XPK Start: Sat Apr 18 14:30:52 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 14:31:16.734311: 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 14:31:16.917311 134224774457152 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 14:31:25.958578 134224774457152 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 14:31:25.960871 134224774457152 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-true-rs-sjsif-slice-job-0-0.mt-16-zarr3-true-rs-sjsif:8482 I0418 14:31:26.592352 134224774457152 max_utils.py:284] Jax distributed system initialized! I0418 14:31:32.780244 134224774457152 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 14:31:32.780346 134224774457152 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 14:31:32.780385 134224774457152 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 14:31:32.780418 134224774457152 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 14:31:33.472190 134224774457152 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 14:31:33.472482 134224774457152 checkpointing.py:677] Setting up checkpoint logger... I0418 14:31:33.472537 134224774457152 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 14:31:33.472581 134224774457152 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 14:31:33.472910 134224774457152 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 0x7a1306314200>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 14:31:35.156835 134224774457152 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 14:31:35.157073 134224774457152 checkpoint_manager.py:708] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79fed0442a20>}, handler_registry=None I0418 14:31:35.157308 134224774457152 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79fed0442a20>` 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 14:31:35.157356 134224774457152 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79fed0442120>` 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 14:31:35.157392 134224774457152 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 0x79fed0442a20>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79fed0442a20>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79fed0442120>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79fed0442120>}). I0418 14:31:35.157737 134224774457152 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 14:31:35.157806 134224774457152 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79fde076c180> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 14:31:37.501302 134103693252352 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776522479032192089, 'commit_timestamp_nsecs': 1776522524566640015, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/5/_CHECKPOINT_METADATA I0418 14:31:37.763128 134103693252352 checkpoint_manager.py:1871] Missing metrics for step 5 I0418 14:31:38.198780 134224774457152 checkpoint_manager.py:1812] Found 3 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints I0418 14:31:39.055514 134224774457152 checkpoint_manager.py:929] [process=5][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_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79fed04412e0> I0418 14:31:39.055707 134224774457152 checkpointing.py:301] Checkpoint manager created! I0418 14:31:39.976650 134224774457152 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 14:31:39.976764 134224774457152 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 14:31:40.351954 134224774457152 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 14:31:40.352042 134224774457152 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 14:31:40.368063 134224774457152 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 14:31:40.368121 134224774457152 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 14:31:40.391590 134224774457152 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 14:31:40.391653 134224774457152 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 14:31:40.407892 134224774457152 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 14:31:40.407952 134224774457152 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 14:31:40.424106 134224774457152 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 14:31:40.424166 134224774457152 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 14:31:40.449354 134224774457152 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 14:31:40.449426 134224774457152 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 14:31:40.470139 134224774457152 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 14:31:40.470203 134224774457152 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 14:31:40.683372 134224774457152 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 14:31:40.683491 134224774457152 checkpointing.py:581] restoring from this run's directory step 9 I0418 14:31:40.935314 134224774457152 event_tracking.py:70] [process=5] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/9. I0418 14:31:41.186502 134224774457152 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/9. I0418 14:31:41.186934 134224774457152 event_tracking.py:125] [process=5] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/9. I0418 14:31:42.586800 1447 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 14:31:48.055993 134224774457152 jax_array_handlers.py:843] [process=5] /jax/orbax/read/worker/io/requested throughput: 288.635 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.474980592727661 s) (per-host) I0418 14:31:48.060823 134224774457152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/checkpoint/read/gbytes_per_sec: 269.061 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 5.881914138793945 s) (per-host) I0418 14:31:51.154550 134224774457152 event_tracking.py:138] [process=5] [sync] Finished load in 10.22 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/9 I0418 14:31:51.154704 134224774457152 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints', 'checkpointer_start_time': 1776522700.93522, 'checkpointer_duration_secs': 10.219454288482666, 'checkpoint_manager_start_time': 1776522700.6837358, 'checkpoint_manager_duration_secs': 10.470940589904785} fsdp: 32 I0418 14:31:51.565157 134224774457152 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 14:31:51.565282 134224774457152 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 14:31:51.565335 134224774457152 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 14:31:51.565392 134224774457152 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 14:31:51.565442 134224774457152 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 14:31:51.565496 134224774457152 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 14:31:51.565550 134224774457152 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 14:31:51.565603 134224774457152 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 14:31:51.565643 134224774457152 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 14:31:51.565679 134224774457152 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 14:31:51.565729 134224774457152 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 14:31:51.565777 134224774457152 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 14:31:52.051737 134224774457152 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 14:31:52.051832 134224774457152 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 14:31:52.067028 134224774457152 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 14:31:52.067086 134224774457152 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 14:32:03.022832 134224774457152 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 14:32:03.023651 134224774457152 metric_logger.py:301] number parameters: 1.104 billion I0418 14:32:14.321513 134224774457152 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0418 14:32:14.477281 134224774457152 checkpointing.py:776] Waited 0.15574955940246582 seconds for step 10 to finish before starting checkpointing. I0418 14:32:14.479548 134224774457152 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 14:32:14.481377 134224774457152 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 10 I0418 14:32:14.482962 134224774457152 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/10. I0418 14:32:15.253809 134224774457152 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 14:32:15.254914 134224774457152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 14:32:15.254972 134224774457152 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 14:32:15.535707 134224774457152 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.281972s I0418 14:32:15.535874 134224774457152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 5.396 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2858603000640869 s) (per-host) I0418 14:32:15.535930 134224774457152 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285923s (batch_requests_ready=0.002186s, total_serialization_initiated=0.283666s, others=0.000072s) I0418 14:32:15.536006 134224774457152 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289967s (all_items=0.000024s, per_item={'items': '0.00002408'}, temp_paths=0.289943) I0418 14:32:15.536727 134224774457152 event_tracking.py:125] [process=5] [async] Finished blocking save in 1.06 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/10. I0418 14:32:15.537024 134103156381440 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 14:52:15.536993 I0418 14:32:15.543127 134224774457152 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0418 14:32:15.543372 134101562554112 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0418 14:32:15.543507 134224774457152 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776522734.4795294, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776522734.4814138, 'checkpointer_blocking_duration_secs': 1.0557618141174316, 'get_old_steps_start_time': 1776522735.5371988, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776522734.4777908, 'checkpoint_manager_blocking_duration_secs': 1.0656812191009521} I0418 14:32:15.543617 134224774457152 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0418 14:32:15.543672 134224774457152 max_utils.py:750] Memstats: After params initialized: I0418 14:32:15.543719 134224774457152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0418 14:32:15.543751 134224774457152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0418 14:32:15.543781 134224774457152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0418 14:32:15.543804 134224774457152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0418 14:32:15.857160 134224774457152 metric_logger.py:196] completed step: 10, seconds: 11.298, TFLOP/s/device: 1.203, Tokens/s/device: 181.275, total_weights: 65536, loss: 6.833, lm_loss: 6.833, perplexity: 927.869 I0418 14:32:16.035878 134224774457152 metric_logger.py:196] completed step: 11, seconds: 1.534, TFLOP/s/device: 8.856, Tokens/s/device: 1334.902, total_weights: 65536, loss: 6.527, lm_loss: 6.527, perplexity: 683.659 I0418 14:32:16.183352 134224774457152 metric_logger.py:196] completed step: 12, seconds: 0.031, TFLOP/s/device: 431.666, Tokens/s/device: 65065.447, total_weights: 65536, loss: 6.265, lm_loss: 6.265, perplexity: 526.068 I0418 14:32:16.330651 134224774457152 metric_logger.py:196] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.709, Tokens/s/device: 13371.114, total_weights: 65536, loss: 6.045, lm_loss: 6.045, perplexity: 422.001 I0418 14:32:16.335170 134224774457152 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0418 14:32:16.624834 134224774457152 checkpointing.py:776] Waited 0.2896299362182617 seconds for step 15 to finish before starting checkpointing. I0418 14:32:16.627734 134224774457152 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 14:32:20.462283 134102585939712 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/10/items/array_metadatas/process_5 I0418 14:32:57.427839 134103156381440 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.451 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.17778778076172 s) (per-host) I0418 14:32:57.427966 134103156381440 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.890830s. I0418 14:33:05.079255 134103156381440 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 49.542102s. I0418 14:33:05.079561 134101562554112 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 14:33:05.079694 134101562554112 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0418 14:33:05.079745 134101562554112 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0418 14:33:05.081231 134101562554112 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0418 14:33:05.081421 134224774457152 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0418 14:33:05.081575 134224774457152 checkpoint_manager.py:1452] Waiting for previous save to complete took 48.453839 seconds. If this number is high, consider checkpointing less frequently. I0418 14:33:05.083594 134224774457152 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 15 I0418 14:33:05.085962 134224774457152 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/15. I0418 14:33:05.818817 134224774457152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 14:33:05.818911 134224774457152 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 14:33:05.844745 134224774457152 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027030s I0418 14:33:05.844887 134224774457152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 50.589 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.030492067337036133 s) (per-host) I0418 14:33:05.844942 134224774457152 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030551s (batch_requests_ready=0.001795s, total_serialization_initiated=0.028691s, others=0.000065s) I0418 14:33:05.845022 134224774457152 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035640s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=0.035627) I0418 14:33:05.845685 134224774457152 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/15. I0418 14:33:05.846009 134096680359680 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 14:53:05.845970 I0418 14:33:05.859494 134224774457152 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0418 14:33:05.859800 134101562554112 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0418 14:33:05.859960 134224774457152 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776522736.6277053, 'wait_for_prev_duration_secs': 48.45383906364441, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776522785.0836318, 'checkpointer_blocking_duration_secs': 0.7625267505645752, 'get_old_steps_start_time': 1776522785.8461819, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776522736.625408, 'checkpoint_manager_blocking_duration_secs': 49.23451852798462} I0418 14:33:05.860081 134224774457152 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0418 14:33:05.860917 134224774457152 metric_logger.py:196] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.132, Tokens/s/device: 13887.194, total_weights: 65536, loss: 5.864, lm_loss: 5.864, perplexity: 352.193 I0418 14:33:05.867044 134224774457152 metric_logger.py:196] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.264, Tokens/s/device: 13907.091, total_weights: 65536, loss: 5.719, lm_loss: 5.719, perplexity: 304.661 I0418 14:33:10.993437 134102585939712 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/15/items/array_metadatas/process_5 I0418 14:33:26.591469 134224774457152 metric_logger.py:196] completed step: 16, seconds: 49.530, TFLOP/s/device: 0.274, Tokens/s/device: 41.348, total_weights: 65536, loss: 5.605, lm_loss: 5.605, perplexity: 271.872 I0418 14:33:26.738779 134224774457152 metric_logger.py:196] completed step: 17, seconds: 20.578, TFLOP/s/device: 0.660, Tokens/s/device: 99.522, total_weights: 65536, loss: 5.517, lm_loss: 5.517, perplexity: 248.975 I0418 14:33:26.886191 134224774457152 metric_logger.py:196] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.135, Tokens/s/device: 13435.323, total_weights: 65536, loss: 5.448, lm_loss: 5.448, perplexity: 232.333 I0418 14:33:27.032916 134224774457152 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0418 14:33:27.033644 134224774457152 checkpointing.py:776] Waited 0.0007483959197998047 seconds for step 19 to finish before starting checkpointing. I0418 14:33:27.035610 134224774457152 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 14:33:48.165082 134096680359680 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.298 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.350650787353516 s) (per-host) I0418 14:33:48.165197 134096680359680 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.319069s. I0418 14:33:54.886879 134096680359680 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 49.040735s. I0418 14:33:54.887165 134101562554112 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 14:33:54.887286 134101562554112 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0418 14:33:54.887336 134101562554112 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0418 14:33:54.888851 134101562554112 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0418 14:33:54.889044 134224774457152 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0418 14:33:54.889179 134224774457152 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.853574 seconds. If this number is high, consider checkpointing less frequently. I0418 14:33:54.890790 134224774457152 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 19 I0418 14:33:54.892856 134224774457152 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/19. I0418 14:33:56.049067 134224774457152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 14:33:56.049161 134224774457152 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 14:33:56.074341 134224774457152 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026204s I0418 14:33:56.074486 134224774457152 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 51.996 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.029667377471923828 s) (per-host) I0418 14:33:56.074535 134224774457152 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.029726s (batch_requests_ready=0.001651s, total_serialization_initiated=0.028001s, others=0.000074s) I0418 14:33:56.074610 134224774457152 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034077s (all_items=0.000011s, per_item={'items': '0.00001073'}, temp_paths=0.034066) I0418 14:33:56.075258 134224774457152 event_tracking.py:125] [process=5] [async] Finished blocking save in 1.18 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/19. I0418 14:33:56.075590 134101562554112 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 14:53:56.075553 I0418 14:33:56.507807 134224774457152 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0418 14:33:56.508183 134095622629120 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0418 14:33:56.508359 134224774457152 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776522807.0355809, 'wait_for_prev_duration_secs': 27.8535737991333, 'time_between_consecutive_saves_sec': 21.954307079315186, 'checkpointer_blocking_start_time': 1776522834.8908298, 'checkpointer_blocking_duration_secs': 1.1849052906036377, 'get_old_steps_start_time': 1776522836.0757518, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1776522807.0338786, 'checkpoint_manager_blocking_duration_secs': 29.47444438934326} I0418 14:33:56.508487 134224774457152 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0418 14:33:56.508532 134224774457152 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 14:34:02.046145 134102585939712 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260418_125719/linen_xpk_main_20260418_125719_16_zarr3_true/checkpoints/19/items/array_metadatas/process_5 I0418 14:34:38.588140 134101562554112 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.129 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.54328179359436 s) (per-host) I0418 14:34:38.588259 134101562554112 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.512556s. I0418 14:34:46.976239 134101562554112 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 50.900517s. I0418 14:34:46.976535 134095622629120 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 14:34:46.976655 134095622629120 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0418 14:34:46.976701 134095622629120 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0418 14:34:46.978516 134095622629120 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0418 14:34:46.982784 134224774457152 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0418 14:34:46.982938 134224774457152 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 14:34:46.983854 134224774457152 metric_logger.py:196] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.363, Tokens/s/device: 13921.934, 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 14:34:57 UTC 2026 EXIT_CODE=0