XPK Start: Sun Apr 19 05:27:26 UTC 2026 2026-04-19 05:27:30.941878: 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 00:00:1776576450.954535 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776576450.958271 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776576450.969650 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776576450.969669 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776576450.969671 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776576450.969673 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 05:27:50.193914: 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) I0419 05:27:50.725338 133349908354880 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 05:27:59,765:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 05:27:59.765481 133349908354880 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 05:27:59,767:jax._src.distributed:157: Connecting to JAX distributed service on mt-10-shardy-true-e8hdw-slice-job-0-0.mt-10-shardy-true-e8hdw:8482 I0419 05:27:59.767722 133349908354880 distributed.py:157] Connecting to JAX distributed service on mt-10-shardy-true-e8hdw-slice-job-0-0.mt-10-shardy-true-e8hdw:8482 I0419 05:28:00.950672 133349908354880 max_utils.py:284] Jax distributed system initialized! I0419 05:28:07.885967 133349908354880 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 05:28:07.886076 133349908354880 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 05:28:07.886117 133349908354880 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 I0419 05:28:07.886155 133349908354880 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 05:28:08.508279 133349908354880 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 05:28:08.625148 133349908354880 checkpointing.py:688] Setting up checkpoint logger... I0419 05:28:08.625302 133349908354880 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 05:28:08.625365 133349908354880 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 05:28:08.625604 133349908354880 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 0x79476d975160>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 05:28:12.756737 133349908354880 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 05:28:12.756974 133349908354880 checkpoint_manager.py:709] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7932d8740650>}, handler_registry=None I0419 05:28:12.757223 133349908354880 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7932d8740650>` 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`. I0419 05:28:12.757288 133349908354880 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79340b958170>` 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`. I0419 05:28:12.757326 133349908354880 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 0x7932d8740650>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7932d8740650>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79340b958170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79340b958170>}). I0419 05:28:12.757670 133349908354880 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 05:28:12.757746 133349908354880 async_checkpointer.py:177] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7932d87e4f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 05:28:13.927644 133349908354880 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints I0419 05:28:14.766550 133349908354880 checkpoint_manager.py:929] [process=3][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_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79340b95a7b0> I0419 05:28:14.766741 133349908354880 checkpointing.py:302] Checkpoint manager created! I0419 05:28:15.697139 133349908354880 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 05:28:15.697277 133349908354880 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 05:28:17.173784 133349908354880 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 05:28:17.173913 133349908354880 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 05:28:17.173970 133349908354880 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 05:28:17.174015 133349908354880 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 05:28:17.174053 133349908354880 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 05:28:17.174089 133349908354880 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 05:28:17.174127 133349908354880 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:28:17.174165 133349908354880 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 05:28:17.174202 133349908354880 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:28:17.174248 133349908354880 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:28:17.174287 133349908354880 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 05:28:17.174322 133349908354880 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 05:28:17.571333 133349908354880 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 05:28:17.571488 133349908354880 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:28:17.576992 133349908354880 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 05:28:17.577089 133349908354880 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:28:17.733784 133349908354880 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 05:28:17.733920 133349908354880 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:28:17.887483 133349908354880 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 05:28:17.887618 133349908354880 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:28:18.053078 133349908354880 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 05:28:18.053217 133349908354880 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:28:18.210514 133349908354880 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 05:28:18.210650 133349908354880 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:28:18.367619 133349908354880 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 05:28:18.367756 133349908354880 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:28:18.533996 133349908354880 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 05:28:18.534139 133349908354880 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:28:18.557428 133349908354880 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 05:28:18.557527 133349908354880 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 05:28:29.660376 133349908354880 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. I0419 05:28:29.662733 133349908354880 metric_logger.py:289] number parameters: 1.104 billion I0419 05:28:40.642117 133349908354880 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 05:28:40.793275 133349908354880 checkpointing.py:798] Waited 0.1511092185974121 seconds for step 0 to finish before starting checkpointing. I0419 05:28:40.795656 133349908354880 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 05:28:40.797489 133349908354880 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 0 I0419 05:28:40.800167 133349908354880 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints/0. I0419 05:28:42.203174 133349908354880 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 05:28:42.204453 133349908354880 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 05:28:42.204513 133349908354880 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 I0419 05:28:42.451663 133349908354880 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.249611s I0419 05:28:42.451834 133349908354880 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.039 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5076069831848145 s) (per-host) I0419 05:28:42.451887 133349908354880 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.507669s (batch_requests_ready=0.251500s, total_serialization_initiated=0.256100s, others=0.000070s) I0419 05:28:42.451984 133349908354880 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.511952s (all_items=0.000022s, per_item={'items': '0.00002170'}, temp_paths=0.511930) I0419 05:28:42.452987 133227503138560 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0419 05:28:42.453162 133349908354880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.655603s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints/0. I0419 05:28:43.311182 133349908354880 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 05:28:43.311610 133223728256768 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0419 05:28:43.311784 133349908354880 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776576520.7956393, 'wait_for_prev_duration_secs': 6.461143493652344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776576520.7975268, 'checkpointer_blocking_duration_secs': 1.6557588577270508, 'get_old_steps_start_time': 1776576522.4533076, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776576520.7936707, 'checkpoint_manager_blocking_duration_secs': 2.51806902885437} I0419 05:28:43.311898 133349908354880 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 05:28:43.311973 133349908354880 max_utils.py:750] Memstats: After params initialized: I0419 05:28:43.312032 133349908354880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_10(process=3,(2,2,0,0)) I0419 05:28:43.312067 133349908354880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_11(process=3,(3,2,0,0)) I0419 05:28:43.312096 133349908354880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_14(process=3,(2,3,0,0)) I0419 05:28:43.312129 133349908354880 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_15(process=3,(3,3,0,0)) I0419 05:28:43.583127 133349908354880 metric_logger.py:185] completed step: 0, seconds: 10.978, TFLOP/s/device: 1.238, Tokens/s/device: 186.549, total_weights: 65536, loss: 10.874 I0419 05:28:43.750035 133349908354880 metric_logger.py:185] completed step: 1, seconds: 2.939, TFLOP/s/device: 4.623, Tokens/s/device: 696.787, total_weights: 65536, loss: 10.874 I0419 05:28:44.163506 133349908354880 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 656.383, Tokens/s/device: 98937.198, total_weights: 65536, loss: 10.262 I0419 05:28:44.311104 133349908354880 metric_logger.py:185] completed step: 3, seconds: 0.404, TFLOP/s/device: 33.598, Tokens/s/device: 5064.243, total_weights: 65536, loss: 9.732 I0419 05:28:44.607162 133349908354880 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.146, Tokens/s/device: 12532.739, total_weights: 65536, loss: 9.274 I0419 05:28:44.614478 133349908354880 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.331, Tokens/s/device: 13917.204, total_weights: 65536, loss: 8.888 I0419 05:29:05.916730 133349908354880 metric_logger.py:185] completed step: 6, seconds: 0.297, TFLOP/s/device: 45.823, Tokens/s/device: 6906.948, total_weights: 65536, loss: 8.589 I0419 05:29:06.064335 133349908354880 metric_logger.py:185] completed step: 7, seconds: 21.156, TFLOP/s/device: 0.642, Tokens/s/device: 96.804, total_weights: 65536, loss: 8.380 I0419 05:29:06.211985 133349908354880 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.517, Tokens/s/device: 13342.280, total_weights: 65536, loss: 8.251 I0419 05:29:06.360886 133349908354880 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 05:29:06.361519 133349908354880 checkpointing.py:798] Waited 0.0006535053253173828 seconds for step 9 to finish before starting checkpointing. I0419 05:29:06.364012 133349908354880 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 05:29:08.663802 2626 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 05:29:11.464167 133227511531264 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints/0/items/array_metadatas/process_3 I0419 05:29:42.265964 133227503138560 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 26.186 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 60.32169723510742 s) (per-host) I0419 05:29:42.266083 133227503138560 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 59.812984s. I0419 05:29:54.127982 133227503138560 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 71.674865s. I0419 05:29:54.128284 133223728256768 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 05:29:54.128418 133223728256768 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0419 05:29:54.128471 133223728256768 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 05:29:54.130122 133223728256768 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 05:29:54.130306 133349908354880 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 05:29:54.130444 133349908354880 checkpoint_manager.py:1458] Waiting for previous save to complete took 47.766436 seconds. If this number is high, consider checkpointing less frequently. I0419 05:29:54.132361 133349908354880 checkpoint_manager.py:1518] [process=3] Saving checkpoint at step 9 I0419 05:29:54.135815 133349908354880 async_checkpointer.py:452] [process=3] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints/9. I0419 05:29:55.521142 133349908354880 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 05:29:55.521308 133349908354880 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 I0419 05:29:55.565216 133349908354880 base_pytree_checkpoint_handler.py:153] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.046047s I0419 05:29:55.565398 133349908354880 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.081 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30362558364868164 s) (per-host) I0419 05:29:55.565447 133349908354880 base_pytree_checkpoint_handler.py:737] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.303683s (batch_requests_ready=0.252888s, total_serialization_initiated=0.050730s, others=0.000066s) I0419 05:29:55.565536 133349908354880 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.308056s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.308039) I0419 05:29:55.566419 133224797579008 async_checkpointer.py:79] [process=3][thread=async_save] Background save thread started. I0419 05:29:55.566542 133349908354880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.434114s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints/9. I0419 05:29:56.422484 133349908354880 checkpoint_manager.py:1566] [process=3][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 05:29:56.422882 133223728256768 async_checkpointer.py:265] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0419 05:29:56.423054 133349908354880 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776576546.3639836, 'wait_for_prev_duration_secs': 47.766435861587524, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776576594.1323984, 'checkpointer_blocking_duration_secs': 1.434309482574463, 'get_old_steps_start_time': 1776576595.5667267, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1776576546.3617404, 'checkpoint_manager_blocking_duration_secs': 50.06127595901489} I0419 05:29:56.423168 133349908354880 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 05:29:56.423214 133349908354880 checkpoint_manager.py:2024] [process=3][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 05:30:03.161912 133227511531264 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_10_shardy_true/checkpoints/9/items/array_metadatas/process_3 I0419 05:30:39.563462 133224797579008 base_pytree_checkpoint_handler.py:129] [process=3] /jax/checkpoint/write/gbytes_per_sec: 35.655 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.30165243148804 s) (per-host) I0419 05:30:39.563574 133224797579008 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.996957s. I0419 05:30:49.203415 133224797579008 async_checkpointer.py:144] [process=3][thread=async_save] Background save thread done. Time taken: 53.636784s. I0419 05:30:49.203724 133223728256768 async_checkpointer.py:273] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 05:30:49.203847 133223728256768 async_checkpointer.py:283] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0419 05:30:49.203905 133223728256768 checkpoint_manager.py:2133] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 05:30:49.205664 133223728256768 checkpoint_manager.py:2142] [process=3][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 05:30:49.205826 133349908354880 checkpoint_manager.py:2036] [process=3][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 05:30:49.205994 133349908354880 checkpoint_manager.py:2013] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 05:30:49.206689 133349908354880 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.441, Tokens/s/device: 13933.679, 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: Sun Apr 19 05:31:02 UTC 2026 EXIT_CODE=0