XPK Start: Sat Apr 25 16:40:04 UTC 2026 2026-04-25 16:40:08.656126: 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 00:00:1777135208.669143 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1777135208.672905 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1777135208.684582 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777135208.684602 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777135208.684605 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1777135208.684607 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-25 16:40:27.868773: 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) I0425 16:40:28.402892 139885733779264 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-25 16:40:37,444:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0425 16:40:37.444792 139885733779264 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-25 16:40:37,447:jax._src.distributed:157: Connecting to JAX distributed service on mt-03-dropout-fisaq-slice-job-0-0.mt-03-dropout-fisaq:8482 I0425 16:40:37.447207 139885733779264 distributed.py:157] Connecting to JAX distributed service on mt-03-dropout-fisaq-slice-job-0-0.mt-03-dropout-fisaq:8482 I0425 16:40:38.434606 139885733779264 max_utils.py:284] Jax distributed system initialized! I0425 16:40:44.294988 139885733779264 max_utils.py:800] System Information: Jax Version: 0.8.1 I0425 16:40:44.295110 139885733779264 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0425 16:40:44.295155 139885733779264 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 I0425 16:40:44.295191 139885733779264 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0425 16:40:44.921741 139885733779264 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0425 16:40:45.038794 139885733779264 checkpointing.py:688] Setting up checkpoint logger... I0425 16:40:45.038911 139885733779264 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0425 16:40:45.038956 139885733779264 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0425 16:40:45.039174 139885733779264 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 0x7f392b179610>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0425 16:40:47.919489 139885733779264 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0425 16:40:47.919838 139885733779264 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 0x7f37ea2ffd70>}, handler_registry=None I0425 16:40:47.920076 139885733779264 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f37ea2ffd70>` 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`. I0425 16:40:47.920140 139885733779264 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f24fc6a4590>` 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`. I0425 16:40:47.920178 139885733779264 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 0x7f37ea2ffd70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f37ea2ffd70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f24fc6a4590>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f24fc6a4590>}). I0425 16:40:47.920534 139885733779264 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0425 16:40:47.920611 139885733779264 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f2646b54f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0425 16:40:49.071231 139885733779264 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints I0425 16:40:49.496004 139885733779264 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f2646fecd10> I0425 16:40:49.496205 139885733779264 checkpointing.py:302] Checkpoint manager created! I0425 16:40:49.850414 139885733779264 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0425 16:40:49.850532 139885733779264 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0425 16:40:51.309386 139885733779264 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0425 16:40:51.309510 139885733779264 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 16:40:51.309566 139885733779264 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0425 16:40:51.309617 139885733779264 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0425 16:40:51.309657 139885733779264 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 16:40:51.309693 139885733779264 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0425 16:40:51.309732 139885733779264 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:40:51.309771 139885733779264 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0425 16:40:51.309811 139885733779264 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:40:51.309849 139885733779264 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0425 16:40:51.309884 139885733779264 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0425 16:40:51.309918 139885733779264 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0425 16:40:51.720168 139885733779264 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:40:51.720320 139885733779264 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 16:40:51.725716 139885733779264 nnx_decoders.py:616] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0425 16:40:51.725813 139885733779264 nnx_decoders.py:616] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 16:40:51.875781 139885733779264 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:40:51.875917 139885733779264 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 16:40:52.026967 139885733779264 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0425 16:40:52.027203 139885733779264 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0425 16:40:52.183215 139885733779264 attentions.py:1150] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:40:52.183362 139885733779264 attentions.py:1150] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 16:40:52.327429 139885733779264 attentions.py:1151] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:40:52.327561 139885733779264 attentions.py:1151] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 16:40:52.471298 139885733779264 attentions.py:1152] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0425 16:40:52.471435 139885733779264 attentions.py:1152] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 16:40:52.623366 139885733779264 attentions.py:1193] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0425 16:40:52.623508 139885733779264 attentions.py:1193] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0425 16:40:52.652390 139885733779264 linears.py:541] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0425 16:40:52.652490 139885733779264 linears.py:541] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None). I0425 16:40:56.160503 139885733779264 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. I0425 16:40:56.162901 139885733779264 metric_logger.py:289] number parameters: 1.104 billion I0425 16:40:59.994524 139885733779264 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0425 16:41:00.064725 139885733779264 checkpointing.py:798] Waited 0.0701911449432373 seconds for step 0 to finish before starting checkpointing. I0425 16:41:00.067161 139885733779264 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:41:00.069290 139885733779264 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0425 16:41:00.072050 139885733779264 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints/0. I0425 16:41:00.615558 139885733779264 signaling_client.py:364] Using JaxDistributedSignalingClient I0425 16:41:00.616672 139885733779264 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 16:41:00.616732 139885733779264 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 I0425 16:41:00.864935 139885733779264 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.250529s I0425 16:41:00.865129 139885733779264 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.018 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5111472606658936 s) (per-host) I0425 16:41:00.865184 139885733779264 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.511212s (batch_requests_ready=0.254085s, total_serialization_initiated=0.257042s, others=0.000085s) I0425 16:41:00.865289 139885733779264 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.515511s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.515491) I0425 16:41:00.866158 139756893632256 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:41:00.866266 139885733779264 async_checkpointer.py:561] Finished blocking save. Time taken: 0.796908s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints/0. I0425 16:41:00.870912 139885733779264 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:41:00.871140 139758479066880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:41:00.871256 139885733779264 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777135260.0671432, 'wait_for_prev_duration_secs': 6.103515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777135260.0693305, 'checkpointer_blocking_duration_secs': 0.7970583438873291, 'get_old_steps_start_time': 1777135260.8664112, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1777135260.0651202, 'checkpoint_manager_blocking_duration_secs': 0.8060998916625977} I0425 16:41:00.871367 139885733779264 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0425 16:41:00.871435 139885733779264 max_utils.py:750] Memstats: After params initialized: I0425 16:41:00.871483 139885733779264 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_24(process=6,(0,6,0,0)) I0425 16:41:00.871515 139885733779264 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_25(process=6,(1,6,0,0)) I0425 16:41:00.871543 139885733779264 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_28(process=6,(0,7,0,0)) I0425 16:41:00.871567 139885733779264 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_29(process=6,(1,7,0,0)) I0425 16:41:01.140543 139885733779264 metric_logger.py:185] completed step: 0, seconds: 3.830, TFLOP/s/device: 0.209, Tokens/s/device: 33.416, total_weights: 4096, loss: 10.874 I0425 16:41:01.235900 139885733779264 metric_logger.py:185] completed step: 1, seconds: 1.144, TFLOP/s/device: 0.700, Tokens/s/device: 111.856, total_weights: 4096, loss: 10.882 I0425 16:41:01.657568 139885733779264 metric_logger.py:185] completed step: 2, seconds: 0.027, TFLOP/s/device: 29.709, Tokens/s/device: 4748.303, total_weights: 4096, loss: 8.922 I0425 16:41:01.727679 139885733779264 metric_logger.py:185] completed step: 3, seconds: 0.411, TFLOP/s/device: 1.948, Tokens/s/device: 311.367, total_weights: 4096, loss: 7.313 I0425 16:41:01.869004 139885733779264 metric_logger.py:185] completed step: 4, seconds: 0.087, TFLOP/s/device: 9.211, Tokens/s/device: 1472.093, total_weights: 4096, loss: 6.002 I0425 16:41:01.876133 139885733779264 metric_logger.py:185] completed step: 5, seconds: 0.070, TFLOP/s/device: 11.523, Tokens/s/device: 1841.700, total_weights: 4096, loss: 4.934 I0425 16:41:11.121143 139885733779264 metric_logger.py:185] completed step: 6, seconds: 0.142, TFLOP/s/device: 5.641, Tokens/s/device: 901.561, total_weights: 4096, loss: 4.196 I0425 16:41:11.191190 139885733779264 metric_logger.py:185] completed step: 7, seconds: 9.176, TFLOP/s/device: 0.087, Tokens/s/device: 13.950, total_weights: 4096, loss: 3.620 I0425 16:41:11.261522 139885733779264 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 I0425 16:41:11.332649 139885733779264 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0425 16:41:11.333246 139885733779264 checkpointing.py:798] Waited 0.0006036758422851562 seconds for step 9 to finish before starting checkpointing. I0425 16:41:11.335689 139885733779264 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:41:12.676453 2634 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0425 16:41:14.524492 139758496462592 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints/0/items/array_metadatas/process_6 I0425 16:41:41.211501 139756893632256 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 38.661 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.857481956481934 s) (per-host) I0425 16:41:41.211610 139756893632256 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.345390s. I0425 16:41:49.259131 139756893632256 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 48.392866s. I0425 16:41:49.259347 139758479066880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:41:49.259406 139758479066880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:41:49.259455 139758479066880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:41:49.262157 139758479066880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0425 16:41:49.262368 139885733779264 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. W0425 16:41:49.262501 139885733779264 checkpoint_manager.py:1458] Waiting for previous save to complete took 37.926819 seconds. If this number is high, consider checkpointing less frequently. I0425 16:41:49.264550 139885733779264 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0425 16:41:49.267929 139885733779264 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints/9. I0425 16:41:50.219940 139885733779264 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0425 16:41:50.220088 139885733779264 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 I0425 16:41:50.265990 139885733779264 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.048161s I0425 16:41:50.266175 139885733779264 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.102 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3023552894592285 s) (per-host) I0425 16:41:50.266231 139885733779264 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.302421s (batch_requests_ready=0.249436s, total_serialization_initiated=0.052912s, others=0.000073s) I0425 16:41:50.266330 139885733779264 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.306676s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.306660) I0425 16:41:50.267215 139760098072320 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0425 16:41:50.267373 139885733779264 async_checkpointer.py:561] Finished blocking save. Time taken: 1.002755s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints/9. I0425 16:41:50.278451 139885733779264 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0425 16:41:50.278682 139758479066880 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0425 16:41:50.278811 139885733779264 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777135271.335658, 'wait_for_prev_duration_secs': 37.92681908607483, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777135309.2645888, 'checkpointer_blocking_duration_secs': 1.0028889179229736, 'get_old_steps_start_time': 1777135310.267496, 'get_old_steps_duration_secs': 2.7179718017578125e-05, 'checkpoint_manager_blocking_start_time': 1777135271.3334653, 'checkpoint_manager_blocking_duration_secs': 38.94531273841858} I0425 16:41:50.278919 139885733779264 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0425 16:41:50.278963 139885733779264 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0425 16:41:54.947083 139733510297344 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_20260425_150412/nnx_xpk_feat_nnx_set_defaults_true_20260425_150412_03_dropout/checkpoints/9/items/array_metadatas/process_6 I0425 16:42:30.309088 139760098072320 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 39.152 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.34522604942322 s) (per-host) I0425 16:42:30.309238 139760098072320 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.041907s. I0425 16:42:37.088536 139760098072320 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 46.821196s. I0425 16:42:37.088739 139758479066880 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0425 16:42:37.088796 139758479066880 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0425 16:42:37.088844 139758479066880 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0425 16:42:37.090791 139758479066880 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0425 16:42:37.090984 139885733779264 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. I0425 16:42:37.091171 139885733779264 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0425 16:42:37.091836 139885733779264 metric_logger.py:185] completed step: 9, seconds: 0.070, TFLOP/s/device: 11.514, Tokens/s/device: 1840.217, 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: Sat Apr 25 16:42:50 UTC 2026 EXIT_CODE=0