Upload folder using huggingface_hub

.gitattributes

@@ -424,3 +424,4 @@ checkpoints/step2838528_20260911/pretrain_step02838528_from02838083_20260911T013
 checkpoints/step2838879_20260911/pretrain_step02838879_from02838528_20260911T022116_1789093276338073781_3552287_000001.pt.tokenizer.json filter=lfs diff=lfs merge=lfs -text
 checkpoints/step2840270_20260911/pretrain_step02840270_from02838528_20260911T031313_1789096393688971963_3552287_000002.pt.tokenizer.json filter=lfs diff=lfs merge=lfs -text
 live400B/step2841749_20260911T0416Z/pretrain_step02841749_from02840270_20260911T041419_1789100059328894287_317951_000001.pt.tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoints/step2842431_20260911/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.tokenizer.json filter=lfs diff=lfs merge=lfs -text

checkpoints/step2842431_20260911/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.provenance.json

@@ -0,0 +1,590 @@
+{
+  "batch_size": 56,
+  "block_size": 2048,
+  "checkpoint_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt",
+  "checkpoint_type": "full",
+  "created_at_iso": "2026-09-11T04:35:48Z",
+  "created_at_unix": 1789101348.5924158,
+  "created_from_scratch": false,
+  "dataset_provenance": {
+    "chat": false,
+    "dataset_field_text": "text",
+    "dataset_hotload_config_path": "/workspace/agillm43_dataset_hotload.json",
+    "dataset_hotload_config_sha256": "3f48fe33b475c719d43e14cc4d0867c19e17dfa1d2cb304cd291eea6a879a7c8",
+    "dataset_hotload_config_stamp": [
+      50,
+      1124990893,
+      1789091829464680808,
+      7210
+    ],
+    "dataset_hotload_config_used": true,
+    "dataset_hotload_enabled": true,
+    "dataset_hotload_generation": 13,
+    "dataset_hotload_schema": "agillm.dataset_hotload.config.v2",
+    "dataset_revision_map_sha256": "8361995a504a2365136b20e1a879ba78944a06309f13f3f013bbd0cdfc72225b",
+    "dataset_revisions": {
+      "EleutherAI/proof-pile-2": "901a9273a770e9d4138c5ddd91802f9c5c6cdc4b",
+      "HuggingFaceFW/fineweb": "9bb295ddab0e05d785b879661af7260fed5140fc",
+      "HuggingFaceFW/fineweb-edu": "87f09149ef4734204d70ed1d046ddc9ca3f2b8f9",
+      "HuggingFaceTB/cosmopedia": "0ae6ec63f91742bd2d1eaef4f02232c55d719385",
+      "HuggingFaceTB/finemath": "e92b25a616738fe95dc186b64dfb19f9c8525594",
+      "HuggingFaceTB/smollm-corpus": "3ba9d605774198c5868892d7a8deda78031a781f",
+      "Skylion007/openwebtext": "79d93d786212f7344586290adb811d4ae6a1762c",
+      "allenai/c4": "1588ec454efa1a09f29cd18ddd04fe05fc8653a2",
+      "allenai/dolma": "7f48140530a023e9ea4c5cfb141160922727d4d3",
+      "allenai/dolmino-mix-1124": "a319f19eef1e257417b11ea8c30da266ae175557",
+      "codeparrot/codeparrot-clean": "35a59fb025bc0a102f7d96eac09d145b896d487b",
+      "mlfoundations/dclm-baseline-1.0": "a3b142c183aebe5af344955ae20836eb34dcf69b",
+      "wikimedia/wikipedia": "b04c8d1ceb2f5cd4588862100d08de323dccfbaa"
+    },
+    "dolma_payload_integrity": "external_url_manifest_pinned_payload_hashes_unavailable",
+    "dolma_url_list_sha256": "bc6ef505be920c94bb9e7147445638f606fdc3f5af622cd1d887dc48e21a2d5b",
+    "dolma_url_manifest_raw_sha256": "6fe8c40d51e5bebc9303453d7f533b232d69f7ae122123621a82621b1cd779b7",
+    "has_language_mix": true,
+    "has_numeracy": false,
+    "hot_config_expected_sha256": "",
+    "hot_config_has_append": true,
+    "hot_config_has_datasets": false,
+    "hot_config_mtime": 1789092253.2119184,
+    "hot_config_path": "/workspace/agillm43_production_v14/contracts/hot_config_owner_v25.json",
+    "hot_config_used": true,
+    "hot_reload_capable": true,
+    "phase": "pretrain",
+    "proof_pile_lfs_manifest_sha256": "d020dfb796b02a75564e4e11d9e03185a55b3f9c6639f87830cfbe10737ed329",
+    "proof_pile_paths_sha256": "8a30018017e3b3d9a9bcfa06a3c404ccd113297b0b1d0c48e2e2aadf113e52b0",
+    "schema": "agillm.dataset_provenance.v1",
+    "sequence_prefetch_depth": 256,
+    "source_count": 23,
+    "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,json:/workspace/agillm44_ultradata_20260911/ultra_fineweb.jsonl|0.45,json:/workspace/agillm44_ultradata_20260911/ultra_fineweb_l3.jsonl|0.35,json:/workspace/agillm44_ultradata_20260911/ultradata_code.jsonl|0.35,json:/workspace/agillm44_ultradata_20260911/ultradata_math.jsonl|0.35,json:/workspace/agillm44_ultradata_20260911/ultradata_rl.jsonl|0.2,json:/workspace/agillm44_ultradata_20260911/ultradata_sft_agent.jsonl|0.2,json:/workspace/agillm44_ultradata_20260911/ultrax.jsonl|0.35",
+    "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",
+    "source_specs": [
+      "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",
+      "json:/workspace/agillm44_ultradata_20260911/ultra_fineweb.jsonl|0.45",
+      "json:/workspace/agillm44_ultradata_20260911/ultra_fineweb_l3.jsonl|0.35",
+      "json:/workspace/agillm44_ultradata_20260911/ultradata_code.jsonl|0.35",
+      "json:/workspace/agillm44_ultradata_20260911/ultradata_math.jsonl|0.35",
+      "json:/workspace/agillm44_ultradata_20260911/ultradata_rl.jsonl|0.2",
+      "json:/workspace/agillm44_ultradata_20260911/ultradata_sft_agent.jsonl|0.2",
+      "json:/workspace/agillm44_ultradata_20260911/ultrax.jsonl|0.35"
+    ],
+    "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,json:/workspace/agillm44_ultradata_20260911/ultra_fineweb.jsonl|0.45,json:/workspace/agillm44_ultradata_20260911/ultra_fineweb_l3.jsonl|0.35,json:/workspace/agillm44_ultradata_20260911/ultradata_code.jsonl|0.35,json:/workspace/agillm44_ultradata_20260911/ultradata_math.jsonl|0.35,json:/workspace/agillm44_ultradata_20260911/ultradata_rl.jsonl|0.2,json:/workspace/agillm44_ultradata_20260911/ultradata_sft_agent.jsonl|0.2,json:/workspace/agillm44_ultradata_20260911/ultrax.jsonl|0.35",
+    "val_source_requested": ""
+  },
+  "effective_global_step": 5682701,
+  "effective_seen_tok": 282878121984,
+  "global_origin_seen_tok": 0,
+  "global_origin_step": 2840270,
+  "gpu": {
+    "allocated_gb": 7.13,
+    "peak_allocated_gb": 12.05,
+    "reserved_gb": 13.34
+  },
+  "hostname": "ead6505f92c4",
+  "lane": "pretrain",
+  "launch": {
+    "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_checkpoint_skip_tail",
+      "0",
+      "--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_step02840270_from02838528_20260911T031313_1789096393688971963_3552287_000002.pt",
+      "--save_dir",
+      "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B",
+      "--data_seed",
+      "2840312",
+      "--batch_size",
+      "56",
+      "--dblock_blocks",
+      "14",
+      "--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"
+    ],
+    "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_checkpoint_skip_tail 0 --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_step02840270_from02838528_20260911T031313_1789096393688971963_3552287_000002.pt --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --data_seed 2840312 --batch_size 56 --dblock_blocks 14 --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",
+    "cwd": "/workspace/agillm44_throughput_80k_20260910",
+    "env": {
+      "AGILLM43_DBLOCK_CC_NOISE_SCALE": "0.0",
+      "AGILLM43_DBLOCK_CLEAN_CONTEXT": "e2e",
+      "AGILLM43_DBLOCK_CLEAN_CONTEXT_NOISE_SCALE": "0.0",
+      "AGILLM43_DBLOCK_DEPTH_ANCHOR_EVERY": "50",
+      "AGILLM43_DBLOCK_FULLSTACK_ANCHOR_CAP": "256",
+      "AGILLM43_DBLOCK_FULLSTACK_AR_ROWS": "1",
+      "AGILLM43_DBLOCK_LOCAL_ROWS": "2",
+      "AGILLM43_DBLOCK_TARGET_ROUTE_ACTIVATE_FROM_CHECKPOINT": "1",
+      "AGILLM43_DBLOCK_TARGET_ROUTE_ACTIVATION_CHECKPOINT_STEP": "2717509",
+      "AGILLM43_DBLOCK_TARGET_ROUTE_ACTIVATION_PARENT_POINTER_SHA256": "04638cfbc19c0f9db6e6dfa03b2509d50a0bf3c56048a2cbd6e67154bc85c827",
+      "AGILLM43_DBLOCK_TARGET_ROUTE_OFFSET": "1",
+      "AGILLM43_DBLOCK_TARGET_ROUTE_POLICY": "xor_fold56_v1",
+      "AGILLM43_DECOMPRESS_CACHE": "0",
+      "AGILLM43_DETACHABLE50M_ANCHOR_EVERY": "128",
+      "AGILLM43_DETACHABLE50M_ANCHOR_TOKENS": "<redacted>",
+      "AGILLM43_DETACHABLE50M_BATCH": "1",
+      "AGILLM43_DETACHABLE50M_CREATE_FROM_LEGACY": "1",
+      "AGILLM43_DETACHABLE50M_DISTILL_WEIGHT": "1.0",
+      "AGILLM43_DETACHABLE50M_ENABLED": "1",
+      "AGILLM43_DETACHABLE50M_EVERY": "1",
+      "AGILLM43_DETACHABLE50M_GATE_V2713_START_CHECKPOINT_SHA256": "39f442fc0e6f9fad6f875e067f63db67f33a41a53e8cbc49ce1591010bb4aec3",
+      "AGILLM43_DETACHABLE50M_HIDDEN_DISTILL_WEIGHT": "0.0",
+      "AGILLM43_DETACHABLE50M_LOGIT_DISTILL_ALPHA": "0.30",
+      "AGILLM43_DETACHABLE50M_LOGIT_DISTILL_TEMPERATURE": "1.0",
+      "AGILLM43_DETACHABLE50M_LOGIT_DISTILL_TOKENS": "<redacted>",
+      "AGILLM43_DETACHABLE50M_LOG_EVERY": "10",
+      "AGILLM43_DETACHABLE50M_LOSS_TOKENS": "<redacted>",
+      "AGILLM43_DETACHABLE50M_LR": "0.0001",
+      "AGILLM43_DETACHABLE50M_OPTIMIZER": "paged_adamw8bit",
+      "AGILLM43_DETACHABLE50M_SEED": "20260830",
+      "AGILLM43_DETACHABLE50M_SEED_FROM_MATURE": "1",
+      "AGILLM43_DETACHABLE50M_TEACHER_EVERY": "1",
+      "AGILLM43_DETACHABLE50M_TELEMETRY_KEEP": "128",
+      "AGILLM43_DETACHABLE50M_TOKENS": "<redacted>",
+      "AGILLM43_DETACHABLE50M_UPDATE_MODE": "sublayer",
+      "AGILLM43_DETACHABLE50M_UPGRADE_LOGIT_DISTILL": "1",
+      "AGILLM43_DETACHABLE50M_UPGRADE_MULTIMODE": "1",
+      "AGILLM43_DETACHABLE50M_UPGRADE_SAT_GATE_V2713": "1",
+      "AGILLM43_DETACHABLE50M_WEIGHT_DECAY": "0.0",
+      "AGILLM43_SEQUENCE_PREFETCH_DEPTH": "256",
+      "AGILLM43_SINGLEFILE_PROFILE": "intelligence-v25-satvar-pretrain",
+      "AGILLM43_SUPERVISE_SET_ARGS_JSON": "{\"--batch_size\":\"56\",\"--dblock_blocks\":\"14\",\"--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\"}",
+      "AGILLM43_WEIGHT_ACCEL_MODE": "shadow",
+      "AGILLM_DATASETS_STREAMING_READ_MAX_RETRIES": "2",
+      "AGILLM_DATASETS_STREAMING_READ_RETRY_INTERVAL": "1",
+      "AGILLM_DATASET_AGENT_ROUTER": "0",
+      "AGILLM_DATASET_HOTLOAD_CONFIG": "/workspace/agillm43_dataset_hotload.json",
+      "AGILLM_DATASET_HOT_RELOAD_CHECK_TOKENS": "<redacted>",
+      "AGILLM_DATASET_HOT_RELOAD_SEC": "5",
+      "AGILLM_DATASET_NN_ROUTER": "1",
+      "AGILLM_DATASET_ROUTER_BLEND": "0.25",
+      "AGILLM_DATASET_ROUTER_EXPLORE": "0.03",
+      "AGILLM_DATASET_ROUTER_LR": "0.005",
+      "AGILLM_DATASET_ROUTER_MIN_SCORE": "0.15",
+      "AGILLM_DATASET_ROUTER_SHARPNESS": "2.0",
+      "AGILLM_ENABLE_CUPY": "0",
+      "AGILLM_HF_HUB_DOWNLOAD_TIMEOUT": "30",
+      "AGILLM_HF_TRANSPORT": "geth_hf_relay",
+      "AGILLM_HOT_CONFIG": "/workspace/agillm43_production_v14/contracts/hot_config_owner_v25.json",
+      "AGILLM_HOT_CONFIG_SHA256": "",
+      "AGILLM_MAX_EXAMPLE_CHARS": "32768",
+      "AGILLM_MAX_EXAMPLE_TOKENS": "<redacted>",
+      "AGILLM_PARENT_POINTER_SHA256": "5385a11291cc1f9370219fda93a1d4659d70747d9f66d1efa6f4afef559e8974",
+      "AGILLM_REMOTE_SHARDS_PER_ITER": "4",
+      "AGILLM_RUN_ID": "singlefile_intelligence_v25_20260911T031845Z_from_step000002840270",
+      "AGILLM_STREAM_NEXT_TIMEOUT_SEC": "60",
+      "AGILLM_STREAM_OPEN_TIMEOUT_SEC": "120",
+      "AGILLM_STREAM_SOCKET_TIMEOUT_SEC": "30",
+      "AGILLM_STREAM_SOURCE_FATAL_COOLDOWN_SEC": "600",
+      "AGILLM_STREAM_SOURCE_MAX_COOLDOWN_SEC": "120",
+      "AGILLM_SYNTHETIC_TOKENIZER": "<redacted>",
+      "CUDA_MODULE_LOADING": "LAZY",
+      "CUDA_VERSION": "12.1.1",
+      "CUDA_VISIBLE_DEVICES": "0",
+      "HF_HUB_DOWNLOAD_TIMEOUT": "30",
+      "HF_HUB_ETAG_TIMEOUT": "10",
+      "PYTORCH_CUDA_ALLOC_CONF": "backend:cudaMallocAsync",
+      "PYTORCH_VERSION": "2.2.1"
+    },
+    "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",
+    "pid": 317951,
+    "ppid": 317877,
+    "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_checkpoint_skip_tail 0 --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_step02840270_from02838528_20260911T031313_1789096393688971963_3552287_000002.pt --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --data_seed 2840312 --batch_size 56 --dblock_blocks 14 --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",
+    "schema": "agillm.launch.v1"
+  },
+  "local_step": 2842431,
+  "loss": 5.4277472496032715,
+  "nat_mask_contract": "agillm.nat-mask.v1",
+  "nat_mask_contract_source": "explicit-legacy-recovery",
+  "nat_mask_eos_id": 1,
+  "nat_mask_legacy_blank_id": 1,
+  "nat_mask_migrated_from_id": 1,
+  "nat_mask_schema_version": 1,
+  "nat_mask_token_id": 2,
+  "pid": 317951,
+  "schema_version": 1,
+  "seen_tok": 282878121984,
+  "source_is_lane_checkpoint": true,
+  "source_is_master_checkpoint": false,
+  "source_is_non_master_checkpoint": false,
+  "step": 2842431,
+  "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_checkpoint_skip_tail 0 --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_step02840270_from02838528_20260911T031313_1789096393688971963_3552287_000002.pt --save_dir /workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B --data_seed 2840312 --batch_size 56 --dblock_blocks 14 --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",
+  "train_script": "runtime_retain_activations_v2.py",
+  "train_script_sha256": "714ee71c440ba9e9699034d7ea5f1de6f73d65cc80caf9f9350cf30f11a33e28",
+  "warmstart_base_seen_tok": 0,
+  "warmstart_base_step": 2840270,
+  "warmstart_kind": "warmstarted_from_lane_checkpoint",
+  "warmstart_source_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02840270_from02838528_20260911T031313_1789096393688971963_3552287_000002.pt",
+  "warmstart_source_step": 2840270
+}

checkpoints/step2842431_20260911/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/agillm5_detachable_50m_scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2110dfd7b7a6c034fb62244a6f8f76a22d9936d55b6d0fe133e43e65b38631c
+size 149

checkpoints/step2842431_20260911/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2110dfd7b7a6c034fb62244a6f8f76a22d9936d55b6d0fe133e43e65b38631c
+size 149

checkpoints/step2842431_20260911/training_latest.json

@@ -0,0 +1,455 @@
+{
+  "checkpoint_main_nbytes": 18533985,
+  "checkpoint_main_sha256": "2c08b804296dd26ec8c7360b2b6e1541db4f8a573294c5d87efe97e30aa40e11",
+  "checkpoint_name": "pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt",
+  "checkpoint_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt",
+  "continuation_state_schema": "agillm43.continuation.state.v2",
+  "continuation_state_sha256": "a70bc68bb6af6e732cbdd544e1f282bee977eeff8cf2c3bb5fc1e1bba2881690",
+  "dblock_committed_step": 1112080,
+  "dblock_target_route_clock": 124922,
+  "dblock_target_route_controller_source_sha256": "714ee71c440ba9e9699034d7ea5f1de6f73d65cc80caf9f9350cf30f11a33e28",
+  "dblock_target_route_last_committed_target": {
+    "attempts": 1,
+    "block": 0,
+    "canonical_index": 42,
+    "committed_step_after": 1112080,
+    "first_attempt_step": 124921,
+    "k": 0,
+    "last_attempt_step": 124921,
+    "layer_offset": 1,
+    "phase": 3,
+    "physical_layer": 1,
+    "q": 3,
+    "route_commit_clock": 124921,
+    "route_offset": 1,
+    "selection": {
+      "authority": "learned",
+      "authority_granted": true,
+      "baseline_prediction": 6.0151997356416915,
+      "candidate_hash": "7366174b2e1fbeb670854a846f1be186bf4933755de5e3bcaf7aa6b0bd537baa",
+      "candidate_persisted": true,
+      "clock": 124921,
+      "fault_count": 0,
+      "guard_override": false,
+      "guard_reason": null,
+      "nominated": true,
+      "override": false,
+      "phase": 42,
+      "raw_prediction": 3,
+      "reason": "candidate_scheduled",
+      "schema": "agillm44.direct56.controller.v1",
+      "selected": 3,
+      "selected_target": 3,
+      "shadow_override": false,
+      "stage": "learned_full",
+      "teacher_ce": 0.0002824861148837954,
+      "teacher_target": 3,
+      "teacher_training_receipts": 44225,
+      "teacher_value_prediction": 5.667962074279785,
+      "validation_samples": 1665,
+      "value_prediction": 5.667962074279785
+    },
+    "sublayer": "ffn",
+    "target_id": 3,
+    "target_sha256": "197d1396433f9932a647b282125c2390a3859a21a2617ca8b39c1dbfee478c00",
+    "teacher_target_id": 3
+  },
+  "dblock_target_route_offset": 1,
+  "dblock_target_route_payload_sha256": "b0652a84e7ad746b957756fa6934232858f379cc2dd77c55ffa99c9ea74bbe43",
+  "dblock_target_route_pending_target": null,
+  "dblock_target_route_policy_id": "guarded-direct56-route-equiv56-v2",
+  "dblock_target_route_policy_sha256": "d4241a9c3a0c319cab6b268ea3792277c2cf7088b8d11492e459d2e58e2bde47",
+  "dblock_target_route_schema": "agillm43.dblock.target-route.direct56.v1",
+  "dblock_target_route_target_map_sha256": "31e64396588ebf72dd6aa20de6cbea14c20b6b0823269fade6f7471d37cc0836",
+  "dblock_target_route_topology_sha256": "c3b1679046220655820501bb30dc2733efc0628f661273ac3f2217355bdc2102",
+  "direct56_checkpointed": true,
+  "direct56_payload_digest": "ffa168a046bc5a1ea98d493d1a989325689ba295dcdbe68fd11c9477ad8ce6aa",
+  "direct56_schema": "agillm44.direct56.controller.v1",
+  "package_format": "block-sharded",
+  "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt",
+  "quality_gate_used_for_training": false,
+  "record_sha256": "e1748da356603380e755840c2297d8ed7d9a07c2a0c0284ce6c427123c4f5de0",
+  "role": "training",
+  "saved_at": "2026-09-11T04:36:44Z",
+  "schema": "agillm43.training.resume.v2",
+  "seen_tok": 282878121984,
+  "shard_dir": "pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards",
+  "shards": [
+    {
+      "name": "core_base.pt",
+      "nbytes": 608664897,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_base.pt",
+      "sha256": "c24cd2f9966572a26f377da3c6f8490e70347999c4a2f0cbf1642b8348cc7e56"
+    },
+    {
+      "name": "core_block_0000.pt",
+      "nbytes": 139953853,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0000.pt",
+      "sha256": "f879804bc1df365197cea73e363d80fa0040ea516e9cb951b16fc14823610988"
+    },
+    {
+      "name": "core_block_0001.pt",
+      "nbytes": 140210879,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0001.pt",
+      "sha256": "a4c27622659e1795b68b42a88165f68e8061bf08f4ae00211adffacca6717d2e"
+    },
+    {
+      "name": "core_block_0002.pt",
+      "nbytes": 140184109,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0002.pt",
+      "sha256": "6c89b0ff1b9ccdfaf8e5af3ab0c131c4c13597c9d9653c2d14eed791d50e0554"
+    },
+    {
+      "name": "core_block_0003.pt",
+      "nbytes": 140171963,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0003.pt",
+      "sha256": "5f350cfa2dedd4c4d7099d0b6678353e40e11050c512edf41226f784068446a0"
+    },
+    {
+      "name": "core_block_0004.pt",
+      "nbytes": 140325754,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0004.pt",
+      "sha256": "57e0f40994be27dd16b269a0d8a80dfa47a91aba561fd0ebd4c92c217aeb63c3"
+    },
+    {
+      "name": "core_block_0005.pt",
+      "nbytes": 140270251,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0005.pt",
+      "sha256": "fa60046be6f950f90cd71409c6b1e3717abae558f69e4e4f0ca0d0c4c0123fbc"
+    },
+    {
+      "name": "core_block_0006.pt",
+      "nbytes": 140622230,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0006.pt",
+      "sha256": "3b4ca41267a6febc67c0941eca5a535204386c634464a3aea1511c1f067af32e"
+    },
+    {
+      "name": "core_block_0007.pt",
+      "nbytes": 140778584,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0007.pt",
+      "sha256": "3abc58d751637d492483797bd8046a9d6d72aca52d29a8d2e37c644e5183b02b"
+    },
+    {
+      "name": "core_block_0008.pt",
+      "nbytes": 140654532,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0008.pt",
+      "sha256": "11a0f9b85d69440e21d18d8bab62eaa77fd458060800a97c65a7d9bf16e13381"
+    },
+    {
+      "name": "core_block_0009.pt",
+      "nbytes": 140488440,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0009.pt",
+      "sha256": "90effda072682a2d143583a55f4bf462af2cd5cc11f7afc1f2577d72ed3ceea9"
+    },
+    {
+      "name": "core_block_0010.pt",
+      "nbytes": 140517485,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0010.pt",
+      "sha256": "db3abcb0ec48814d086d32dec3fcc46d0c8268c2a3d62077d00443d78ef99170"
+    },
+    {
+      "name": "core_block_0011.pt",
+      "nbytes": 140592446,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0011.pt",
+      "sha256": "ac599ff4962faa2846c121c0668d7107181dd2b2da0ce1eeac7b4e90ae1bfd06"
+    },
+    {
+      "name": "core_block_0012.pt",
+      "nbytes": 140608290,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0012.pt",
+      "sha256": "3213335608b5f231a464583138559303a36092b77307d8e50b8ccb1b6645c06a"
+    },
+    {
+      "name": "core_block_0013.pt",
+      "nbytes": 140408682,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0013.pt",
+      "sha256": "c2785ba03a575320bac3c89e102ae18b858a15f78f10b2aea67adced87632992"
+    },
+    {
+      "name": "core_block_0014.pt",
+      "nbytes": 140540506,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0014.pt",
+      "sha256": "c69961dc1d3d216c198e394d802dca8a350abd3822cf533fed44713abc607657"
+    },
+    {
+      "name": "core_block_0015.pt",
+      "nbytes": 140218454,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0015.pt",
+      "sha256": "793f41184c6a645dbb7b189f9bd3ac00b6f1c29b0923e272b847de469cfa08a6"
+    },
+    {
+      "name": "core_block_0016.pt",
+      "nbytes": 140477709,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0016.pt",
+      "sha256": "b8c0324b60d7ba6048886402fd3fb0bf0b3cf3f3cc1ed636f2b3b5a527785906"
+    },
+    {
+      "name": "core_block_0017.pt",
+      "nbytes": 140096271,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0017.pt",
+      "sha256": "3b162664a19f738c4b8c4c1ab86089f9f2036439e6c3c126934366d1f3853d67"
+    },
+    {
+      "name": "core_block_0018.pt",
+      "nbytes": 140427185,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0018.pt",
+      "sha256": "caffa722de10393195183c80f863cd91ec6dbd4611fe40a4e107cb75c7903402"
+    },
+    {
+      "name": "core_block_0019.pt",
+      "nbytes": 140275374,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0019.pt",
+      "sha256": "c781b8e952788186cea7a18e920fe43d0dfb45964f3f9ef2d98a7bf76757be6f"
+    },
+    {
+      "name": "core_block_0020.pt",
+      "nbytes": 140352837,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0020.pt",
+      "sha256": "4859fd6fb5cfa828c8fdc1465376d6b5fc16a027f7b98fe3519f52430416f2c9"
+    },
+    {
+      "name": "core_block_0021.pt",
+      "nbytes": 140164311,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0021.pt",
+      "sha256": "fa80d263f62956f46bd73fd4c7a405f97a9ce4e30ff42824476d4d136f37510d"
+    },
+    {
+      "name": "core_block_0022.pt",
+      "nbytes": 140518804,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0022.pt",
+      "sha256": "7b9e0a779950bdbbd8f1f7779d3be10836b5bb3fde2ba004c5a8e24db4bb3dea"
+    },
+    {
+      "name": "core_block_0023.pt",
+      "nbytes": 140772327,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0023.pt",
+      "sha256": "997b82b46b2b8325423915054b6fcab642d5a4fa1d71eaf515c1f49317f8edc2"
+    },
+    {
+      "name": "core_block_0024.pt",
+      "nbytes": 139689385,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0024.pt",
+      "sha256": "14b0f151441a0dfe6bc5787beea7332df3c8a460045e10c242e9a41d9563ad77"
+    },
+    {
+      "name": "core_block_0025.pt",
+      "nbytes": 139715116,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0025.pt",
+      "sha256": "3b1335fde8d0214a3ff426c3eb898f69316fbb1c5684c8f70d32076c090892d2"
+    },
+    {
+      "name": "core_block_0026.pt",
+      "nbytes": 139653428,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0026.pt",
+      "sha256": "adfe207c84afd0a7f667cb0fafe5d0d6aa7b20b750c439edf4d4049b3010afb2"
+    },
+    {
+      "name": "core_block_0027.pt",
+      "nbytes": 139831404,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/core_block_0027.pt",
+      "sha256": "6400727ff3f65f6fa8ead493e3df2f32c5e11dfb41ddeeb7846b76de5c62b7ae"
+    },
+    {
+      "name": "detachable_50m_base.pt",
+      "nbytes": 122167777,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_base.pt",
+      "sha256": "9e1199deba836ca8e470308512f0020f9be1771002c63a5edee9ed5f7ccf7951"
+    },
+    {
+      "name": "detachable_50m_block_0000.pt",
+      "nbytes": 2721464,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0000.pt",
+      "sha256": "1c63818f82265a3603d0b0bd343300d57212e98867c7eaab23cf0421e01f87f7"
+    },
+    {
+      "name": "detachable_50m_block_0001.pt",
+      "nbytes": 2699389,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0001.pt",
+      "sha256": "3b729e0575d67f9f6fe6210c4a930337d35601cc55248e32c99f93523a89ed00"
+    },
+    {
+      "name": "detachable_50m_block_0002.pt",
+      "nbytes": 2703536,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0002.pt",
+      "sha256": "a1c233877207b959cb5bb86912557e7a3c8c5d8b35ca2ddd569ce150edf8fea8"
+    },
+    {
+      "name": "detachable_50m_block_0003.pt",
+      "nbytes": 2703543,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0003.pt",
+      "sha256": "ea8bf499384ba87f6a928f374f98d1c0ea63324c36b0dca5c96eda369d13c803"
+    },
+    {
+      "name": "detachable_50m_block_0004.pt",
+      "nbytes": 2704984,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0004.pt",
+      "sha256": "69864aa89cc6ae0214c29693f6464d48cbb63ab0db289948123c63fddc454ec8"
+    },
+    {
+      "name": "detachable_50m_block_0005.pt",
+      "nbytes": 2710973,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0005.pt",
+      "sha256": "ba33a723b41b111ca301d34e6c8c42fb23fde58b521369c478f8541ed5fb6578"
+    },
+    {
+      "name": "detachable_50m_block_0006.pt",
+      "nbytes": 2720176,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0006.pt",
+      "sha256": "542ccf54197cbe27591f4ff6025c77b049bbdc890236184d7ff83d4ac17206cd"
+    },
+    {
+      "name": "detachable_50m_block_0007.pt",
+      "nbytes": 2708940,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0007.pt",
+      "sha256": "c78f79d8400a82b4fd14eaea1e709f54c43e0a5af8ef67821e7982262216fa64"
+    },
+    {
+      "name": "detachable_50m_block_0008.pt",
+      "nbytes": 2718141,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0008.pt",
+      "sha256": "9f082729fdc14347f65a307354a505196b1727ca5689de5dfc8fe51e6766a654"
+    },
+    {
+      "name": "detachable_50m_block_0009.pt",
+      "nbytes": 2714435,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0009.pt",
+      "sha256": "bbbb82162577a747b4f5383e18a62f0fbe423769adb328827dd3690c12ff7add"
+    },
+    {
+      "name": "detachable_50m_block_0010.pt",
+      "nbytes": 2708672,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0010.pt",
+      "sha256": "a116c6c916dbaa6df83f5ddf19b47868f6150075fee46575d042ae876c249f85"
+    },
+    {
+      "name": "detachable_50m_block_0011.pt",
+      "nbytes": 2719841,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0011.pt",
+      "sha256": "a48f31c0099a1b311022f59c887bfe1522d60420544a88abcb47690554b0bdef"
+    },
+    {
+      "name": "detachable_50m_block_0012.pt",
+      "nbytes": 2702589,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0012.pt",
+      "sha256": "7370ec79e68ab652988657fe830bed5c5f1879174e44213264dc1048c6407191"
+    },
+    {
+      "name": "detachable_50m_block_0013.pt",
+      "nbytes": 2698072,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0013.pt",
+      "sha256": "4db5953b836bcd36935eb4c6e56039eb529222ddf3f7af564d06ba9b566816b5"
+    },
+    {
+      "name": "detachable_50m_block_0014.pt",
+      "nbytes": 2704520,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0014.pt",
+      "sha256": "4c520525f1fa9da44f016c37b3e25cab618ae1ba972f49bf11a898daced8ca4c"
+    },
+    {
+      "name": "detachable_50m_block_0015.pt",
+      "nbytes": 2701222,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0015.pt",
+      "sha256": "46e17ce933dc26a528cc2fb4ecbf0e26a600937686273621969d61d43e83ad75"
+    },
+    {
+      "name": "detachable_50m_block_0016.pt",
+      "nbytes": 2699258,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0016.pt",
+      "sha256": "e543fc2cfe821abdaf21ad6d21efc70471d5c8dfea47b47fb65175c297dff92e"
+    },
+    {
+      "name": "detachable_50m_block_0017.pt",
+      "nbytes": 2702530,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0017.pt",
+      "sha256": "4c453b3f5a48cc49f516544b1fab5c53bf21e4dd1cb2d8996ac5ec3aef219c83"
+    },
+    {
+      "name": "detachable_50m_block_0018.pt",
+      "nbytes": 2708973,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0018.pt",
+      "sha256": "5889314f1b41efb0d21021d6a66242858459802c8851eea558b25f17b6f77a70"
+    },
+    {
+      "name": "detachable_50m_block_0019.pt",
+      "nbytes": 1970385,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0019.pt",
+      "sha256": "a215f5a2a513a2fd4ef1f656c8440829deced9da76f489b32f51171248a62f45"
+    },
+    {
+      "name": "detachable_50m_block_0020.pt",
+      "nbytes": 1968812,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0020.pt",
+      "sha256": "8b5a579a017de1df49f9245faa369d142fc119c2973b06b6000d2c547e4b986d"
+    },
+    {
+      "name": "detachable_50m_block_0021.pt",
+      "nbytes": 1970032,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/detachable_50m_block_0021.pt",
+      "sha256": "3c442c3285c34a1c24381a68419516915c4b5a1d561e8f1bedefada3fbb94848"
+    },
+    {
+      "name": "ar.pt",
+      "nbytes": 608655940,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/ar.pt",
+      "sha256": "e028c3b2cefb903adcdd88b1ee8883087d7f4b72d5e9af25d0db0e000c7f1f01"
+    },
+    {
+      "name": "sat.pt",
+      "nbytes": 608665321,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/sat.pt",
+      "sha256": "fabbc80e45391e1c73b7fd90fdf28ee6c3bec8b64c97be2b9c300e98690c6aaa"
+    },
+    {
+      "name": "nat.pt",
+      "nbytes": 608655940,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/nat.pt",
+      "sha256": "c8383219c442149473a7da8db4ef5d7b3763a2cfb8062c2627ae1928411fedb8"
+    },
+    {
+      "name": "opt.pt",
+      "nbytes": 1927264311,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/opt.pt",
+      "sha256": "d982e8f30adf03970349a5e7aa909c0eef30f2958d7a1c6336fdc50dbd382276"
+    },
+    {
+      "name": "scaler.pt",
+      "nbytes": 149,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/scaler.pt",
+      "sha256": "c2110dfd7b7a6c034fb62244a6f8f76a22d9936d55b6d0fe133e43e65b38631c"
+    },
+    {
+      "name": "agillm5_detachable_50m_bridge.pt",
+      "nbytes": 1224944,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/agillm5_detachable_50m_bridge.pt",
+      "sha256": "2ce02a3be18435aa6c5cc2c19ab25638d4f6e4971064056174f12f26da5864f7"
+    },
+    {
+      "name": "agillm5_detachable_50m_opt.pt",
+      "nbytes": 64974462,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/agillm5_detachable_50m_opt.pt",
+      "sha256": "6669e62a22ebc8ca78eb10b7e685c8ca88332e96c6cc28ada1a4cbc46f99d392"
+    },
+    {
+      "name": "agillm5_detachable_50m_scaler.pt",
+      "nbytes": 149,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/agillm5_detachable_50m_scaler.pt",
+      "sha256": "c2110dfd7b7a6c034fb62244a6f8f76a22d9936d55b6d0fe133e43e65b38631c"
+    },
+    {
+      "name": "agillm5_detachable_50m_rng.pt",
+      "nbytes": 3516,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/agillm5_detachable_50m_rng.pt",
+      "sha256": "fcd903ec1a26b4fe6ba9f0673814f22e7e53a73cad984090386dfd2c6e41e7e4"
+    },
+    {
+      "name": "agillm5_detachable_50m_counters.pt",
+      "nbytes": 30346,
+      "path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.shards/agillm5_detachable_50m_counters.pt",
+      "sha256": "6c41c4301bfc62e46b61ba3074fac7487ad1b1460e0a1440683e0acf9a5f7629"
+    }
+  ],
+  "step": 2842431,
+  "tokenizer_sidecar_nbytes": 17465552,
+  "tokenizer_sidecar_path": "/workspace/agillm43_production_v14/production_quality_v18_hard_spike_gates/checkpoints_continue400B_from200B/pretrain_step02842431_from02840270_20260911T043548_1789101348573779930_317951_000002.pt.tokenizer.json",
+  "tokenizer_sidecar_sha256": "2cccfc08e2419b199a97e7590a9f37ed886d2ebeb4e8c46d6bbc4948f74ba7a7"
+}