XPK Start: Thu Apr 23 18:06:41 UTC 2026 2026-04-23 18:06:45.086506: 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 00:00:1776967605.098956 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776967605.102619 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776967605.114045 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776967605.114067 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776967605.114069 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776967605.114071 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 18:07:04.289558: 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 18:07:04.820545 139817221502784 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 18:07:13,861:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 18:07:13.861887 139817221502784 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 18:07:13,864:jax._src.distributed:157: Connecting to JAX distributed service on mt-11-optimizer-offload-cfbcr-slice-job-0-0.mt-11-optimizer-offload-cfbcr:8482 I0423 18:07:13.864191 139817221502784 distributed.py:157] Connecting to JAX distributed service on mt-11-optimizer-offload-cfbcr-slice-job-0-0.mt-11-optimizer-offload-cfbcr:8482 I0423 18:07:14.911472 139817221502784 max_utils.py:284] Jax distributed system initialized! I0423 18:07:21.579877 139817221502784 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 18:07:21.579983 139817221502784 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 18:07:21.580023 139817221502784 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 18:07:21.580062 139817221502784 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 18:07:22.197790 139817221502784 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 18:07:22.313898 139817221502784 checkpointing.py:688] Setting up checkpoint logger... I0423 18:07:22.314016 139817221502784 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 18:07:22.314068 139817221502784 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 18:07:22.314274 139817221502784 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 0x7f2939f54bf0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 18:07:25.192251 139817221502784 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 18:07:25.192592 139817221502784 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 0x7f163e0e43e0>}, handler_registry=None I0423 18:07:25.192838 139817221502784 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f163e0e43e0>` 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 18:07:25.192887 139817221502784 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f163fc37da0>` 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 18:07:25.192924 139817221502784 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 0x7f163e0e43e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f163e0e43e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f163fc37da0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f163fc37da0>}). I0423 18:07:25.193258 139817221502784 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 18:07:25.193334 139817221502784 async_checkpointer.py:177] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f163fc0cf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 18:07:26.793422 139817221502784 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints I0423 18:07:26.795598 139817221502784 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f16e825ce90> I0423 18:07:26.795722 139817221502784 checkpointing.py:302] Checkpoint manager created! I0423 18:07:27.725913 139817221502784 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 18:07:27.726035 139817221502784 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 18:07:29.146555 139817221502784 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0423 18:07:29.146687 139817221502784 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:07:29.146744 139817221502784 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:07:29.146789 139817221502784 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0423 18:07:29.146827 139817221502784 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:07:29.146863 139817221502784 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:07:29.146901 139817221502784 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:07:29.146938 139817221502784 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0423 18:07:29.146974 139817221502784 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:07:29.147014 139817221502784 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:07:29.147048 139817221502784 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0423 18:07:29.147082 139817221502784 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0423 18:07:29.554426 139817221502784 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:07:29.554579 139817221502784 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:07:29.560039 139817221502784 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:07:29.560137 139817221502784 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:07:29.718268 139817221502784 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:07:29.718402 139817221502784 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:07:29.875219 139817221502784 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:07:29.875352 139817221502784 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:07:30.045443 139817221502784 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:07:30.045595 139817221502784 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:07:30.204602 139817221502784 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:07:30.204752 139817221502784 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:07:30.360862 139817221502784 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 18:07:30.360995 139817221502784 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:07:30.527040 139817221502784 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 18:07:30.527180 139817221502784 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:07:30.550397 139817221502784 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 18:07:30.550498 139817221502784 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 18:07:41.543784 139817221502784 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. I0423 18:07:41.546180 139817221502784 metric_logger.py:289] number parameters: 1.104 billion I0423 18:07:52.784317 139817221502784 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 18:07:52.935363 139817221502784 checkpointing.py:798] Waited 0.15103411674499512 seconds for step 0 to finish before starting checkpointing. I0423 18:07:52.937633 139817221502784 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:07:52.939230 139817221502784 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 0 I0423 18:07:52.942013 139817221502784 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints/0. I0423 18:07:53.909520 139817221502784 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 18:07:53.910534 139817221502784 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:07:53.910592 139817221502784 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 18:07:54.156720 139817221502784 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.248328s I0423 18:07:54.156889 139817221502784 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.025 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5099959373474121 s) (per-host) I0423 18:07:54.156943 139817221502784 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.510059s (batch_requests_ready=0.255186s, total_serialization_initiated=0.254801s, others=0.000072s) I0423 18:07:54.157043 139817221502784 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.514152s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.514132) I0423 18:07:54.158114 139692435552000 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0423 18:07:54.158292 139817221502784 async_checkpointer.py:561] Finished blocking save. Time taken: 1.218989s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints/0. I0423 18:07:54.216184 139817221502784 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:07:54.216545 139696747304704 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:07:54.216739 139817221502784 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776967672.9376159, 'wait_for_prev_duration_secs': 7.915496826171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776967672.939267, 'checkpointer_blocking_duration_secs': 1.219149112701416, 'get_old_steps_start_time': 1776967674.1584406, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776967672.935781, 'checkpoint_manager_blocking_duration_secs': 1.2809176445007324} I0423 18:07:54.216867 139817221502784 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 18:07:54.216983 139817221502784 max_utils.py:750] Memstats: After params initialized: I0423 18:07:54.217057 139817221502784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_18(process=5,(2,4,0,0)) I0423 18:07:54.217103 139817221502784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_19(process=5,(3,4,0,0)) I0423 18:07:54.217135 139817221502784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_22(process=5,(2,5,0,0)) I0423 18:07:54.217174 139817221502784 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_23(process=5,(3,5,0,0)) I0423 18:07:54.486234 139817221502784 metric_logger.py:185] completed step: 0, seconds: 11.237, TFLOP/s/device: 1.209, Tokens/s/device: 182.254, total_weights: 65536, loss: 10.874 I0423 18:07:54.653367 139817221502784 metric_logger.py:185] completed step: 1, seconds: 1.700, TFLOP/s/device: 7.992, Tokens/s/device: 1204.569, total_weights: 65536, loss: 10.874 I0423 18:07:55.077328 139817221502784 metric_logger.py:185] completed step: 2, seconds: 0.022, TFLOP/s/device: 627.320, Tokens/s/device: 94556.535, total_weights: 65536, loss: 10.262 I0423 18:07:55.224643 139817221502784 metric_logger.py:185] completed step: 3, seconds: 0.415, TFLOP/s/device: 32.737, Tokens/s/device: 4934.476, total_weights: 65536, loss: 9.732 I0423 18:07:55.520138 139817221502784 metric_logger.py:185] completed step: 4, seconds: 0.162, TFLOP/s/device: 83.666, Tokens/s/device: 12611.070, total_weights: 65536, loss: 9.274 I0423 18:07:55.527252 139817221502784 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.162, Tokens/s/device: 13891.715, total_weights: 65536, loss: 8.888 I0423 18:08:14.584966 139817221502784 metric_logger.py:185] completed step: 6, seconds: 0.295, TFLOP/s/device: 46.009, Tokens/s/device: 6934.968, total_weights: 65536, loss: 8.589 I0423 18:08:14.732243 139817221502784 metric_logger.py:185] completed step: 7, seconds: 18.911, TFLOP/s/device: 0.718, Tokens/s/device: 108.294, total_weights: 65536, loss: 8.380 I0423 18:08:14.880001 139817221502784 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.516, Tokens/s/device: 13342.106, total_weights: 65536, loss: 8.251 I0423 18:08:15.028393 139817221502784 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 18:08:15.029023 139817221502784 checkpointing.py:798] Waited 0.0006403923034667969 seconds for step 9 to finish before starting checkpointing. I0423 18:08:15.030876 139817221502784 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:08:15.727988 2620 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 18:08:17.656737 139696755697408 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints/0/items/array_metadatas/process_5 I0423 18:08:50.149693 139692435552000 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 27.956 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 56.50274872779846 s) (per-host) I0423 18:08:50.149808 139692435552000 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.991579s. I0423 18:08:55.885118 139692435552000 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 61.726873s. I0423 18:08:55.885411 139696747304704 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:08:55.885529 139696747304704 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:08:55.885579 139696747304704 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:08:55.887152 139696747304704 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 18:08:55.887336 139817221502784 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 18:08:55.887472 139817221502784 checkpoint_manager.py:1458] Waiting for previous save to complete took 40.856601 seconds. If this number is high, consider checkpointing less frequently. I0423 18:08:55.890621 139817221502784 checkpoint_manager.py:1518] [process=5] Saving checkpoint at step 9 I0423 18:08:55.894167 139817221502784 async_checkpointer.py:452] [process=5] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints/9. I0423 18:08:57.150154 139817221502784 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:08:57.150311 139817221502784 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 18:08:57.193697 139817221502784 base_pytree_checkpoint_handler.py:153] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.045703s I0423 18:08:57.193869 139817221502784 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.596 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5941629409790039 s) (per-host) I0423 18:08:57.193924 139817221502784 base_pytree_checkpoint_handler.py:737] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.594228s (batch_requests_ready=0.543815s, total_serialization_initiated=0.050339s, others=0.000074s) I0423 18:08:57.194022 139817221502784 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.598390s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.598373) I0423 18:08:57.195010 139692980815616 async_checkpointer.py:79] [process=5][thread=async_save] Background save thread started. I0423 18:08:57.195141 139817221502784 async_checkpointer.py:561] Finished blocking save. Time taken: 1.304436s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints/9. I0423 18:08:57.254722 139817221502784 checkpoint_manager.py:1566] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:08:57.255055 139696747304704 async_checkpointer.py:265] [process=5][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:08:57.255229 139817221502784 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776967695.0308468, 'wait_for_prev_duration_secs': 40.856600761413574, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776967735.8906724, 'checkpointer_blocking_duration_secs': 1.3046493530273438, 'get_old_steps_start_time': 1776967737.1953413, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776967695.0292437, 'checkpoint_manager_blocking_duration_secs': 42.22594904899597} I0423 18:08:57.255350 139817221502784 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 18:08:57.255395 139817221502784 checkpoint_manager.py:2024] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:09:01.665734 139696755697408 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_11_optimizer_offload_false/checkpoints/9/items/array_metadatas/process_5 I0423 18:09:38.853884 139692980815616 base_pytree_checkpoint_handler.py:129] [process=5] /jax/checkpoint/write/gbytes_per_sec: 37.383 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.25413227081299 s) (per-host) I0423 18:09:38.854034 139692980815616 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.658812s. I0423 18:09:44.731091 139692980815616 async_checkpointer.py:144] [process=5][thread=async_save] Background save thread done. Time taken: 47.535854s. I0423 18:09:44.731386 139696747304704 async_checkpointer.py:273] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:09:44.731503 139696747304704 async_checkpointer.py:283] [process=5][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:09:44.731551 139696747304704 checkpoint_manager.py:2133] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:09:44.733741 139696747304704 checkpoint_manager.py:2142] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 18:09:44.733925 139817221502784 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 18:09:44.734093 139817221502784 checkpoint_manager.py:2013] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:09:44.734749 139817221502784 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.494, Tokens/s/device: 13941.647, 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: Thu Apr 23 18:09:56 UTC 2026 EXIT_CODE=0