XPK Start: Mon Apr 20 20:59:58 UTC 2026 2026-04-20 21:00:02.188009: 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 00:00:1776718802.201895 9 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776718802.205937 9 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776718802.217454 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776718802.217472 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776718802.217475 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776718802.217477 9 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-20 21:00:21.460366: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 21:00:21.987616 137368461580096 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-20 21:00:31,028:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0420 21:00:31.028847 137368461580096 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-20 21:00:31,031:jax._src.distributed:157: Connecting to JAX distributed service on mt-03-dropout-t6y0d-slice-job-0-0.mt-03-dropout-t6y0d:8482 I0420 21:00:31.031191 137368461580096 distributed.py:157] Connecting to JAX distributed service on mt-03-dropout-t6y0d-slice-job-0-0.mt-03-dropout-t6y0d:8482 I0420 21:00:31.889081 137368461580096 max_utils.py:284] Jax distributed system initialized! I0420 21:00:38.821431 137368461580096 max_utils.py:800] System Information: Jax Version: 0.8.1 I0420 21:00:38.821538 137368461580096 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0420 21:00:38.821580 137368461580096 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0420 21:00:38.821616 137368461580096 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 21:00:39.448349 137368461580096 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 21:00:39.563092 137368461580096 checkpointing.py:688] Setting up checkpoint logger... I0420 21:00:39.563214 137368461580096 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 21:00:39.563260 137368461580096 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0420 21:00:39.563466 137368461580096 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 0x7cef11f55010>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 21:00:42.563813 137368461580096 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 21:00:42.564164 137368461580096 checkpoint_manager.py:709] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cdabc0adeb0>}, handler_registry=None I0420 21:00:42.564401 137368461580096 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cdabc0adeb0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 21:00:42.564450 137368461580096 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cda645cb170>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 21:00:42.564488 137368461580096 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 0x7cdabc0adeb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7cdabc0adeb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cda645cb170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7cda645cb170>}). I0420 21:00:42.564833 137368461580096 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0420 21:00:42.564912 137368461580096 async_checkpointer.py:177] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7cda641b8f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0420 21:00:43.709697 137368461580096 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints I0420 21:00:43.750082 137368461580096 checkpoint_manager.py:929] [process=7][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7cdabc029b20> I0420 21:00:43.750228 137368461580096 checkpointing.py:302] Checkpoint manager created! I0420 21:00:44.101053 137368461580096 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 21:00:44.101176 137368461580096 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 21:00:45.529560 137368461580096 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 21:00:45.529705 137368461580096 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:00:45.529761 137368461580096 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 21:00:45.529808 137368461580096 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 21:00:45.530034 137368461580096 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:00:45.530092 137368461580096 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 21:00:45.530137 137368461580096 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:00:45.530178 137368461580096 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 21:00:45.530217 137368461580096 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:00:45.530257 137368461580096 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 21:00:45.530293 137368461580096 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 21:00:45.530328 137368461580096 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 21:00:45.934886 137368461580096 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:00:45.935050 137368461580096 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0420 21:00:45.940348 137368461580096 nnx_decoders.py:616] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 21:00:45.940449 137368461580096 nnx_decoders.py:616] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0420 21:00:46.089285 137368461580096 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:00:46.089427 137368461580096 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0420 21:00:46.242024 137368461580096 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 21:00:46.242162 137368461580096 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0420 21:00:46.403544 137368461580096 attentions.py:1150] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:00:46.403698 137368461580096 attentions.py:1150] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0420 21:00:46.553867 137368461580096 attentions.py:1151] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:00:46.554001 137368461580096 attentions.py:1151] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0420 21:00:46.703646 137368461580096 attentions.py:1152] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 21:00:46.703805 137368461580096 attentions.py:1152] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0420 21:00:46.864109 137368461580096 attentions.py:1193] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 21:00:46.864246 137368461580096 attentions.py:1193] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0420 21:00:46.893057 137368461580096 linears.py:541] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 21:00:46.893158 137368461580096 linears.py:541] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None). I0420 21:00:50.453306 137368461580096 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. I0420 21:00:50.455648 137368461580096 metric_logger.py:289] number parameters: 1.104 billion I0420 21:00:54.200890 137368461580096 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 21:00:54.440198 137368461580096 checkpointing.py:798] Waited 0.23929834365844727 seconds for step 0 to finish before starting checkpointing. I0420 21:00:54.442785 137368461580096 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:00:54.444762 137368461580096 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 0 I0420 21:00:54.447453 137368461580096 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints/0. I0420 21:00:55.015967 137368461580096 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 21:00:55.017025 137368461580096 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:00:55.017084 137368461580096 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 21:00:55.272602 137368461580096 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.257778s I0420 21:00:55.272796 137368461580096 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.965 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5201857089996338 s) (per-host) I0420 21:00:55.272847 137368461580096 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.520247s (batch_requests_ready=0.255932s, total_serialization_initiated=0.264245s, others=0.000071s) I0420 21:00:55.272940 137368461580096 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.524556s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.524538) I0420 21:00:55.273933 137240147650304 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:00:55.274113 137368461580096 async_checkpointer.py:561] Finished blocking save. Time taken: 0.829284s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints/0. I0420 21:00:55.276039 137368461580096 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:00:55.276334 137239640143616 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:00:55.276490 137368461580096 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776718854.4427674, 'wait_for_prev_duration_secs': 6.604194641113281e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776718854.4447987, 'checkpointer_blocking_duration_secs': 0.8294260501861572, 'get_old_steps_start_time': 1776718855.2742445, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776718854.4405925, 'checkpoint_manager_blocking_duration_secs': 0.8358590602874756} I0420 21:00:55.276603 137368461580096 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 21:00:55.276685 137368461580096 max_utils.py:750] Memstats: After params initialized: I0420 21:00:55.276739 137368461580096 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_26(process=7,(2,6,0,0)) I0420 21:00:55.276774 137368461580096 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_27(process=7,(3,6,0,0)) I0420 21:00:55.276803 137368461580096 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_30(process=7,(2,7,0,0)) I0420 21:00:55.276830 137368461580096 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_31(process=7,(3,7,0,0)) I0420 21:00:55.547634 137368461580096 metric_logger.py:185] completed step: 0, seconds: 3.744, TFLOP/s/device: 0.214, Tokens/s/device: 34.186, total_weights: 4096, loss: 10.874 I0420 21:00:55.644810 137368461580096 metric_logger.py:185] completed step: 1, seconds: 1.345, TFLOP/s/device: 0.595, Tokens/s/device: 95.169, total_weights: 4096, loss: 10.882 I0420 21:00:56.065100 137368461580096 metric_logger.py:185] completed step: 2, seconds: 0.028, TFLOP/s/device: 28.287, Tokens/s/device: 4520.891, total_weights: 4096, loss: 8.922 I0420 21:00:56.135460 137368461580096 metric_logger.py:185] completed step: 3, seconds: 0.410, TFLOP/s/device: 1.952, Tokens/s/device: 311.916, total_weights: 4096, loss: 7.313 I0420 21:00:56.276702 137368461580096 metric_logger.py:185] completed step: 4, seconds: 0.087, TFLOP/s/device: 9.222, Tokens/s/device: 1473.856, total_weights: 4096, loss: 6.002 I0420 21:00:56.283995 137368461580096 metric_logger.py:185] completed step: 5, seconds: 0.070, TFLOP/s/device: 11.475, Tokens/s/device: 1833.916, total_weights: 4096, loss: 4.934 I0420 21:01:05.442475 137368461580096 metric_logger.py:185] completed step: 6, seconds: 0.142, TFLOP/s/device: 5.645, Tokens/s/device: 902.152, total_weights: 4096, loss: 4.196 I0420 21:01:05.512708 137368461580096 metric_logger.py:185] completed step: 7, seconds: 9.089, TFLOP/s/device: 0.088, Tokens/s/device: 14.082, total_weights: 4096, loss: 3.620 I0420 21:01:05.583229 137368461580096 metric_logger.py:185] completed step: 8, seconds: 0.076, TFLOP/s/device: 10.514, Tokens/s/device: 1680.363, total_weights: 4096, loss: 3.238 I0420 21:01:05.654359 137368461580096 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 21:01:05.654954 137368461580096 checkpointing.py:798] Waited 0.0006031990051269531 seconds for step 9 to finish before starting checkpointing. I0420 21:01:05.657326 137368461580096 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:01:07.039775 2632 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 21:01:09.554748 137241731073792 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints/0/items/array_metadatas/process_7 I0420 21:01:36.337666 137240147650304 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 37.985 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.58500623703003 s) (per-host) I0420 21:01:36.337785 137240147650304 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.063733s. I0420 21:01:44.940597 137240147650304 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.666527s. I0420 21:01:44.940904 137239640143616 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:01:44.941024 137239640143616 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:01:44.941074 137239640143616 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:01:44.942857 137239640143616 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 21:01:44.943014 137368461580096 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 21:01:44.943155 137368461580096 checkpoint_manager.py:1458] Waiting for previous save to complete took 39.285837 seconds. If this number is high, consider checkpointing less frequently. I0420 21:01:44.945281 137368461580096 checkpoint_manager.py:1518] [process=7] Saving checkpoint at step 9 I0420 21:01:44.948766 137368461580096 async_checkpointer.py:452] [process=7] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints/9. I0420 21:01:45.503020 137368461580096 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0420 21:01:45.503177 137368461580096 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 21:01:45.552831 137368461580096 base_pytree_checkpoint_handler.py:153] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.051777s I0420 21:01:45.552998 137368461580096 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.815 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.32038116455078125 s) (per-host) I0420 21:01:45.553049 137368461580096 base_pytree_checkpoint_handler.py:737] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.320441s (batch_requests_ready=0.263945s, total_serialization_initiated=0.056427s, others=0.000068s) I0420 21:01:45.553143 137368461580096 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.324442s (all_items=0.000014s, per_item={'items': '0.00001407'}, temp_paths=0.324428) I0420 21:01:45.554020 137241731073792 async_checkpointer.py:79] [process=7][thread=async_save] Background save thread started. I0420 21:01:45.554183 137368461580096 async_checkpointer.py:561] Finished blocking save. Time taken: 0.608830s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints/9. I0420 21:01:45.973897 137368461580096 checkpoint_manager.py:1566] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 21:01:45.974259 137239640143616 async_checkpointer.py:265] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0420 21:01:45.974432 137368461580096 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776718865.6572921, 'wait_for_prev_duration_secs': 39.285836696624756, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776718904.9453197, 'checkpointer_blocking_duration_secs': 0.6089746952056885, 'get_old_steps_start_time': 1776718905.5543132, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776718865.6551769, 'checkpoint_manager_blocking_duration_secs': 40.3192195892334} I0420 21:01:45.974550 137368461580096 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 21:01:45.974597 137368461580096 checkpoint_manager.py:2024] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 21:01:50.815372 137234079418112 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260420_190404/nnx_xpk_feat_nnx_set_defaults_true_20260420_190404_03_dropout/checkpoints/9/items/array_metadatas/process_7 I0420 21:02:26.678216 137241731073792 base_pytree_checkpoint_handler.py:129] [process=7] /jax/checkpoint/write/gbytes_per_sec: 38.113 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.445557594299316 s) (per-host) I0420 21:02:26.678342 137241731073792 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.124209s. I0420 21:02:35.372134 137241731073792 async_checkpointer.py:144] [process=7][thread=async_save] Background save thread done. Time taken: 49.817985s. I0420 21:02:35.372373 137239640143616 async_checkpointer.py:273] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 21:02:35.372436 137239640143616 async_checkpointer.py:283] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0420 21:02:35.372485 137239640143616 checkpoint_manager.py:2133] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 21:02:35.374056 137239640143616 checkpoint_manager.py:2142] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 21:02:35.374217 137368461580096 checkpoint_manager.py:2036] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 21:02:35.374386 137368461580096 checkpoint_manager.py:2013] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 21:02:35.375103 137368461580096 metric_logger.py:185] completed step: 9, seconds: 0.070, TFLOP/s/device: 11.467, Tokens/s/device: 1832.761, 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: Mon Apr 20 21:02:43 UTC 2026 EXIT_CODE=0