XPK Start: Wed Apr 22 21:52:14 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. `rope_parameters`'s factor field must be a float >= 1, got 40 `rope_parameters`'s beta_fast field must be a float, got 32 `rope_parameters`'s beta_slow field must be a float, got 1 DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 2026-04-22 21:52:39.109158: 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) I0422 21:52:39.320524 138179387090752 max_utils.py:273] Attempting to initialize the jax distributed system... I0422 21:52:48.362895 138179387090752 distributed.py:149] Starting JAX distributed service on [::]:8482 I0422 21:52:48.365235 138179387090752 distributed.py:172] Connecting to JAX distributed service on mt-08-checkpoint-async--q0y3a-slice-job-0-0.mt-08-checkpoint-async--q0y3a:8482 I0422 21:52:49.338028 138179387090752 max_utils.py:284] Jax distributed system initialized! I0422 21:52:55.646547 138179387090752 max_utils.py:800] System Information: Jax Version: 0.9.2 I0422 21:52:55.646671 138179387090752 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0422 21:52:55.646712 138179387090752 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 I0422 21:52:55.646748 138179387090752 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 21:52:56.331813 138179387090752 maxtext_utils.py:1565] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 21:52:56.332091 138179387090752 checkpointing.py:677] Setting up checkpoint logger... I0422 21:52:56.332145 138179387090752 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 21:52:56.332188 138179387090752 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0422 21:52:56.332518 138179387090752 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7dabd21066c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 21:52:57.670626 138179387090752 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0422 21:52:57.670862 138179387090752 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 0x7d977c5e5730>}, handler_registry=None I0422 21:52:57.671111 138179387090752 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d977c5e5730>` 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`. I0422 21:52:57.671160 138179387090752 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d977c5e95b0>` 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`. I0422 21:52:57.671195 138179387090752 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 0x7d977c5e5730>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d977c5e5730>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d977c5e95b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d977c5e95b0>}). I0422 21:52:57.671505 138179387090752 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0422 21:52:57.671571 138179387090752 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d977c41b560> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0422 21:53:00.392771 138058317928192 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776894519212869878, 'commit_timestamp_nsecs': 1776894562457269810, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/0/_CHECKPOINT_METADATA I0422 21:53:00.641177 138058317928192 checkpoint_manager.py:1871] Missing metrics for step 0 I0422 21:53:00.939502 138179387090752 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints I0422 21:53:01.948263 138179387090752 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_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d977c5e7f50> I0422 21:53:01.948439 138179387090752 checkpointing.py:301] Checkpoint manager created! I0422 21:53:02.859600 138179387090752 nnx_wrappers.py:453] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 21:53:02.859720 138179387090752 nnx_wrappers.py:453] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 21:53:03.245815 138179387090752 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 21:53:03.245911 138179387090752 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 21:53:03.262928 138179387090752 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 21:53:03.262988 138179387090752 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 21:53:03.287141 138179387090752 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 21:53:03.287212 138179387090752 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 21:53:03.304236 138179387090752 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 21:53:03.304296 138179387090752 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 21:53:03.321338 138179387090752 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 21:53:03.321397 138179387090752 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 21:53:03.346934 138179387090752 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 21:53:03.347013 138179387090752 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 21:53:03.369504 138179387090752 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 21:53:03.369586 138179387090752 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 21:53:03.583229 138179387090752 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0422 21:53:03.583342 138179387090752 checkpointing.py:581] restoring from this run's directory step 9 I0422 21:53:03.827931 138179387090752 event_tracking.py:70] [process=0] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/9. I0422 21:53:04.078493 138179387090752 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/9. I0422 21:53:04.078943 138179387090752 event_tracking.py:125] [process=0] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/9. I0422 21:53:05.923192 1452 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 21:53:11.735547 138179387090752 jax_array_handlers.py:843] [process=0] /jax/orbax/read/worker/io/requested throughput: 271.589 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.818623065948486 s) (per-host) I0422 21:53:11.741459 138179387090752 base_pytree_checkpoint_handler.py:130] [process=0] /jax/checkpoint/read/gbytes_per_sec: 237.907 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.652166366577148 s) (per-host) I0422 21:53:13.357256 138179387090752 event_tracking.py:138] [process=0] [sync] Finished load in 9.53 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/9 I0422 21:53:13.357425 138179387090752 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints', 'checkpointer_start_time': 1776894783.8278315, 'checkpointer_duration_secs': 9.529560804367065, 'checkpoint_manager_start_time': 1776894783.583622, 'checkpoint_manager_duration_secs': 9.77377200126648} fsdp: 32 I0422 21:53:13.783220 138179387090752 maxtext_utils.py:1668] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0422 21:53:13.783369 138179387090752 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0422 21:53:13.783446 138179387090752 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0422 21:53:13.783537 138179387090752 maxtext_utils.py:1668] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0422 21:53:13.783633 138179387090752 maxtext_utils.py:1668] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0422 21:53:13.783699 138179387090752 maxtext_utils.py:1668] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0422 21:53:13.783781 138179387090752 maxtext_utils.py:1668] 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) I0422 21:53:13.783860 138179387090752 maxtext_utils.py:1668] 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') I0422 21:53:13.783921 138179387090752 maxtext_utils.py:1668] 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) I0422 21:53:13.783979 138179387090752 maxtext_utils.py:1668] 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) I0422 21:53:13.784055 138179387090752 maxtext_utils.py:1668] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0422 21:53:13.784141 138179387090752 maxtext_utils.py:1668] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0422 21:53:14.279738 138179387090752 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 21:53:14.279830 138179387090752 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0422 21:53:14.295747 138179387090752 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0422 21:53:14.295804 138179387090752 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0422 21:53:25.273079 138179387090752 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. I0422 21:53:25.281032 138179387090752 metric_logger.py:301] number parameters: 1.104 billion I0422 21:53:37.693058 138179387090752 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0422 21:53:37.846517 138179387090752 checkpointing.py:776] Waited 0.15343666076660156 seconds for step 10 to finish before starting checkpointing. I0422 21:53:37.849134 138179387090752 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 21:53:37.851251 138179387090752 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 10 I0422 21:53:37.852967 138179387090752 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10. I0422 21:53:38.573773 138179387090752 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 21:53:38.586483 138179387090752 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 21:53:38.586550 138179387090752 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 I0422 21:53:38.821835 138056287909632 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10 I0422 21:53:38.858781 138179387090752 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.272972s I0422 21:53:38.859646 138179387090752 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.549 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2780137062072754 s) (per-host) I0422 21:53:38.859724 138179387090752 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278114s (batch_requests_ready=0.002320s, total_serialization_initiated=0.275002s, others=0.000792s) I0422 21:53:38.859807 138179387090752 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.282088s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.282070) I0422 21:53:38.860571 138179387090752 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10. I0422 21:53:38.860940 138057923667712 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 22:13:38.860894 I0422 21:53:38.873287 138179387090752 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0422 21:53:38.873612 138053008631552 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 21:53:38.873769 138179387090752 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776894817.849109, 'wait_for_prev_duration_secs': 6.651878356933594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776894817.8513343, 'checkpointer_blocking_duration_secs': 1.0097577571868896, 'get_old_steps_start_time': 1776894818.8611155, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776894817.8470716, 'checkpoint_manager_blocking_duration_secs': 1.0266609191894531} I0422 21:53:38.873874 138179387090752 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0422 21:53:38.873924 138179387090752 max_utils.py:750] Memstats: After params initialized: I0422 21:53:38.873975 138179387090752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0422 21:53:38.874009 138179387090752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0422 21:53:38.874036 138179387090752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0422 21:53:38.874059 138179387090752 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0422 21:53:39.212519 138179387090752 metric_logger.py:196] completed step: 10, seconds: 12.105, TFLOP/s/device: 1.122, Tokens/s/device: 169.192, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0422 21:53:39.357564 138179387090752 metric_logger.py:196] completed step: 11, seconds: 1.518, TFLOP/s/device: 8.953, Tokens/s/device: 1349.560, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0422 21:53:39.505228 138179387090752 metric_logger.py:196] completed step: 12, seconds: 0.011, TFLOP/s/device: 1260.752, Tokens/s/device: 190034.332, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0422 21:53:39.652505 138179387090752 metric_logger.py:196] completed step: 13, seconds: 0.145, TFLOP/s/device: 93.418, Tokens/s/device: 14081.021, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0422 21:53:39.661811 138179387090752 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0422 21:53:39.946271 138179387090752 checkpointing.py:776] Waited 0.28430867195129395 seconds for step 15 to finish before starting checkpointing. I0422 21:53:39.949285 138179387090752 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 21:53:40.592365 138053561591552 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10/items I0422 21:53:40.762883 138052529813248 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776894820340230480, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776894820340230480, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10/_CHECKPOINT_METADATA I0422 21:53:43.766818 138053025416960 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10/items/array_metadatas/process_0 I0422 21:54:24.590713 138053017024256 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.610686s (commit=41.362871s, array_metadata_write=1.247814s) I0422 21:54:24.592184 138057923667712 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.331 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.010515213012695 s) (per-host) I0422 21:54:24.592353 138057923667712 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.731298s. I0422 21:54:26.355815 138057923667712 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000410s. I0422 21:54:26.434882 138057923667712 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0422 21:54:26.511372 138052529813248 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776894820340230480, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10/_CHECKPOINT_METADATA I0422 21:54:27.401340 138057923667712 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.936682s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10/items I0422 21:54:27.403722 138057923667712 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10/items I0422 21:54:27.946723 138057923667712 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10 I0422 21:54:29.340091 138057923667712 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10`. I0422 21:54:29.341232 138057923667712 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.49 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/10 I0422 21:54:29.342448 138057923667712 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.481388s. I0422 21:54:29.342679 138053008631552 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 21:54:29.342783 138053008631552 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 21:54:29.342923 138053008631552 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0422 21:54:29.345200 138053008631552 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0422 21:54:29.345532 138179387090752 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. W0422 21:54:29.345734 138179387090752 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.396451 seconds. If this number is high, consider checkpointing less frequently. I0422 21:54:29.347661 138179387090752 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 15 I0422 21:54:29.350094 138179387090752 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15. I0422 21:54:30.064445 138179387090752 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 21:54:30.064594 138179387090752 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 I0422 21:54:30.095162 138179387090752 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.031797s I0422 21:54:30.096041 138179387090752 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 37.971 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04062485694885254 s) (per-host) I0422 21:54:30.096155 138179387090752 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040756s (batch_requests_ready=0.004347s, total_serialization_initiated=0.035590s, others=0.000819s) I0422 21:54:30.096277 138179387090752 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045712s (all_items=0.000024s, per_item={'items': '0.00002384'}, temp_paths=0.045688) I0422 21:54:30.097310 138179387090752 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15. I0422 21:54:30.097689 138053008631552 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 22:14:30.097651 I0422 21:54:30.100194 138179387090752 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0422 21:54:30.100595 138053025416960 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 21:54:30.100826 138179387090752 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776894819.9492407, 'wait_for_prev_duration_secs': 49.396450996398926, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776894869.3477268, 'checkpointer_blocking_duration_secs': 0.75014328956604, 'get_old_steps_start_time': 1776894870.0978985, 'get_old_steps_duration_secs': 4.76837158203125e-05, 'checkpoint_manager_blocking_start_time': 1776894819.9469702, 'checkpoint_manager_blocking_duration_secs': 50.15380001068115} I0422 21:54:30.100981 138179387090752 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0422 21:54:30.304126 138049845458688 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15 I0422 21:54:32.056978 138053561591552 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15/items I0422 21:54:34.951546 138053569984256 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15/items/array_metadatas/process_0 I0422 21:54:52.036179 138179387090752 metric_logger.py:196] completed step: 14, seconds: 0.148, TFLOP/s/device: 91.791, Tokens/s/device: 13835.688, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0422 21:54:52.045359 138179387090752 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.526, Tokens/s/device: 13946.583, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0422 21:54:52.191008 138179387090752 metric_logger.py:196] completed step: 16, seconds: 72.382, TFLOP/s/device: 0.188, Tokens/s/device: 28.294, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0422 21:54:52.338355 138179387090752 metric_logger.py:196] completed step: 17, seconds: 0.007, TFLOP/s/device: 1909.646, Tokens/s/device: 287842.586, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0422 21:54:52.485634 138179387090752 metric_logger.py:196] completed step: 18, seconds: 0.147, TFLOP/s/device: 92.617, Tokens/s/device: 13960.178, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0422 21:54:52.632304 138179387090752 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0422 21:54:52.633080 138179387090752 checkpointing.py:776] Waited 0.0007967948913574219 seconds for step 19 to finish before starting checkpointing. I0422 21:54:52.635133 138179387090752 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 21:55:16.123186 138053553198848 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.952347s (commit=41.845423s, array_metadata_write=1.106924s) I0422 21:55:16.124301 138053008631552 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.288 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.068864583969116 s) (per-host) I0422 21:55:16.124432 138053008631552 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.026598s. I0422 21:55:18.350272 138053008631552 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000188s. I0422 21:55:18.714522 138053008631552 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0422 21:55:19.817711 138053008631552 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.469931s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15/items I0422 21:55:19.820212 138053008631552 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15/items I0422 21:55:20.345194 138053008631552 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15 I0422 21:55:22.181917 138053008631552 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15`. I0422 21:55:22.182790 138053008631552 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.84 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/15 I0422 21:55:22.184336 138053008631552 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.086500s. I0422 21:55:22.184472 138053025416960 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 21:55:22.184533 138053025416960 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 21:55:22.184610 138053025416960 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0422 21:55:22.187457 138053025416960 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0422 21:55:22.187632 138179387090752 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0422 21:55:22.187769 138179387090752 checkpoint_manager.py:1452] Waiting for previous save to complete took 29.552649 seconds. If this number is high, consider checkpointing less frequently. I0422 21:55:22.189221 138179387090752 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 19 I0422 21:55:22.191191 138179387090752 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19. I0422 21:55:22.904445 138179387090752 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0422 21:55:22.904544 138179387090752 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 I0422 21:55:22.930329 138179387090752 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026883s I0422 21:55:22.931041 138179387090752 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.268 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.031310081481933594 s) (per-host) I0422 21:55:22.931107 138179387090752 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031389s (batch_requests_ready=0.001916s, total_serialization_initiated=0.028811s, others=0.000661s) I0422 21:55:22.931177 138179387090752 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035320s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.035304) I0422 21:55:22.931846 138179387090752 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19. I0422 21:55:22.932187 138053025416960 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 22:15:22.932147 I0422 21:55:22.934212 138179387090752 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0422 21:55:22.934504 138053553198848 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 21:55:22.934652 138179387090752 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776894892.6350925, 'wait_for_prev_duration_secs': 29.552648782730103, 'time_between_consecutive_saves_sec': 23.28982162475586, 'checkpointer_blocking_start_time': 1776894922.1892598, 'checkpointer_blocking_duration_secs': 0.743070125579834, 'get_old_steps_start_time': 1776894922.9323485, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776894892.6333237, 'checkpoint_manager_blocking_duration_secs': 30.301294803619385} I0422 21:55:22.934808 138179387090752 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0422 21:55:22.934852 138179387090752 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 21:55:23.569507 138053008631552 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19 I0422 21:55:25.298249 138052521420544 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19/items I0422 21:55:28.288669 138053569984256 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19/items/array_metadatas/process_0 I0422 21:56:08.700194 138053561591552 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.184839s (commit=40.971407s, array_metadata_write=1.213432s) I0422 21:56:08.701273 138053025416960 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.488 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.801518201828 s) (per-host) I0422 21:56:08.701382 138053025416960 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.769077s. I0422 21:56:10.506190 138053025416960 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0422 21:56:10.599030 138053025416960 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000245s. I0422 21:56:11.464269 138053025416960 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.884682s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19/items I0422 21:56:11.465947 138053025416960 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19/items I0422 21:56:11.992913 138053025416960 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19 I0422 21:56:13.451353 138053025416960 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19`. I0422 21:56:13.452217 138053025416960 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 51.26 seconds @ gs://lance-maxtext/linen_ckpt_xpk_test_pipeline_scan_nnx_20260422_212603/linen_xpk_test_pipeline_scan_nnx_20260422_212603_08_checkpoint_async_true/checkpoints/19 I0422 21:56:13.453851 138053025416960 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.521541s. I0422 21:56:13.454087 138053553198848 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 21:56:13.454204 138053553198848 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 21:56:13.454256 138053553198848 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0422 21:56:13.455792 138053553198848 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0422 21:56:13.455982 138179387090752 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. I0422 21:56:13.456081 138179387090752 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 21:56:13.457150 138179387090752 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.192, Tokens/s/device: 13896.145, 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: Wed Apr 22 21:56:24 UTC 2026 EXIT_CODE=0