XPK Start: Mon Apr 20 20:44:58 UTC 2026 2026-04-20 20:45:02.040952: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1776717902.053400 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776717902.057088 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776717902.068290 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776717902.068309 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776717902.068311 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776717902.068313 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 20:45:21.222894: 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) I0420 20:45:21.749771 138014997006144 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 20:45:30,790:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 20:45:30.790949 138014997006144 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 20:45:30,795:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-false-sv-4sbgn-slice-job-0-0.mt-16-zarr3-false-sv-4sbgn:8482 I0420 20:45:30.795481 138014997006144 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-false-sv-4sbgn-slice-job-0-0.mt-16-zarr3-false-sv-4sbgn:8482 I0420 20:45:31.657427 138014997006144 max_utils.py:284] Jax distributed system initialized! I0420 20:45:38.450385 138014997006144 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 20:45:38.450490 138014997006144 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 20:45:38.450529 138014997006144 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0420 20:45:38.450566 138014997006144 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 20:45:39.077009 138014997006144 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:45:39.077598 138014997006144 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 20:45:39.077792 138014997006144 checkpointing.py:688] Setting up checkpoint logger... I0420 20:45:39.077844 138014997006144 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0420 20:45:39.077889 138014997006144 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 20:45:39.078237 138014997006144 base_pytree_checkpoint_handler.py:415] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7d8595775400>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 20:45:42.068264 138014997006144 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 20:45:42.068601 138014997006144 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d85d8957da0>}, handler_registry=None I0420 20:45:42.068850 138014997006144 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d85d8957da0>` 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`. I0420 20:45:42.068901 138014997006144 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d72d3cb2e10>` 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`. I0420 20:45:42.068940 138014997006144 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 0x7d85d8957da0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d85d8957da0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d72d3cb2e10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d72d3cb2e10>}). I0420 20:45:42.069453 138014997006144 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 20:45:42.069532 138014997006144 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d72d2d94180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 20:45:43.190130 138014997006144 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints I0420 20:45:43.614914 138014997006144 checkpoint_manager.py:929] [process=7][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_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d72d3cadf40> I0420 20:45:43.615100 138014997006144 checkpointing.py:302] Checkpoint manager created! I0420 20:45:44.500715 138014997006144 nnx_wrappers.py:455] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 20:45:44.500821 138014997006144 nnx_wrappers.py:455] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:45:44.869063 138014997006144 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:45:44.869160 138014997006144 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:45:44.885055 138014997006144 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 20:45:44.885115 138014997006144 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 20:45:44.907930 138014997006144 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:45:44.908004 138014997006144 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:45:44.923984 138014997006144 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:45:44.924045 138014997006144 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:45:44.939863 138014997006144 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 20:45:44.939924 138014997006144 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:45:44.964684 138014997006144 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 20:45:44.964755 138014997006144 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 20:45:44.984539 138014997006144 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 20:45:44.984608 138014997006144 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 20:45:45.173216 138014997006144 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 20:45:45.173328 138014997006144 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 20:45:46.602081 138014997006144 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0420 20:45:46.602212 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0420 20:45:46.602268 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0420 20:45:46.602327 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0420 20:45:46.602379 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0420 20:45:46.602417 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0420 20:45:46.602470 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0420 20:45:46.602523 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0420 20:45:46.602564 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0420 20:45:46.602600 138014997006144 maxtext_utils.py:1740] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0420 20:45:46.602650 138014997006144 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', None) I0420 20:45:46.602768 138014997006144 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: (None, 'fsdp') I0420 20:45:47.110051 138014997006144 train.py:158] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:45:47.110145 138014997006144 train.py:158] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:45:47.125265 138014997006144 train.py:165] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 20:45:47.125331 138014997006144 train.py:165] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 20:45:57.865434 138014997006144 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. I0420 20:45:57.866456 138014997006144 metric_logger.py:289] number parameters: 1.104 billion I0420 20:46:09.190367 138014997006144 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 20:46:09.344333 138014997006144 checkpointing.py:798] Waited 0.1539468765258789 seconds for step 0 to finish before starting checkpointing. I0420 20:46:09.347003 138014997006144 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:46:09.348619 138014997006144 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0420 20:46:09.351437 138014997006144 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/0. I0420 20:46:09.972034 138014997006144 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 20:46:09.973097 138014997006144 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:46:09.973153 138014997006144 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 I0420 20:46:10.223332 138014997006144 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.251377s I0420 20:46:10.223509 138014997006144 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.824 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5462057590484619 s) (per-host) I0420 20:46:10.223564 138014997006144 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.546269s (batch_requests_ready=0.291271s, total_serialization_initiated=0.254926s, others=0.000072s) I0420 20:46:10.223671 138014997006144 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.550545s (all_items=0.000023s, per_item={'items': '0.00002337'}, temp_paths=0.550522) I0420 20:46:10.224649 137893368092416 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:46:10.224840 138014997006144 async_checkpointer.py:561] Finished blocking save. Time taken: 0.876135s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/0. I0420 20:46:10.400345 138014997006144 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:46:10.400756 137890176685824 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:46:10.400926 138014997006144 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776717969.3469844, 'wait_for_prev_duration_secs': 6.389617919921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776717969.348672, 'checkpointer_blocking_duration_secs': 0.8762824535369873, 'get_old_steps_start_time': 1776717970.2249744, 'get_old_steps_duration_secs': 2.5510787963867188e-05, 'checkpoint_manager_blocking_start_time': 1776717969.3449674, 'checkpoint_manager_blocking_duration_secs': 1.0559189319610596} I0420 20:46:10.401041 138014997006144 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 20:46:10.401095 138014997006144 max_utils.py:750] Memstats: After params initialized: I0420 20:46:10.401158 138014997006144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0420 20:46:10.401193 138014997006144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0420 20:46:10.401221 138014997006144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0420 20:46:10.401247 138014997006144 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0420 20:46:10.660634 138014997006144 metric_logger.py:185] completed step: 0, seconds: 11.324, TFLOP/s/device: 1.200, Tokens/s/device: 180.858, total_weights: 65536, loss: 10.874 I0420 20:46:10.851071 138014997006144 metric_logger.py:185] completed step: 1, seconds: 1.469, TFLOP/s/device: 9.249, Tokens/s/device: 1394.158, total_weights: 65536, loss: 10.874 I0420 20:46:11.285354 138014997006144 metric_logger.py:185] completed step: 2, seconds: 0.043, TFLOP/s/device: 315.210, Tokens/s/device: 47511.890, total_weights: 65536, loss: 10.560 I0420 20:46:11.432909 138014997006144 metric_logger.py:185] completed step: 3, seconds: 0.435, TFLOP/s/device: 31.248, Tokens/s/device: 4710.071, total_weights: 65536, loss: 9.980 I0420 20:46:11.437328 138014997006144 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0420 20:46:11.727640 138014997006144 checkpointing.py:798] Waited 0.29027700424194336 seconds for step 5 to finish before starting checkpointing. I0420 20:46:11.730525 138014997006144 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:46:13.079152 2574 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 20:46:15.057346 137893376485120 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/0/items/array_metadatas/process_7 I0420 20:46:51.461965 137893368092416 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.803 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.78463888168335 s) (per-host) I0420 20:46:51.462070 137893368092416 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.237292s. I0420 20:46:58.699236 137893368092416 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 48.474442s. I0420 20:46:58.699532 137890176685824 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:46:58.699649 137890176685824 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:46:58.699715 137890176685824 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:46:58.702566 137890176685824 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 20:46:58.702768 138014997006144 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 20:46:58.702911 138014997006144 checkpoint_manager.py:1458] Waiting for previous save to complete took 46.972398 seconds. If this number is high, consider checkpointing less frequently. I0420 20:46:58.704994 138014997006144 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 5 I0420 20:46:58.709083 138014997006144 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/5. I0420 20:46:59.269578 138014997006144 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:46:59.269693 138014997006144 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 I0420 20:46:59.316473 138014997006144 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048082s I0420 20:46:59.316633 138014997006144 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.887 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.31566548347473145 s) (per-host) I0420 20:46:59.316696 138014997006144 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.315737s (batch_requests_ready=0.265766s, total_serialization_initiated=0.049893s, others=0.000078s) I0420 20:46:59.316793 138014997006144 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.320565s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=0.320551) I0420 20:46:59.317737 137890713556736 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:46:59.317897 138014997006144 async_checkpointer.py:561] Finished blocking save. Time taken: 0.612831s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/5. I0420 20:46:59.319725 138014997006144 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:46:59.320002 137890176685824 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:46:59.320134 138014997006144 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776717971.7304873, 'wait_for_prev_duration_secs': 46.97239804267883, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776718018.705037, 'checkpointer_blocking_duration_secs': 0.6129701137542725, 'get_old_steps_start_time': 1776718019.3180234, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776717971.7283, 'checkpoint_manager_blocking_duration_secs': 47.59180235862732} I0420 20:46:59.320300 138014997006144 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0420 20:46:59.320831 138014997006144 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 89.056, Tokens/s/device: 13423.523, total_weights: 65536, loss: 9.460 I0420 20:46:59.326510 138014997006144 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.194, Tokens/s/device: 13896.522, total_weights: 65536, loss: 8.959 I0420 20:47:18.614618 138014997006144 metric_logger.py:185] completed step: 6, seconds: 47.888, TFLOP/s/device: 0.284, Tokens/s/device: 42.766, total_weights: 65536, loss: 8.469 I0420 20:47:18.762027 138014997006144 metric_logger.py:185] completed step: 7, seconds: 19.142, TFLOP/s/device: 0.710, Tokens/s/device: 106.989, total_weights: 65536, loss: 8.003 I0420 20:47:18.909691 138014997006144 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.377, Tokens/s/device: 13471.823, total_weights: 65536, loss: 7.572 I0420 20:47:19.056648 138014997006144 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 20:47:19.057369 138014997006144 checkpointing.py:798] Waited 0.0007421970367431641 seconds for step 9 to finish before starting checkpointing. I0420 20:47:19.059291 138014997006144 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:47:23.142322 137893376485120 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/5/items/array_metadatas/process_7 I0420 20:47:59.751519 137890713556736 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 26.001 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.75051236152649 s) (per-host) I0420 20:47:59.751643 137890713556736 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.433794s. I0420 20:48:09.432895 137890713556736 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 70.115031s. I0420 20:48:09.433174 137890176685824 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:48:09.433295 137890176685824 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:48:09.433342 137890176685824 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:48:09.434994 137890176685824 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0420 20:48:09.435160 138014997006144 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0420 20:48:09.435276 138014997006144 checkpoint_manager.py:1458] Waiting for previous save to complete took 50.375998 seconds. If this number is high, consider checkpointing less frequently. I0420 20:48:09.438346 138014997006144 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0420 20:48:09.441704 138014997006144 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9. I0420 20:48:09.999594 138014997006144 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0420 20:48:09.999702 138014997006144 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 I0420 20:48:10.044079 138014997006144 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045524s I0420 20:48:10.044245 138014997006144 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.062 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3047299385070801 s) (per-host) I0420 20:48:10.044292 138014997006144 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304787s (batch_requests_ready=0.257206s, total_serialization_initiated=0.047519s, others=0.000062s) I0420 20:48:10.044379 138014997006144 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308731s (all_items=0.000012s, per_item={'items': '0.00001192'}, temp_paths=0.308719) I0420 20:48:10.045279 137893368092416 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 20:48:10.045444 138014997006144 async_checkpointer.py:561] Finished blocking save. Time taken: 0.607027s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9. I0420 20:48:10.076509 138014997006144 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 20:48:10.076810 137890176685824 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 20:48:10.076985 138014997006144 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776718039.0592606, 'wait_for_prev_duration_secs': 50.37599849700928, 'time_between_consecutive_saves_sec': 20.35665464401245, 'checkpointer_blocking_start_time': 1776718089.4383864, 'checkpointer_blocking_duration_secs': 0.6071650981903076, 'get_old_steps_start_time': 1776718090.0455642, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1776718039.0575879, 'checkpoint_manager_blocking_duration_secs': 51.01935863494873} I0420 20:48:10.077111 138014997006144 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 20:48:10.077159 138014997006144 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 20:48:15.103349 137893376485120 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/linen_xpk_feat_nnx_set_defaults_true_20260420_190404_16_zarr3_false/checkpoints/9/items/array_metadatas/process_7 I0420 20:48:52.353382 137893368092416 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.068 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.61382699012756 s) (per-host) I0420 20:48:52.353505 137893368092416 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.308112s. I0420 20:48:58.890183 137893368092416 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 48.844774s. I0420 20:48:58.890477 137890176685824 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 20:48:58.890609 137890176685824 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 20:48:58.890676 137890176685824 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 20:48:58.893116 137890176685824 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 20:48:58.893291 138014997006144 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 20:48:58.893445 138014997006144 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 20:48:58.894128 138014997006144 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.184, Tokens/s/device: 13895.014, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 20:49:09 UTC 2026 EXIT_CODE=0