_0006.pt",
|
| 296 |
+
"nbytes": 2720179,
|
| 297 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0006.pt",
|
| 298 |
+
"sha256": "5a6f147e7815f2a39b764e036b2cf73dbeaaf86aba0faf11774ab3f114c8b85d"
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"name": "detachable_50m_block_0007.pt",
|
| 302 |
+
"nbytes": 2708807,
|
| 303 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0007.pt",
|
| 304 |
+
"sha256": "8e33658a4b143086af2f2e85499eed18c990ef8385432a035692be712a86a1da"
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"name": "detachable_50m_block_0008.pt",
|
| 308 |
+
"nbytes": 2718042,
|
| 309 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0008.pt",
|
| 310 |
+
"sha256": "e57bc5b71315d09e41eb69e09a5bd2db790f41859e024b25fc7c984906364bde"
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"name": "detachable_50m_block_0009.pt",
|
| 314 |
+
"nbytes": 2714065,
|
| 315 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0009.pt",
|
| 316 |
+
"sha256": "ef27517c36d3d601ef3cdcaa7b126aed5f41acad0460aeb0294861686b53f0d7"
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"name": "detachable_50m_block_0010.pt",
|
| 320 |
+
"nbytes": 2708790,
|
| 321 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0010.pt",
|
| 322 |
+
"sha256": "b7eddb66a807a1d8f510ff4176ee01ca7d2eb30a0856d55eda173760f7060183"
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"name": "detachable_50m_block_0011.pt",
|
| 326 |
+
"nbytes": 2719439,
|
| 327 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0011.pt",
|
| 328 |
+
"sha256": "f1ae6db9a614601b32346d652b18befda84310fe34261af0380f57cb8871ce06"
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"name": "detachable_50m_block_0012.pt",
|
| 332 |
+
"nbytes": 2702747,
|
| 333 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0012.pt",
|
| 334 |
+
"sha256": "36785f6af39921e41e548a8518ce4fb8e611be63bb0f37910493fb676e01a1fb"
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"name": "detachable_50m_block_0013.pt",
|
| 338 |
+
"nbytes": 2698035,
|
| 339 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0013.pt",
|
| 340 |
+
"sha256": "7d1b5a4a673330587b8e50e205c8af7020de832dde414c8f51757d191953681c"
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"name": "detachable_50m_block_0014.pt",
|
| 344 |
+
"nbytes": 2704281,
|
| 345 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0014.pt",
|
| 346 |
+
"sha256": "202baae70cd1f9e56b046b751c1528af61abe528d5fdc62f01d8e2d085d6528f"
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"name": "detachable_50m_block_0015.pt",
|
| 350 |
+
"nbytes": 2701156,
|
| 351 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0015.pt",
|
| 352 |
+
"sha256": "24b15efde13edb5f7c76601790d0e8dbb27a8e2e12ad9ca29ca8d200aa45a9a2"
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"name": "detachable_50m_block_0016.pt",
|
| 356 |
+
"nbytes": 2699013,
|
| 357 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0016.pt",
|
| 358 |
+
"sha256": "398f4fd2533b3f5743a17ded1ca697b8ba9910b75db3b03b31e536bd4b1dc514"
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"name": "detachable_50m_block_0017.pt",
|
| 362 |
+
"nbytes": 2702298,
|
| 363 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0017.pt",
|
| 364 |
+
"sha256": "bb04eacf83f2433661cf7611d6a6eea34d276851332b2c4b0b228e0a99411fab"
|
| 365 |
+
},
|
| 366 |
+
{
|
| 367 |
+
"name": "detachable_50m_block_0018.pt",
|
| 368 |
+
"nbytes": 2709050,
|
| 369 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0018.pt",
|
| 370 |
+
"sha256": "b0410ca9e8b62c1707b0bd748815552c5a2f7453f781bc2fe9cff89cc92e27ca"
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"name": "detachable_50m_block_0019.pt",
|
| 374 |
+
"nbytes": 1970442,
|
| 375 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0019.pt",
|
| 376 |
+
"sha256": "cc49557b550810a019f80e4f266fb78ba8bb3d3fc7c605013139643840bd623b"
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"name": "detachable_50m_block_0020.pt",
|
| 380 |
+
"nbytes": 1968771,
|
| 381 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0020.pt",
|
| 382 |
+
"sha256": "9a1b64ab149bb674349daa0ab02db243209218c25bd5414fcc98a04b6821a84f"
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"name": "detachable_50m_block_0021.pt",
|
| 386 |
+
"nbytes": 1969870,
|
| 387 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/detachable_50m_block_0021.pt",
|
| 388 |
+
"sha256": "40ab687303d9c3d948a78c3baea6bbecb2e4dd606c78f5734261d9eb80cb7013"
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"name": "ar.pt",
|
| 392 |
+
"nbytes": 608649226,
|
| 393 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/ar.pt",
|
| 394 |
+
"sha256": "6eb038eab02b8a7975d868cae51e8665a17e526703c248b692ba227df54cbd36"
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"name": "sat.pt",
|
| 398 |
+
"nbytes": 608658757,
|
| 399 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/sat.pt",
|
| 400 |
+
"sha256": "39491f3c4626dfde4155ba3fd73079a432db5547b677e57f004a2461c0870f33"
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"name": "nat.pt",
|
| 404 |
+
"nbytes": 608649226,
|
| 405 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/nat.pt",
|
| 406 |
+
"sha256": "6eb038eab02b8a7975d868cae51e8665a17e526703c248b692ba227df54cbd36"
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"name": "opt.pt",
|
| 410 |
+
"nbytes": 1898513769,
|
| 411 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/opt.pt",
|
| 412 |
+
"sha256": "0f2b7deddd8640e7ed13c10422d83ba00fb040bab0841cd648590db48fcec349"
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"name": "scaler.pt",
|
| 416 |
+
"nbytes": 149,
|
| 417 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/scaler.pt",
|
| 418 |
+
"sha256": "c2110dfd7b7a6c034fb62244a6f8f76a22d9936d55b6d0fe133e43e65b38631c"
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"name": "agillm5_detachable_50m_bridge.pt",
|
| 422 |
+
"nbytes": 1225015,
|
| 423 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/agillm5_detachable_50m_bridge.pt",
|
| 424 |
+
"sha256": "048693d7144c509e31729e68ed35c2831b64f59d12a823aedd20b858d452eab8"
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"name": "agillm5_detachable_50m_opt.pt",
|
| 428 |
+
"nbytes": 65092766,
|
| 429 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/agillm5_detachable_50m_opt.pt",
|
| 430 |
+
"sha256": "21dbf45f487f708907c4cc68b24862a52a22e94d90242ccabc33aa5bea5a9a07"
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"name": "agillm5_detachable_50m_scaler.pt",
|
| 434 |
+
"nbytes": 149,
|
| 435 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/agillm5_detachable_50m_scaler.pt",
|
| 436 |
+
"sha256": "c2110dfd7b7a6c034fb62244a6f8f76a22d9936d55b6d0fe133e43e65b38631c"
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"name": "agillm5_detachable_50m_rng.pt",
|
| 440 |
+
"nbytes": 3553,
|
| 441 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/agillm5_detachable_50m_rng.pt",
|
| 442 |
+
"sha256": "68ba40d73d8f27b16aa666a86e2dea813ab2807df5c053cf132aa2b6e236511b"
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"name": "agillm5_detachable_50m_counters.pt",
|
| 446 |
+
"nbytes": 30067,
|
| 447 |
+
"path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.shards/agillm5_detachable_50m_counters.pt",
|
| 448 |
+
"sha256": "3b581cd5c7c81030dee159ba749e445052b0773b420dfce65bad781cf7f8f76f"
|
| 449 |
+
}
|
| 450 |
+
],
|
| 451 |
+
"step": 2838528,
|
| 452 |
+
"tokenizer_sidecar_nbytes": 17465554,
|
| 453 |
+
"tokenizer_sidecar_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02838528_from02838083_20260911T013820_1789090700317818808_3427606_000003.pt.tokenizer.json",
|
| 454 |
+
"tokenizer_sidecar_sha256": "e8b0e152efe5446bd1a4aabe1b9171ce934a1b5af10602b0e160e0c3288c7e02"
|
| 455 |
+
}
|