XPK Start: Wed Apr 22 17:33:08 UTC 2026 2026-04-22 17:33:12.373734: 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 00:00:1776879192.386598 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776879192.390295 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776879192.401569 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879192.401586 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879192.401589 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776879192.401591 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-22 17:33:31.484330: 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 17:33:32.004682 136450511308608 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-22 17:33:41,044:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0422 17:33:41.044955 136450511308608 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-22 17:33:41,047:jax._src.distributed:157: Connecting to JAX distributed service on mt-02-synthetic-meyp6-slice-job-0-0.mt-02-synthetic-meyp6:8482 I0422 17:33:41.047502 136450511308608 distributed.py:157] Connecting to JAX distributed service on mt-02-synthetic-meyp6-slice-job-0-0.mt-02-synthetic-meyp6:8482 I0422 17:33:42.365638 136450511308608 max_utils.py:284] Jax distributed system initialized! I0422 17:33:49.418242 136450511308608 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 17:33:49.418352 136450511308608 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 17:33:49.418394 136450511308608 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 17:33:49.418428 136450511308608 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 17:33:50.032951 136450511308608 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 17:33:50.147779 136450511308608 checkpointing.py:688] Setting up checkpoint logger... I0422 17:33:50.147902 136450511308608 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 17:33:50.147947 136450511308608 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0422 17:33:50.148158 136450511308608 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 0x7c195a7753d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 17:33:53.494420 136450511308608 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0422 17:33:53.494786 136450511308608 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 0x7c04e834cfe0>}, handler_registry=None I0422 17:33:53.495033 136450511308608 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c04e834cfe0>` 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 17:33:53.495084 136450511308608 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c04e843f3b0>` 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 17:33:53.495120 136450511308608 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 0x7c04e834cfe0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c04e834cfe0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c04e843f3b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c04e843f3b0>}). I0422 17:33:53.495452 136450511308608 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0422 17:33:53.495528 136450511308608 async_checkpointer.py:177] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c04a8778f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 17:33:54.182811 136450511308608 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_02_synthetic/checkpoints I0422 17:33:54.249984 136450511308608 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/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c04e8439430> I0422 17:33:54.250147 136450511308608 checkpointing.py:302] Checkpoint manager created! I0422 17:33:55.183673 136450511308608 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0422 17:33:55.183797 136450511308608 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0422 17:33:56.601475 136450511308608 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0422 17:33:56.601598 136450511308608 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 17:33:56.601667 136450511308608 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0422 17:33:56.601716 136450511308608 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0422 17:33:56.601756 136450511308608 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 17:33:56.601793 136450511308608 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0422 17:33:56.601831 136450511308608 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:33:56.601867 136450511308608 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0422 17:33:56.601903 136450511308608 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:33:56.601938 136450511308608 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0422 17:33:56.601976 136450511308608 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0422 17:33:56.602010 136450511308608 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0422 17:33:57.004442 136450511308608 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:33:57.004586 136450511308608 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:33:57.009965 136450511308608 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 17:33:57.010057 136450511308608 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:33:57.157969 136450511308608 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:33:57.158100 136450511308608 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:33:57.305391 136450511308608 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 17:33:57.305521 136450511308608 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0422 17:33:57.462561 136450511308608 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:33:57.462729 136450511308608 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:33:57.609073 136450511308608 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:33:57.609280 136450511308608 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:33:57.754508 136450511308608 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 17:33:57.754664 136450511308608 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:33:57.908530 136450511308608 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 17:33:57.908690 136450511308608 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0422 17:33:57.931110 136450511308608 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 17:33:57.931208 136450511308608 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0422 17:34:08.972315 136450511308608 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. I0422 17:34:08.974749 136450511308608 metric_logger.py:289] number parameters: 1.104 billion I0422 17:34:19.834862 136450511308608 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0422 17:34:19.985149 136450511308608 checkpointing.py:798] Waited 0.1502704620361328 seconds for step 0 to finish before starting checkpointing. I0422 17:34:19.987750 136450511308608 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:34:19.989516 136450511308608 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 0 I0422 17:34:19.992246 136450511308608 async_checkpointer.py:452] [process=5] 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_02_synthetic/checkpoints/0. I0422 17:34:20.569702 136450511308608 signaling_client.py:364] Using JaxDistributedSignalingClient I0422 17:34:20.570941 136450511308608 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 17:34:20.571005 136450511308608 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 17:34:20.823679 136450511308608 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.255114s I0422 17:34:20.823858 136450511308608 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.008 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5128984451293945 s) (per-host) I0422 17:34:20.823910 136450511308608 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.512962s (batch_requests_ready=0.251241s, total_serialization_initiated=0.261649s, others=0.000071s) I0422 17:34:20.824016 136450511308608 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.516949s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=0.516930) I0422 17:34:20.825024 136324627535616 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0422 17:34:20.825196 136450511308608 async_checkpointer.py:561] Finished blocking save. Time taken: 0.835611s. 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_02_synthetic/checkpoints/0. I0422 17:34:21.215998 136450511308608 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:34:21.216383 136323042100992 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:34:21.216550 136450511308608 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_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879259.9877317, 'wait_for_prev_duration_secs': 6.270408630371094e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879259.989553, 'checkpointer_blocking_duration_secs': 0.8357574939727783, 'get_old_steps_start_time': 1776879260.8253303, 'get_old_steps_duration_secs': 2.4557113647460938e-05, 'checkpoint_manager_blocking_start_time': 1776879259.9855509, 'checkpoint_manager_blocking_duration_secs': 1.230957269668579} I0422 17:34:21.216678 136450511308608 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0422 17:34:21.216754 136450511308608 max_utils.py:750] Memstats: After params initialized: I0422 17:34:21.216811 136450511308608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0422 17:34:21.216848 136450511308608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0422 17:34:21.216877 136450511308608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0422 17:34:21.216903 136450511308608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0422 17:34:21.483829 136450511308608 metric_logger.py:185] completed step: 0, seconds: 10.859, TFLOP/s/device: 1.251, Tokens/s/device: 188.598, total_weights: 65536, loss: 10.874 I0422 17:34:21.661364 136450511308608 metric_logger.py:185] completed step: 1, seconds: 1.647, TFLOP/s/device: 8.248, Tokens/s/device: 1243.267, total_weights: 65536, loss: 10.874 I0422 17:34:22.087174 136450511308608 metric_logger.py:185] completed step: 2, seconds: 0.032, TFLOP/s/device: 430.054, Tokens/s/device: 64822.435, total_weights: 65536, loss: 10.262 I0422 17:34:22.234474 136450511308608 metric_logger.py:185] completed step: 3, seconds: 0.416, TFLOP/s/device: 32.622, Tokens/s/device: 4917.202, total_weights: 65536, loss: 9.732 I0422 17:34:22.530143 136450511308608 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.276, Tokens/s/device: 12552.250, total_weights: 65536, loss: 9.274 I0422 17:34:22.537420 136450511308608 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.534, Tokens/s/device: 13947.723, total_weights: 65536, loss: 8.888 I0422 17:34:41.622120 136450511308608 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.864, Tokens/s/device: 6913.150, total_weights: 65536, loss: 8.589 I0422 17:34:41.769708 136450511308608 metric_logger.py:185] completed step: 7, seconds: 18.939, TFLOP/s/device: 0.717, Tokens/s/device: 108.138, total_weights: 65536, loss: 8.380 I0422 17:34:41.917185 136450511308608 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.931, Tokens/s/device: 13404.633, total_weights: 65536, loss: 8.251 I0422 17:34:42.065566 136450511308608 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0422 17:34:42.066128 136450511308608 checkpointing.py:798] Waited 0.0005750656127929688 seconds for step 9 to finish before starting checkpointing. I0422 17:34:42.068193 136450511308608 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:34:42.512925 2604 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0422 17:34:44.375576 136324635928320 array_metadata_store.py:203] [process=5][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_02_synthetic/checkpoints/0/items/array_metadatas/process_5 I0422 17:35:16.458792 136324627535616 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 28.133 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 56.14779281616211 s) (per-host) I0422 17:35:16.458912 136324627535616 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.633777s. I0422 17:35:24.081903 136324627535616 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 63.256751s. I0422 17:35:24.082216 136323042100992 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:35:24.082341 136323042100992 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:35:24.082391 136323042100992 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:35:24.085705 136323042100992 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 17:35:24.085827 136450511308608 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. W0422 17:35:24.085901 136450511308608 checkpoint_manager.py:1458] Waiting for previous save to complete took 42.017716 seconds. If this number is high, consider checkpointing less frequently. I0422 17:35:24.088025 136450511308608 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 9 I0422 17:35:24.091623 136450511308608 async_checkpointer.py:452] [process=5] 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_02_synthetic/checkpoints/9. I0422 17:35:25.084240 136450511308608 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0422 17:35:25.084404 136450511308608 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 17:35:25.130164 136450511308608 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.047924s I0422 17:35:25.130339 136450511308608 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.155 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.29926633834838867 s) (per-host) I0422 17:35:25.130393 136450511308608 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.299330s (batch_requests_ready=0.246572s, total_serialization_initiated=0.052686s, others=0.000072s) I0422 17:35:25.130491 136450511308608 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.303532s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.303516) I0422 17:35:25.131379 136324115842816 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0422 17:35:25.131508 136450511308608 async_checkpointer.py:561] Finished blocking save. Time taken: 1.043412s. 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_02_synthetic/checkpoints/9. I0422 17:35:25.172552 136450511308608 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 17:35:25.172887 136323042100992 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0422 17:35:25.173063 136450511308608 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_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879282.068164, 'wait_for_prev_duration_secs': 42.0177161693573, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879324.0880635, 'checkpointer_blocking_duration_secs': 1.0436279773712158, 'get_old_steps_start_time': 1776879325.1317117, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776879282.0663526, 'checkpoint_manager_blocking_duration_secs': 43.106669425964355} I0422 17:35:25.173179 136450511308608 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0422 17:35:25.173223 136450511308608 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 17:35:30.610853 136324635928320 array_metadata_store.py:203] [process=5][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_02_synthetic/checkpoints/9/items/array_metadatas/process_5 I0422 17:36:08.049776 136324115842816 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 36.549 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 43.21866464614868 s) (per-host) I0422 17:36:08.049894 136324115842816 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.918306s. I0422 17:36:18.683326 136324115842816 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 53.551723s. I0422 17:36:18.683609 136323042100992 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 17:36:18.683744 136323042100992 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0422 17:36:18.683793 136323042100992 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 17:36:18.685342 136323042100992 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 17:36:18.685533 136450511308608 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. I0422 17:36:18.685719 136450511308608 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 17:36:18.686384 136450511308608 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.306, Tokens/s/device: 13913.422, total_weights: 65536, loss: 8.175 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Wed Apr 22 17:36:30 UTC 2026 EXIT_CODE=0