XPK Start: Sat Apr 18 00:58:36 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 00:59:00.429704: 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 00:59:00.965579 133320566032192 max_utils.py:273] Attempting to initialize the jax distributed system... I0418 00:59:10.007154 133320566032192 distributed.py:149] Starting JAX distributed service on [::]:8482 I0418 00:59:10.009399 133320566032192 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-3gzbs-slice-job-0-0.mt-11-optimizer-offload-3gzbs:8482 I0418 00:59:11.367934 133320566032192 max_utils.py:284] Jax distributed system initialized! I0418 00:59:16.608004 133320566032192 max_utils.py:800] System Information: Jax Version: 0.9.2 I0418 00:59:16.608105 133320566032192 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0418 00:59:16.608145 133320566032192 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 00:59:16.608179 133320566032192 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0418 00:59:17.295622 133320566032192 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0418 00:59:17.295915 133320566032192 checkpointing.py:677] Setting up checkpoint logger... I0418 00:59:17.295969 133320566032192 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0418 00:59:17.296015 133320566032192 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0418 00:59:17.296414 133320566032192 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 0x7940653485c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0418 00:59:20.150673 133320566032192 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0418 00:59:20.150913 133320566032192 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 0x792c2c3f5e80>}, handler_registry=None I0418 00:59:20.151158 133320566032192 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x792c2c3f5e80>` 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 00:59:20.151206 133320566032192 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x792c2c53a390>` 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 00:59:20.151240 133320566032192 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 0x792c2c3f5e80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x792c2c3f5e80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x792c2c53a390>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x792c2c53a390>}). I0418 00:59:20.151600 133320566032192 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0418 00:59:20.151670 133320566032192 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x792c2c2ab880> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0418 00:59:22.648817 133320566032192 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints I0418 00:59:22.665813 133320566032192 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=10), 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_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x792c2c52f200> I0418 00:59:22.665955 133320566032192 checkpointing.py:301] Checkpoint manager created! I0418 00:59:23.609428 133320566032192 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0418 00:59:23.609538 133320566032192 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 00:59:23.986197 133320566032192 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 00:59:23.986284 133320566032192 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 00:59:24.002331 133320566032192 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0418 00:59:24.002386 133320566032192 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0418 00:59:24.025854 133320566032192 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 00:59:24.025918 133320566032192 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 00:59:24.042021 133320566032192 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 00:59:24.042082 133320566032192 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 00:59:24.058145 133320566032192 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 00:59:24.058207 133320566032192 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 00:59:24.083222 133320566032192 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0418 00:59:24.083293 133320566032192 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0418 00:59:24.104081 133320566032192 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0418 00:59:24.104142 133320566032192 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0418 00:59:24.321490 133320566032192 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0418 00:59:24.321600 133320566032192 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0418 00:59:25.909931 133320566032192 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0418 00:59:25.910063 133320566032192 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 00:59:25.910117 133320566032192 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 00:59:25.910174 133320566032192 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 00:59:25.910224 133320566032192 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 00:59:25.910263 133320566032192 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 00:59:25.910317 133320566032192 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 00:59:25.910367 133320566032192 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 00:59:25.910405 133320566032192 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 00:59:25.910441 133320566032192 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 00:59:25.910488 133320566032192 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0418 00:59:25.910531 133320566032192 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0418 00:59:26.401260 133320566032192 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 00:59:26.401354 133320566032192 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0418 00:59:26.416715 133320566032192 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0418 00:59:26.416784 133320566032192 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0418 00:59:37.356876 133320566032192 max_utils.py:791] Total memory size: 1.3 GB, Output size: 0.1 GB, Temp size: 1.1 GB, Argument size: 0.1 GB, Host temp size: 0.0 GB. I0418 00:59:37.364804 133320566032192 metric_logger.py:301] number parameters: 1.104 billion I0418 00:59:48.864696 133320566032192 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0418 00:59:49.033522 133320566032192 checkpointing.py:776] Waited 0.16880297660827637 seconds for step 0 to finish before starting checkpointing. I0418 00:59:49.036348 133320566032192 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 00:59:49.038072 133320566032192 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0418 00:59:49.039357 133320566032192 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0. I0418 00:59:49.350886 133320566032192 signaling_client.py:364] Using JaxDistributedSignalingClient I0418 00:59:49.365778 133320566032192 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 00:59:49.365856 133320566032192 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 00:59:49.602812 133185686869760 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0 I0418 00:59:49.631386 133320566032192 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.266343s I0418 00:59:49.632168 133320566032192 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.631 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27402806282043457 s) (per-host) I0418 00:59:49.632242 133320566032192 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.274120s (batch_requests_ready=0.002445s, total_serialization_initiated=0.270964s, others=0.000711s) I0418 00:59:49.632325 133320566032192 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.278090s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.278074) I0418 00:59:49.633133 133320566032192 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.59 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0. I0418 00:59:49.633414 133186077128448 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:19:49.633383 I0418 00:59:49.644599 133320566032192 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0418 00:59:49.644931 133172880729856 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 00:59:49.645095 133320566032192 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776473989.0363297, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776473989.0381117, 'checkpointer_blocking_duration_secs': 0.5954527854919434, 'get_old_steps_start_time': 1776473989.6335902, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776473989.034049, 'checkpoint_manager_blocking_duration_secs': 0.6110012531280518} I0418 00:59:49.645201 133320566032192 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0418 00:59:49.645252 133320566032192 max_utils.py:750] Memstats: After params initialized: I0418 00:59:49.645300 133320566032192 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_0(process=0,(0,0,0,0)) I0418 00:59:49.645332 133320566032192 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_1(process=0,(1,0,0,0)) I0418 00:59:49.645359 133320566032192 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_4(process=0,(0,1,0,0)) I0418 00:59:49.645384 133320566032192 max_utils.py:756] Using (GB) 0.17 / 31.25 (0.544000%) on TPU_5(process=0,(1,1,0,0)) I0418 00:59:49.993126 133320566032192 metric_logger.py:196] completed step: 0, seconds: 11.194, TFLOP/s/device: 1.214, Tokens/s/device: 182.957, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 00:59:49.996188 133320566032192 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/tensorboard/' I0418 00:59:50.423355 133320566032192 metric_logger.py:196] completed step: 1, seconds: 1.127, TFLOP/s/device: 12.061, Tokens/s/device: 1817.949, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0418 00:59:50.584776 133320566032192 metric_logger.py:196] completed step: 2, seconds: 0.431, TFLOP/s/device: 31.527, Tokens/s/device: 4752.104, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0418 00:59:50.747701 133320566032192 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1550.865, Tokens/s/device: 233763.269, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.047 I0418 00:59:52.197688 133172872337152 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776473991782918118, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776473991782918118, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0418 00:59:52.458459 133185670084352 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0/items I0418 00:59:53.583547 2824 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0418 00:59:55.732666 133185403930368 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_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_0 I0418 01:00:12.898975 133320566032192 metric_logger.py:196] completed step: 4, seconds: 0.162, TFLOP/s/device: 83.665, Tokens/s/device: 12610.915, total_weights: 65536, loss: 9.278, lm_loss: 9.278, perplexity: 10694.968 I0418 01:00:12.907426 133320566032192 metric_logger.py:196] completed step: 5, seconds: 0.163, TFLOP/s/device: 83.490, Tokens/s/device: 12584.568, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.520 I0418 01:00:13.069123 133320566032192 metric_logger.py:196] completed step: 6, seconds: 22.150, TFLOP/s/device: 0.613, Tokens/s/device: 92.459, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.688 I0418 01:00:13.231781 133320566032192 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2227.033, Tokens/s/device: 335682.675, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4377.010 I0418 01:00:13.394382 133320566032192 metric_logger.py:196] completed step: 8, seconds: 0.164, TFLOP/s/device: 82.951, Tokens/s/device: 12503.205, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3845.943 I0418 01:00:13.556438 133320566032192 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0418 01:00:13.557182 133320566032192 checkpointing.py:776] Waited 0.0007581710815429688 seconds for step 9 to finish before starting checkpointing. I0418 01:00:13.559172 133320566032192 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:00:31.748306 133176080971520 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.164297s (commit=36.519707s, array_metadata_write=1.644590s) I0418 01:00:31.749387 133186077128448 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.273 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.391204833984375 s) (per-host) I0418 01:00:31.749509 133186077128448 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.115983s. I0418 01:00:32.431337 133186077128448 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776473991782918118, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0418 01:00:33.210261 133172872337152 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776473991782918118, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0418 01:00:33.398736 133186077128448 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 01:00:33.583976 133186077128448 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000265s. I0418 01:00:34.479599 133186077128448 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.921830s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0/items I0418 01:00:34.481381 133186077128448 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0/items I0418 01:00:35.031377 133186077128448 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0 I0418 01:00:36.438704 133186077128448 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_11_optimizer_offload_true/checkpoints/0`. I0418 01:00:36.439549 133186077128448 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 47.40 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/0 I0418 01:00:36.440814 133186077128448 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.807286s. I0418 01:00:36.441041 133172880729856 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:00:36.441160 133172880729856 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:00:36.441225 133172880729856 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:00:36.443101 133172880729856 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0418 01:00:36.443275 133320566032192 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0418 01:00:36.443421 133320566032192 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.884252 seconds. If this number is high, consider checkpointing less frequently. I0418 01:00:36.445338 133320566032192 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0418 01:00:36.447355 133320566032192 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9. I0418 01:00:37.167685 133320566032192 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0418 01:00:37.167777 133320566032192 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 01:00:37.187502 133320566032192 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.020491s I0418 01:00:37.188291 133320566032192 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 62.284 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.024773836135864258 s) (per-host) I0418 01:00:37.188364 133320566032192 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.024864s (batch_requests_ready=0.001807s, total_serialization_initiated=0.022347s, others=0.000709s) I0418 01:00:37.188446 133320566032192 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.028861s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.028843) I0418 01:00:37.189135 133320566032192 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9. I0418 01:00:37.189425 133172880729856 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 01:20:37.189390 I0418 01:00:37.205764 133320566032192 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0418 01:00:37.206034 133176147437312 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0418 01:00:37.206171 133320566032192 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776474013.559135, 'wait_for_prev_duration_secs': 22.884252071380615, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776474036.445379, 'checkpointer_blocking_duration_secs': 0.7441778182983398, 'get_old_steps_start_time': 1776474037.1895816, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776474013.557409, 'checkpoint_manager_blocking_duration_secs': 23.648722887039185} I0418 01:00:37.206278 133320566032192 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0418 01:00:37.206323 133320566032192 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0418 01:00:37.851685 133186077128448 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9 I0418 01:00:40.032958 133166421493504 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9/items I0418 01:00:44.208362 133176164222720 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_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_0 I0418 01:01:23.490829 133176155830016 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.482330s (commit=40.286147s, array_metadata_write=1.196183s) I0418 01:01:23.492008 133172880729856 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.105 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.328460454940796 s) (per-host) I0418 01:01:23.492136 133172880729856 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.302601s. I0418 01:01:25.600065 133172880729856 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0418 01:01:25.814567 133172880729856 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000233s. I0418 01:01:26.685144 133172880729856 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.903684s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9/items I0418 01:01:26.687066 133172880729856 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9/items I0418 01:01:27.256285 133172880729856 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9 I0418 01:01:28.688028 133172880729856 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_11_optimizer_offload_true/checkpoints/9`. I0418 01:01:28.688970 133172880729856 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.24 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_11_optimizer_offload_true/checkpoints/9 I0418 01:01:28.690327 133172880729856 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.500791s. I0418 01:01:28.690491 133176147437312 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0418 01:01:28.690568 133176147437312 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0418 01:01:28.690632 133176147437312 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0418 01:01:28.692703 133176147437312 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0418 01:01:28.692857 133320566032192 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0418 01:01:28.692949 133320566032192 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0418 01:01:28.693911 133320566032192 metric_logger.py:196] completed step: 9, seconds: 0.162, TFLOP/s/device: 83.991, Tokens/s/device: 12660.028, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.739 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sat Apr 18 01:01:37 UTC 2026 EXIT_CODE=0