XPK Start: Thu Apr 23 16:25:38 UTC 2026 2026-04-23 16:25:42.834496: 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 00:00:1776961542.847130 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776961542.850815 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776961542.862090 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776961542.862107 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776961542.862110 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776961542.862112 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 16:26:01.910453: 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) I0423 16:26:02.440970 133146243807040 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 16:26:11,482:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 16:26:11.482547 133146243807040 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 16:26:11,484:jax._src.distributed:157: Connecting to JAX distributed service on mt-09-pdb-lt-1-l7gaj-slice-job-0-0.mt-09-pdb-lt-1-l7gaj:8482 I0423 16:26:11.484853 133146243807040 distributed.py:157] Connecting to JAX distributed service on mt-09-pdb-lt-1-l7gaj-slice-job-0-0.mt-09-pdb-lt-1-l7gaj:8482 I0423 16:26:12.657407 133146243807040 max_utils.py:284] Jax distributed system initialized! I0423 16:26:19.658703 133146243807040 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 16:26:19.658808 133146243807040 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 16:26:19.658848 133146243807040 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 I0423 16:26:19.658882 133146243807040 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 16:26:20.281440 133146243807040 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 8, 1, 1, 1, 1, 4, 1, 1, 1, 1) I0423 16:26:20.282056 133146243807040 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 8, 1, 1, 1, 1, 4, 1, 1, 1, 1) I0423 16:26:20.282242 133146243807040 checkpointing.py:688] Setting up checkpoint logger... I0423 16:26:20.282292 133146243807040 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 16:26:20.282334 133146243807040 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 16:26:20.282692 133146243807040 base_pytree_checkpoint_handler.py:415] 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 0x79180a37d250>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 16:26:23.250629 133146243807040 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 16:26:23.251011 133146243807040 checkpoint_manager.py:709] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79049aaf9790>}, handler_registry=None I0423 16:26:23.251247 133146243807040 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79049aaf9790>` 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`. I0423 16:26:23.251296 133146243807040 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79049aafe2a0>` 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`. I0423 16:26:23.251331 133146243807040 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 0x79049aaf9790>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79049aaf9790>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79049aafe2a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79049aafe2a0>}). I0423 16:26:23.251848 133146243807040 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 16:26:23.251927 133146243807040 async_checkpointer.py:177] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x79049a160180> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 16:26:23.955061 133146243807040 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints I0423 16:26:23.986897 133146243807040 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=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_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79049aaf8170> I0423 16:26:23.987030 133146243807040 checkpointing.py:302] Checkpoint manager created! I0423 16:26:24.882858 133146243807040 nnx_wrappers.py:455] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 16:26:24.882962 133146243807040 nnx_wrappers.py:455] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0423 16:26:25.250970 133146243807040 attentions.py:1084] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 16:26:25.251064 133146243807040 attentions.py:1084] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0423 16:26:25.266860 133146243807040 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 16:26:25.266920 133146243807040 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, 'tensor'). I0423 16:26:25.289523 133146243807040 attentions.py:1150] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:26:25.289592 133146243807040 attentions.py:1150] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0423 16:26:25.305398 133146243807040 attentions.py:1151] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:26:25.305463 133146243807040 attentions.py:1151] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0423 16:26:25.321183 133146243807040 attentions.py:1152] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 16:26:25.321245 133146243807040 attentions.py:1152] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0423 16:26:25.345940 133146243807040 attentions.py:1193] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 16:26:25.346014 133146243807040 attentions.py:1193] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, 'tensor', None). I0423 16:26:25.365464 133146243807040 linears.py:541] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 16:26:25.365531 133146243807040 linears.py:541] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, 'tensor'). I0423 16:26:25.553694 133146243807040 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 16:26:25.553805 133146243807040 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 8 tensor: 4 I0423 16:26:27.008045 133146243807040 maxtext_utils.py:1740] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: ('tensor',) I0423 16:26:27.008178 133146243807040 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, 'tensor') I0423 16:26:27.008239 133146243807040 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, 'tensor') I0423 16:26:27.008299 133146243807040 maxtext_utils.py:1740] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: ('tensor', None, 'fsdp') I0423 16:26:27.008351 133146243807040 maxtext_utils.py:1740] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: ('tensor', None) I0423 16:26:27.008391 133146243807040 maxtext_utils.py:1740] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: ('tensor', None) I0423 16:26:27.008446 133146243807040 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, 'tensor', None) I0423 16:26:27.008497 133146243807040 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: ('tensor', None, None, 'fsdp') I0423 16:26:27.008538 133146243807040 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, 'tensor', None) I0423 16:26:27.008577 133146243807040 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, 'tensor', None) I0423 16:26:27.008628 133146243807040 maxtext_utils.py:1740] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed', 'vocab') Physical: ('fsdp', 'tensor') I0423 16:26:27.008714 133146243807040 maxtext_utils.py:1740] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed') Physical: ('tensor', 'fsdp') I0423 16:26:27.511577 133146243807040 train.py:158] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 16:26:27.511682 133146243807040 train.py:158] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0423 16:26:27.526580 133146243807040 train.py:165] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 16:26:27.526639 133146243807040 train.py:165] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0423 16:26:37.746471 133146243807040 max_utils.py:791] Total memory size: 0.9 GB, Output size: 0.4 GB, Temp size: 0.5 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0423 16:26:37.747359 133146243807040 metric_logger.py:289] number parameters: 1.104 billion I0423 16:26:39.309906 133146243807040 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 16:26:47.856293 133146243807040 checkpointing.py:798] Waited 8.546356201171875 seconds for step 0 to finish before starting checkpointing. I0423 16:26:47.858820 133146243807040 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 16:26:47.860481 133146243807040 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 0 I0423 16:26:47.863255 133146243807040 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints/0. I0423 16:26:48.415292 133146243807040 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 16:26:48.416290 133146243807040 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:26:48.416348 133146243807040 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 I0423 16:26:48.644995 133146243807040 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.229785s I0423 16:26:48.645179 133146243807040 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.133 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4924154281616211 s) (per-host) I0423 16:26:48.645233 133146243807040 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.492479s (batch_requests_ready=0.259107s, total_serialization_initiated=0.233300s, others=0.000071s) I0423 16:26:48.645324 133146243807040 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.496429s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.496409) I0423 16:26:48.646285 133023215363840 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0423 16:26:48.646468 133146243807040 async_checkpointer.py:561] Finished blocking save. Time taken: 0.785915s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints/0. I0423 16:26:48.669754 133146243807040 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:26:48.670078 133020531009280 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:26:48.670248 133146243807040 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776961607.8588018, 'wait_for_prev_duration_secs': 6.508827209472656e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776961607.860522, 'checkpointer_blocking_duration_secs': 0.7860507965087891, 'get_old_steps_start_time': 1776961608.646593, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776961607.8567932, 'checkpoint_manager_blocking_duration_secs': 0.813410758972168} I0423 16:26:48.670350 133146243807040 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 16:26:48.670397 133146243807040 max_utils.py:750] Memstats: After params initialized: I0423 16:26:48.670445 133146243807040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0423 16:26:48.670476 133146243807040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0423 16:26:48.670501 133146243807040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0423 16:26:48.670524 133146243807040 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0423 16:26:48.939248 133146243807040 metric_logger.py:185] completed step: 0, seconds: 1.562, TFLOP/s/device: 2.174, Tokens/s/device: 327.693, total_weights: 16384, loss: 10.862 I0423 16:26:49.023536 133146243807040 metric_logger.py:185] completed step: 1, seconds: 9.628, TFLOP/s/device: 0.353, Tokens/s/device: 53.178, total_weights: 16384, loss: 10.862 I0423 16:26:49.448381 133146243807040 metric_logger.py:185] completed step: 2, seconds: 0.026, TFLOP/s/device: 131.109, Tokens/s/device: 19762.236, total_weights: 16384, loss: 9.763 I0423 16:26:49.507020 133146243807040 metric_logger.py:185] completed step: 3, seconds: 0.425, TFLOP/s/device: 7.985, Tokens/s/device: 1203.607, total_weights: 16384, loss: 8.823 I0423 16:26:49.624641 133146243807040 metric_logger.py:185] completed step: 4, seconds: 0.064, TFLOP/s/device: 53.492, Tokens/s/device: 8062.865, total_weights: 16384, loss: 7.947 I0423 16:26:49.630516 133146243807040 metric_logger.py:185] completed step: 5, seconds: 0.059, TFLOP/s/device: 57.906, Tokens/s/device: 8728.265, total_weights: 16384, loss: 7.212 I0423 16:27:02.667576 133146243807040 metric_logger.py:185] completed step: 6, seconds: 0.118, TFLOP/s/device: 28.719, Tokens/s/device: 4328.785, total_weights: 16384, loss: 6.652 I0423 16:27:02.726324 133146243807040 metric_logger.py:185] completed step: 7, seconds: 12.979, TFLOP/s/device: 0.262, Tokens/s/device: 39.448, total_weights: 16384, loss: 6.269 I0423 16:27:02.784907 133146243807040 metric_logger.py:185] completed step: 8, seconds: 0.063, TFLOP/s/device: 53.889, Tokens/s/device: 8122.729, total_weights: 16384, loss: 6.031 I0423 16:27:02.843016 133146243807040 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 16:27:02.843511 133146243807040 checkpointing.py:798] Waited 0.0005114078521728516 seconds for step 9 to finish before starting checkpointing. I0423 16:27:02.845735 133146243807040 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:27:04.586518 2450 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 16:27:06.535945 133023223756544 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints/0/items/array_metadatas/process_5 I0423 16:27:37.765698 133023215363840 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 31.838 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 49.6128830909729 s) (per-host) I0423 16:27:37.765818 133023215363840 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 49.119410s. I0423 16:27:46.878046 133023215363840 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 58.231622s. I0423 16:27:46.878340 133020531009280 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:27:46.878462 133020531009280 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:27:46.878515 133020531009280 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:27:46.880875 133020531009280 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 16:27:46.881067 133146243807040 checkpoint_manager.py:2036] [process=5][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0423 16:27:46.881195 133146243807040 checkpoint_manager.py:1458] Waiting for previous save to complete took 44.035466 seconds. If this number is high, consider checkpointing less frequently. I0423 16:27:46.883204 133146243807040 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 9 I0423 16:27:46.886772 133146243807040 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints/9. I0423 16:27:47.470207 133146243807040 jax_array_handlers.py:358] Scheduling D2H of 39 prioritized jax.Array. I0423 16:27:47.470310 133146243807040 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 I0423 16:27:47.512999 133146243807040 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043981s I0423 16:27:47.513164 133146243807040 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.063 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3046834468841553 s) (per-host) I0423 16:27:47.513220 133146243807040 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.304748s (batch_requests_ready=0.258868s, total_serialization_initiated=0.045807s, others=0.000072s) I0423 16:27:47.513313 133146243807040 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308828s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.308812) I0423 16:27:47.514152 133012486338304 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0423 16:27:47.514262 133146243807040 async_checkpointer.py:561] Finished blocking save. Time taken: 0.630985s. Continuing background save to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints/9. I0423 16:27:47.519645 133146243807040 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 16:27:47.519898 133020531009280 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0423 16:27:47.520021 133146243807040 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776961622.845705, 'wait_for_prev_duration_secs': 44.03546619415283, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776961666.883244, 'checkpointer_blocking_duration_secs': 0.63112473487854, 'get_old_steps_start_time': 1776961667.5143883, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776961622.8437502, 'checkpoint_manager_blocking_duration_secs': 44.67623567581177} I0423 16:27:47.520124 133146243807040 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 16:27:47.520171 133146243807040 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 16:27:52.886531 132993305736960 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/linen_xpk_feat_nnx_set_defaults_true_20260423_155242_09_pdb_lt_1/checkpoints/9/items/array_metadatas/process_5 I0423 16:28:28.461837 133012486338304 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 38.290 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.253313302993774 s) (per-host) I0423 16:28:28.461962 133012486338304 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.947746s. I0423 16:28:37.689606 133012486338304 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 50.175374s. I0423 16:28:37.689863 133020531009280 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 16:28:37.689919 133020531009280 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0423 16:28:37.689965 133020531009280 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 16:28:37.692195 133020531009280 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 16:28:37.692356 133146243807040 checkpoint_manager.py:2036] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0423 16:28:37.692499 133146243807040 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 16:28:37.693210 133146243807040 metric_logger.py:185] completed step: 9, seconds: 0.059, TFLOP/s/device: 57.615, Tokens/s/device: 8684.295, total_weights: 16384, loss: 5.892 Per train step: Total TFLOPs: 3.40 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 16:28:49 UTC 2026 EXIT_CODE=0