XPK Start: Wed Apr 22 17:36:49 UTC 2026 2026-04-22 17:36:53.941678: 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 00:00:1776879413.954697 9 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776879413.958734 9 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776879413.970333 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879413.970352 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879413.970355 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879413.970357 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 17:37:13.234495: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0422 17:37:13.766803 137177320609600 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 17:37:22,810:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 17:37:22.810433 137177320609600 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 17:37:22,812:jax._src.distributed:157: Connecting to JAX distributed service on mt-03-dropout-exwbm-slice-job-0-0.mt-03-dropout-exwbm:8482 I0422 17:37:22.812714 137177320609600 distributed.py:157] Connecting to JAX distributed service on mt-03-dropout-exwbm-slice-job-0-0.mt-03-dropout-exwbm:8482 I0422 17:37:23.829980 137177320609600 max_utils.py:284] Jax distributed system initialized! I0422 17:37:31.000592 137177320609600 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 17:37:31.000700 137177320609600 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 17:37:31.000741 137177320609600 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 I0422 17:37:31.000774 137177320609600 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 17:37:31.618657 137177320609600 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:37:31.735057 137177320609600 checkpointing.py:688] Setting up checkpoint logger... I0422 17:37:31.735199 137177320609600 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 17:37:31.735249 137177320609600 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 17:37:31.735455 137177320609600 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 0x7cc28e979310>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 17:37:35.114887 137177320609600 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 17:37:35.115146 137177320609600 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cc150cbcf80>}, handler_registry=None I0422 17:37:35.115381 137177320609600 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cc150cbcf80>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0422 17:37:35.115430 137177320609600 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cae242b0e30>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0422 17:37:35.115465 137177320609600 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 0x7cc150cbcf80>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cc150cbcf80>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cae242b0e30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cae242b0e30>}). I0422 17:37:35.115809 137177320609600 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 17:37:35.115886 137177320609600 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7cae24254f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 17:37:35.879705 137177320609600 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints I0422 17:37:35.894803 137177320609600 checkpoint_manager.py:929] [process=6][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/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cae242b1b80> I0422 17:37:35.894929 137177320609600 checkpointing.py:302] Checkpoint manager created! I0422 17:37:36.245651 137177320609600 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 17:37:36.245770 137177320609600 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 17:37:37.669652 137177320609600 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 17:37:37.669774 137177320609600 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 17:37:37.669830 137177320609600 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 17:37:37.669874 137177320609600 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 17:37:37.669912 137177320609600 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 17:37:37.669947 137177320609600 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 17:37:37.669985 137177320609600 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:37:37.670021 137177320609600 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 17:37:37.670057 137177320609600 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:37:37.670092 137177320609600 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:37:37.670142 137177320609600 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 17:37:37.670177 137177320609600 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 17:37:38.075750 137177320609600 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:37:38.075898 137177320609600 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0422 17:37:38.081325 137177320609600 nnx_decoders.py:616] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:37:38.081426 137177320609600 nnx_decoders.py:616] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0422 17:37:38.239236 137177320609600 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:37:38.239371 137177320609600 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0422 17:37:38.398448 137177320609600 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:37:38.398588 137177320609600 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0422 17:37:38.557021 137177320609600 attentions.py:1150] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:37:38.557174 137177320609600 attentions.py:1150] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0422 17:37:38.707703 137177320609600 attentions.py:1151] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:37:38.707837 137177320609600 attentions.py:1151] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0422 17:37:38.858549 137177320609600 attentions.py:1152] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:37:38.858680 137177320609600 attentions.py:1152] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0422 17:37:39.021683 137177320609600 attentions.py:1193] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 17:37:39.021824 137177320609600 attentions.py:1193] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0422 17:37:39.050150 137177320609600 linears.py:541] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 17:37:39.050249 137177320609600 linears.py:541] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None). I0422 17:37:42.605120 137177320609600 max_utils.py:791] Total memory size: 0.8 GB, Output size: 0.4 GB, Temp size: 0.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0422 17:37:42.607426 137177320609600 metric_logger.py:289] number parameters: 1.104 billion I0422 17:37:46.338854 137177320609600 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 17:37:46.420220 137177320609600 checkpointing.py:798] Waited 0.08135485649108887 seconds for step 0 to finish before starting checkpointing. I0422 17:37:46.423972 137177320609600 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:37:46.426008 137177320609600 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0422 17:37:46.428752 137177320609600 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints/0. I0422 17:37:46.953909 137177320609600 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 17:37:46.955089 137177320609600 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 17:37:46.955162 137177320609600 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0422 17:37:47.202925 137177320609600 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250151s I0422 17:37:47.203118 137177320609600 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.090 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4992687702178955 s) (per-host) I0422 17:37:47.203172 137177320609600 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.499344s (batch_requests_ready=0.242427s, total_serialization_initiated=0.256833s, others=0.000083s) I0422 17:37:47.203269 137177320609600 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.503489s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.503470) I0422 17:37:47.204219 137051408893696 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:37:47.204335 137177320609600 async_checkpointer.py:561] Finished blocking save. Time taken: 0.778250s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints/0. I0422 17:37:47.643557 137177320609600 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:37:47.643931 137049319405312 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:37:47.644108 137177320609600 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879466.4239533, 'wait_for_prev_duration_secs': 6.890296936035156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879466.4260495, 'checkpointer_blocking_duration_secs': 0.7783923149108887, 'get_old_steps_start_time': 1776879467.2044616, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1776879466.4206266, 'checkpoint_manager_blocking_duration_secs': 1.2234272956848145} I0422 17:37:47.644218 137177320609600 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 17:37:47.644291 137177320609600 max_utils.py:750] Memstats: After params initialized: I0422 17:37:47.644354 137177320609600 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_24(process=6,(0,6,0,0)) I0422 17:37:47.644390 137177320609600 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_25(process=6,(1,6,0,0)) I0422 17:37:47.644419 137177320609600 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_28(process=6,(0,7,0,0)) I0422 17:37:47.644444 137177320609600 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_29(process=6,(1,7,0,0)) I0422 17:37:47.915077 137177320609600 metric_logger.py:185] completed step: 0, seconds: 3.730, TFLOP/s/device: 0.215, Tokens/s/device: 34.313, total_weights: 4096, loss: 10.874 I0422 17:37:48.004534 137177320609600 metric_logger.py:185] completed step: 1, seconds: 1.574, TFLOP/s/device: 0.509, Tokens/s/device: 81.296, total_weights: 4096, loss: 10.882 I0422 17:37:48.417658 137177320609600 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 38.765, Tokens/s/device: 6195.547, total_weights: 4096, loss: 8.922 I0422 17:37:48.487833 137177320609600 metric_logger.py:185] completed step: 3, seconds: 0.404, TFLOP/s/device: 1.984, Tokens/s/device: 317.157, total_weights: 4096, loss: 7.313 I0422 17:37:48.628879 137177320609600 metric_logger.py:185] completed step: 4, seconds: 0.086, TFLOP/s/device: 9.298, Tokens/s/device: 1486.039, total_weights: 4096, loss: 6.002 I0422 17:37:48.635919 137177320609600 metric_logger.py:185] completed step: 5, seconds: 0.070, TFLOP/s/device: 11.494, Tokens/s/device: 1836.995, total_weights: 4096, loss: 4.934 I0422 17:37:57.679878 137177320609600 metric_logger.py:185] completed step: 6, seconds: 0.142, TFLOP/s/device: 5.657, Tokens/s/device: 904.108, total_weights: 4096, loss: 4.196 I0422 17:37:57.749949 137177320609600 metric_logger.py:185] completed step: 7, seconds: 8.975, TFLOP/s/device: 0.089, Tokens/s/device: 14.262, total_weights: 4096, loss: 3.620 I0422 17:37:57.820276 137177320609600 metric_logger.py:185] completed step: 8, seconds: 0.076, TFLOP/s/device: 10.501, Tokens/s/device: 1678.380, total_weights: 4096, loss: 3.238 I0422 17:37:57.891403 137177320609600 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 17:37:57.891950 137177320609600 checkpointing.py:798] Waited 0.0005557537078857422 seconds for step 9 to finish before starting checkpointing. I0422 17:37:57.893935 137177320609600 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:37:58.876413 2633 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 17:38:00.819604 137049814312704 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints/0/items/array_metadatas/process_6 I0422 17:38:28.617815 137051408893696 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.687 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.913939476013184 s) (per-host) I0422 17:38:28.617933 137051408893696 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.413651s. I0422 17:38:34.693419 137051408893696 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.489122s. I0422 17:38:34.693720 137049319405312 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:38:34.693838 137049319405312 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:38:34.693889 137049319405312 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:38:34.695692 137049319405312 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 17:38:34.695800 137177320609600 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0422 17:38:34.695876 137177320609600 checkpoint_manager.py:1458] Waiting for previous save to complete took 36.801950 seconds. If this number is high, consider checkpointing less frequently. I0422 17:38:34.697880 137177320609600 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0422 17:38:34.701336 137177320609600 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints/9. I0422 17:38:35.254439 137177320609600 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 17:38:35.254605 137177320609600 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0422 17:38:35.296302 137177320609600 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043861s I0422 17:38:35.296474 137177320609600 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.034 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30643391609191895 s) (per-host) I0422 17:38:35.296531 137177320609600 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306500s (batch_requests_ready=0.257841s, total_serialization_initiated=0.048585s, others=0.000074s) I0422 17:38:35.296628 137177320609600 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.310644s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.310628) I0422 17:38:35.297550 137050900838144 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0422 17:38:35.297716 137177320609600 async_checkpointer.py:561] Finished blocking save. Time taken: 0.599766s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints/9. I0422 17:38:35.716508 137177320609600 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:38:35.716889 137049319405312 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:38:35.717060 137177320609600 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879477.8939059, 'wait_for_prev_duration_secs': 36.8019495010376, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879514.697918, 'checkpointer_blocking_duration_secs': 0.5999042987823486, 'get_old_steps_start_time': 1776879515.297841, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776879477.8921845, 'checkpoint_manager_blocking_duration_secs': 37.82484030723572} I0422 17:38:35.717196 137177320609600 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 17:38:35.717242 137177320609600 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:38:40.447843 137044441732864 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_03_dropout/checkpoints/9/items/array_metadatas/process_6 I0422 17:39:16.342925 137050900838144 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.198 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.35284113883972 s) (per-host) I0422 17:39:16.343047 137050900838144 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.045381s. I0422 17:39:23.178513 137050900838144 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 47.880831s. I0422 17:39:23.178820 137049319405312 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:39:23.178941 137049319405312 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:39:23.178989 137049319405312 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:39:23.180637 137049319405312 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 17:39:23.180816 137177320609600 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0422 17:39:23.180988 137177320609600 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:39:23.181658 137177320609600 metric_logger.py:185] completed step: 9, seconds: 0.069, TFLOP/s/device: 11.563, Tokens/s/device: 1848.029, total_weights: 4096, loss: 3.085 Per train step: Total TFLOPs: 0.80 split as 99.60% learnable weight flops and 0.40% attention flops XPK End: Wed Apr 22 17:39:35 UTC 2026 EXIT_CODE=0