XPK Start: Sat Apr 18 01:38:45 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-18 01:39:09.641007: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0418 01:39:10.179515 134348740237120 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 01:39:19.220411 134348740237120 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 01:39:19.222753 134348740237120 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-rs-skwdt-slice-job-0-0.mt-15-ocdbt-false-rs-skwdt:8482 I0418 01:39:20.556404 134348740237120 max_utils.py:284] Jax distributed system initialized! I0418 01:39:26.734730 134348740237120 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 01:39:26.734856 134348740237120 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 01:39:26.734900 134348740237120 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0418 01:39:26.734937 134348740237120 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 01:39:27.420367 134348740237120 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 01:39:27.420643 134348740237120 checkpointing.py:677] Setting up checkpoint logger... I0418 01:39:27.420696 134348740237120 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True I0418 01:39:27.420743 134348740237120 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 01:39:27.421099 134348740237120 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7a2fc9351040>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 01:39:28.667818 134348740237120 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 01:39:28.668085 134348740237120 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1b5c50f230>}, handler_registry=None I0418 01:39:28.668322 134348740237120 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1b5c50f230>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0418 01:39:28.668370 134348740237120 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1b5c50e600>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0418 01:39:28.668406 134348740237120 composite_checkpoint_handler.py:505] Initialized registry DefaultCheckpointHandlerRegistry({('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1b5c50f230>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1b5c50f230>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1b5c50e600>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1b5c50e600>}). I0418 01:39:28.668727 134348740237120 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 01:39:28.668795 134348740237120 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a1b5c2bbc40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 01:39:31.453410 134226485688064 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776476228843135148, 'commit_timestamp_nsecs': 1776476247957189470, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/9/_CHECKPOINT_METADATA I0418 01:39:31.696455 134226485688064 checkpoint_manager.py:1871] Missing metrics for step 9 I0418 01:39:32.104952 134348740237120 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints I0418 01:39:33.406631 134348740237120 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a1b5c50d1f0> I0418 01:39:33.406805 134348740237120 checkpointing.py:301] Checkpoint manager created! I0418 01:39:34.335372 134348740237120 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 01:39:34.335486 134348740237120 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 01:39:34.715110 134348740237120 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 01:39:34.715200 134348740237120 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 01:39:34.731239 134348740237120 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 01:39:34.731296 134348740237120 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 01:39:34.754535 134348740237120 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 01:39:34.754600 134348740237120 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:39:34.770713 134348740237120 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 01:39:34.770774 134348740237120 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:39:34.786858 134348740237120 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0418 01:39:34.786919 134348740237120 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:39:34.816718 134348740237120 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 01:39:34.816819 134348740237120 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 01:39:34.838241 134348740237120 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 01:39:34.838303 134348740237120 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 01:39:35.056432 134348740237120 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 01:39:35.056539 134348740237120 checkpointing.py:581] restoring from this run's directory step 9 I0418 01:39:35.732493 134348740237120 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/9. I0418 01:39:35.975561 134348740237120 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/9. I0418 01:39:35.976008 134348740237120 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.24 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/9. I0418 01:39:39.434611 1497 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 01:39:48.646858 134348740237120 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 151.354 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.440935611724854 s) (per-host) I0418 01:39:48.654829 134348740237120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 143.677 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 11.014924764633179 s) (per-host) I0418 01:39:50.310443 134348740237120 event_tracking.py:138] [process=0] [sync] Finished load in 14.58 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/9 I0418 01:39:50.310601 134348740237120 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints', 'checkpointer_start_time': 1776476375.7323933, 'checkpointer_duration_secs': 14.578176021575928, 'checkpoint_manager_start_time': 1776476375.0567832, 'checkpoint_manager_duration_secs': 15.253788232803345} fsdp: 32 I0418 01:39:50.732170 134348740237120 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 01:39:50.732296 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 01:39:50.732350 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0418 01:39:50.732408 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0418 01:39:50.732465 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 01:39:50.732506 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0418 01:39:50.732561 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0418 01:39:50.732615 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0418 01:39:50.732655 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0418 01:39:50.732692 134348740237120 maxtext_utils.py:1654] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0418 01:39:50.732740 134348740237120 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 01:39:50.732789 134348740237120 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 01:39:51.222281 134348740237120 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 01:39:51.222377 134348740237120 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 01:39:51.237664 134348740237120 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 01:39:51.237728 134348740237120 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 01:40:02.106583 134348740237120 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0418 01:40:02.114420 134348740237120 metric_logger.py:301] number parameters: 1.104 billion I0418 01:40:13.661884 134348740237120 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0418 01:40:13.817305 134348740237120 checkpointing.py:776] Waited 0.15540456771850586 seconds for step 10 to finish before starting checkpointing. I0418 01:40:13.819903 134348740237120 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 01:40:13.821423 134348740237120 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0418 01:40:13.822798 134348740237120 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10. I0418 01:40:14.555209 134348740237120 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 01:40:14.567991 134348740237120 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:40:14.568057 134348740237120 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0418 01:40:14.831959 134223717439232 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10 I0418 01:40:14.849131 134348740237120 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281796s I0418 01:40:14.849926 134348740237120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.377 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28690671920776367 s) (per-host) I0418 01:40:14.850014 134348740237120 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.287017s (batch_requests_ready=0.002266s, total_serialization_initiated=0.284026s, others=0.000726s) I0418 01:40:14.850099 134348740237120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.290943s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.290925) I0418 01:40:14.850884 134348740237120 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.03 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10. I0418 01:40:14.851234 134222115219200 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:00:14.851193 I0418 01:40:14.854489 134348740237120 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:40:14.854796 134223725831936 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:40:14.854962 134348740237120 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776476413.819885, 'wait_for_prev_duration_secs': 5.817413330078125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776476413.8214586, 'checkpointer_blocking_duration_secs': 1.0299310684204102, 'get_old_steps_start_time': 1776476414.851412, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776476413.8178337, 'checkpoint_manager_blocking_duration_secs': 1.0370943546295166} I0418 01:40:14.855059 134348740237120 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0418 01:40:14.855109 134348740237120 max_utils.py:750] Memstats: After params initialized: I0418 01:40:14.855160 134348740237120 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0418 01:40:14.855195 134348740237120 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0418 01:40:14.855221 134348740237120 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0418 01:40:14.855246 134348740237120 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0418 01:40:15.199144 134348740237120 metric_logger.py:196] completed step: 10, seconds: 11.245, TFLOP/s/device: 1.208, Tokens/s/device: 182.118, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0418 01:40:15.344303 134348740237120 metric_logger.py:196] completed step: 11, seconds: 1.535, TFLOP/s/device: 8.849, Tokens/s/device: 1333.880, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0418 01:40:15.491539 134348740237120 metric_logger.py:196] completed step: 12, seconds: 0.011, TFLOP/s/device: 1287.391, Tokens/s/device: 194049.649, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0418 01:40:15.638874 134348740237120 metric_logger.py:196] completed step: 13, seconds: 0.146, TFLOP/s/device: 93.324, Tokens/s/device: 14066.804, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0418 01:40:15.648068 134348740237120 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0418 01:40:15.933098 134348740237120 checkpointing.py:776] Waited 0.2848801612854004 seconds for step 15 to finish before starting checkpointing. I0418 01:40:15.936082 134348740237120 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:40:17.010469 134222157182720 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/items I0418 01:40:17.165166 134222106826496 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776476416739818650, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776476416739818650, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0418 01:40:20.945328 134222140397312 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/items/array_metadatas/process_0 I0418 01:40:28.174547 134222123611904 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 9.810223s (commit=8.677151s, array_metadata_write=1.133072s) I0418 01:40:28.176037 134222115219200 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 116.036 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.612985849380493 s) (per-host) I0418 01:40:28.176212 134222115219200 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 13.324855s. W0418 01:40:29.575350 134222115219200 ocdbt_utils.py:168] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/items, I0418 01:40:29.632514 134222106826496 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776476416739818650, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/_CHECKPOINT_METADATA I0418 01:40:31.279729 134222115219200 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000391s. I0418 01:40:31.279983 134222115219200 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.299671s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/items I0418 01:40:31.282100 134222115219200 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10/items I0418 01:40:32.220276 134222115219200 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10 I0418 01:40:34.485996 134222115219200 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10`. I0418 01:40:34.487149 134222115219200 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 20.67 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/10 I0418 01:40:34.488785 134222115219200 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 19.637424s. I0418 01:40:34.489115 134223725831936 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:40:34.489280 134223725831936 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:40:34.489391 134223725831936 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:40:34.491001 134223725831936 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0418 01:40:34.491321 134348740237120 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0418 01:40:34.491493 134348740237120 checkpoint_manager.py:1452] Waiting for previous save to complete took 18.555425 seconds. If this number is high, consider checkpointing less frequently. I0418 01:40:34.493198 134348740237120 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0418 01:40:34.495613 134348740237120 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15. I0418 01:40:35.208888 134348740237120 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:40:35.209020 134348740237120 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0418 01:40:35.236495 134348740237120 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028614s I0418 01:40:35.237446 134348740237120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.371 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03728628158569336 s) (per-host) I0418 01:40:35.237563 134348740237120 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037427s (batch_requests_ready=0.004214s, total_serialization_initiated=0.032360s, others=0.000854s) I0418 01:40:35.237699 134348740237120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043380s (all_items=0.000023s, per_item={'items': '0.00002265'}, temp_paths=0.043358) I0418 01:40:35.238711 134348740237120 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15. I0418 01:40:35.239164 134223725831936 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:00:35.239124 I0418 01:40:35.241671 134348740237120 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:40:35.242047 134222123611904 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:40:35.242280 134348740237120 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776476415.9360383, 'wait_for_prev_duration_secs': 18.55542492866516, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776476434.493264, 'checkpointer_blocking_duration_secs': 0.7460930347442627, 'get_old_steps_start_time': 1776476435.2393835, 'get_old_steps_duration_secs': 4.458427429199219e-05, 'checkpoint_manager_blocking_start_time': 1776476415.933762, 'checkpoint_manager_blocking_duration_secs': 19.30846357345581} I0418 01:40:35.242433 134348740237120 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0418 01:40:35.880254 134222115219200 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15 I0418 01:40:40.197369 134222140397312 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15/items I0418 01:40:44.345398 134222668875520 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15/items/array_metadatas/process_0 I0418 01:40:52.191562 134222148790016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 10.462240s (commit=8.589261s, array_metadata_write=1.872979s) I0418 01:40:52.192605 134223725831936 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 92.959 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 16.992429494857788 s) (per-host) I0418 01:40:52.192735 134223725831936 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 16.953421s. W0418 01:40:54.023916 134223725831936 ocdbt_utils.py:168] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15/items, I0418 01:40:54.724055 134223725831936 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000267s. I0418 01:40:54.724237 134223725831936 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.295424s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15/items I0418 01:40:54.725641 134223725831936 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15/items I0418 01:40:55.225454 134223725831936 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15 I0418 01:40:56.917109 134223725831936 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15`. I0418 01:40:56.917967 134223725831936 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 22.42 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/15 I0418 01:40:56.919392 134223725831936 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 21.680075s. I0418 01:40:57.208297 134222123611904 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:40:57.208411 134222123611904 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:40:57.208481 134222123611904 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:40:57.210182 134222123611904 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0418 01:40:57.295857 134348740237120 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.372, Tokens/s/device: 13923.259, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0418 01:40:57.304818 134348740237120 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.140, Tokens/s/device: 13888.324, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0418 01:40:57.450365 134348740237120 metric_logger.py:196] completed step: 16, seconds: 41.655, TFLOP/s/device: 0.326, Tokens/s/device: 49.165, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0418 01:40:57.597599 134348740237120 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1921.257, Tokens/s/device: 289592.760, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0418 01:40:57.745075 134348740237120 metric_logger.py:196] completed step: 18, seconds: 0.147, TFLOP/s/device: 92.628, Tokens/s/device: 13961.891, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0418 01:40:57.891720 134348740237120 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0418 01:40:57.892426 134348740237120 checkpointing.py:776] Waited 0.0007259845733642578 seconds for step 19 to finish before starting checkpointing. I0418 01:40:57.894478 134348740237120 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 01:40:57.896002 134348740237120 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0418 01:40:57.897394 134348740237120 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19. I0418 01:40:59.031247 134348740237120 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:40:59.031345 134348740237120 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0418 01:40:59.056672 134348740237120 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026131s I0418 01:40:59.057443 134348740237120 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 50.260 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030691862106323242 s) (per-host) I0418 01:40:59.057520 134348740237120 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030789s (batch_requests_ready=0.001918s, total_serialization_initiated=0.028179s, others=0.000692s) I0418 01:40:59.057604 134348740237120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034692s (all_items=0.000015s, per_item={'items': '0.00001550'}, temp_paths=0.034676) I0418 01:40:59.058305 134348740237120 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.16 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19. I0418 01:40:59.058611 134222123611904 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 02:00:59.058576 I0418 01:40:59.060520 134348740237120 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:40:59.060761 134222140397312 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:40:59.060893 134348740237120 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776476457.8944411, 'wait_for_prev_duration_secs': 7.534027099609375e-05, 'time_between_consecutive_saves_sec': 0.684197187423706, 'checkpointer_blocking_start_time': 1776476457.8960395, 'checkpointer_blocking_duration_secs': 1.1626994609832764, 'get_old_steps_start_time': 1776476459.0587566, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776476457.8926604, 'checkpoint_manager_blocking_duration_secs': 1.1682000160217285} I0418 01:40:59.061006 134348740237120 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0418 01:40:59.061054 134348740237120 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:40:59.259714 134223725831936 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19 I0418 01:41:00.983775 134221058262784 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19/items I0418 01:41:04.240861 134222157182720 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19/items/array_metadatas/process_0 I0418 01:41:12.061521 134222148790016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 9.546952s (commit=8.019243s, array_metadata_write=1.527709s) I0418 01:41:12.062546 134222123611904 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 121.174 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 13.035794258117676 s) (per-host) I0418 01:41:12.062610 134222123611904 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 13.003895s. W0418 01:41:14.742153 134222123611904 ocdbt_utils.py:168] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19/items, I0418 01:41:16.044189 134222123611904 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000244s. I0418 01:41:16.044370 134222123611904 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.915291s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19/items I0418 01:41:16.046068 134222123611904 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19/items I0418 01:41:16.987478 134222123611904 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19 I0418 01:41:19.166189 134222123611904 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19`. I0418 01:41:19.167075 134222123611904 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 21.27 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_15_ocdbt_false/checkpoints/19 I0418 01:41:19.168247 134222123611904 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 20.109529s. I0418 01:41:19.168468 134222140397312 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:41:19.168582 134222140397312 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:41:19.168633 134222140397312 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:41:19.170623 134222140397312 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0418 01:41:19.170770 134348740237120 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0418 01:41:19.170875 134348740237120 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 01:41:19.171872 134348740237120 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.307, Tokens/s/device: 13913.516, total_weights: 65536, loss: 5.391, lm_loss: 5.391, perplexity: 219.349 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 01:41:31 UTC 2026 EXIT_CODE=0