XPK Start: Thu Apr 23 18:47:36 UTC 2026 2026-04-23 18:47:40.811932: 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 00:00:1776970060.824793 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776970060.828589 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776970060.840436 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970060.840457 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970060.840459 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776970060.840461 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-23 18:48:00.048753: 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 18:48:00.574214 139164146304832 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-23 18:48:09,614:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0423 18:48:09.614412 139164146304832 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-23 18:48:09,616:jax._src.distributed:157: Connecting to JAX distributed service on mt-16-zarr3-true-sv-r4vm4-slice-job-0-0.mt-16-zarr3-true-sv-r4vm4:8482 I0423 18:48:09.616801 139164146304832 distributed.py:157] Connecting to JAX distributed service on mt-16-zarr3-true-sv-r4vm4-slice-job-0-0.mt-16-zarr3-true-sv-r4vm4:8482 I0423 18:48:10.663903 139164146304832 max_utils.py:284] Jax distributed system initialized! I0423 18:48:17.447896 139164146304832 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 18:48:17.448003 139164146304832 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 18:48:17.448045 139164146304832 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 18:48:17.448080 139164146304832 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 18:48:18.070848 139164146304832 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 18:48:18.186009 139164146304832 checkpointing.py:688] Setting up checkpoint logger... I0423 18:48:18.186188 139164146304832 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 18:48:18.186241 139164146304832 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0423 18:48:18.186462 139164146304832 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 0x7e912bb75580>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 18:48:21.201304 139164146304832 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0423 18:48:21.201656 139164146304832 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 0x7e7cd43a5ee0>}, handler_registry=None I0423 18:48:21.201895 139164146304832 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e7cd43a5ee0>` 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 18:48:21.201944 139164146304832 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e7dd55cf380>` 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 18:48:21.201981 139164146304832 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 0x7e7cd43a5ee0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e7cd43a5ee0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e7dd55cf380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e7dd55cf380>}). I0423 18:48:21.202334 139164146304832 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0423 18:48:21.202411 139164146304832 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e7dd4becf40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 18:48:22.743192 139164146304832 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_16_zarr3_true/checkpoints I0423 18:48:22.745452 139164146304832 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=5), 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_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e7dd6a035c0> I0423 18:48:22.745576 139164146304832 checkpointing.py:302] Checkpoint manager created! I0423 18:48:23.676709 139164146304832 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0423 18:48:23.676830 139164146304832 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0423 18:48:25.079501 139164146304832 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0423 18:48:25.079625 139164146304832 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:48:25.079704 139164146304832 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0423 18:48:25.079779 139164146304832 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0423 18:48:25.079909 139164146304832 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:48:25.079977 139164146304832 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0423 18:48:25.080045 139164146304832 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:48:25.080120 139164146304832 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0423 18:48:25.080184 139164146304832 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:48:25.080247 139164146304832 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0423 18:48:25.080321 139164146304832 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0423 18:48:25.080385 139164146304832 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0423 18:48:25.480934 139164146304832 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:48:25.481087 139164146304832 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:48:25.486645 139164146304832 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 18:48:25.486743 139164146304832 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:48:25.649410 139164146304832 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:48:25.649561 139164146304832 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:48:25.809071 139164146304832 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 18:48:25.809226 139164146304832 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0423 18:48:25.978461 139164146304832 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 18:48:25.978601 139164146304832 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:48:26.141299 139164146304832 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 18:48:26.141435 139164146304832 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:48:26.303361 139164146304832 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 18:48:26.303493 139164146304832 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:48:26.472610 139164146304832 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 18:48:26.472753 139164146304832 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0423 18:48:26.497169 139164146304832 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 18:48:26.497268 139164146304832 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0423 18:48:37.585281 139164146304832 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 18:48:37.587784 139164146304832 metric_logger.py:289] number parameters: 1.104 billion I0423 18:48:48.814287 139164146304832 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0423 18:48:48.965370 139164146304832 checkpointing.py:798] Waited 0.15107083320617676 seconds for step 0 to finish before starting checkpointing. I0423 18:48:48.967678 139164146304832 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:48:48.969467 139164146304832 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0 I0423 18:48:48.972259 139164146304832 async_checkpointer.py:452] [process=6] 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_16_zarr3_true/checkpoints/0. I0423 18:48:49.583919 139164146304832 signaling_client.py:364] Using JaxDistributedSignalingClient I0423 18:48:49.585146 139164146304832 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:48:49.585205 139164146304832 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 18:48:49.827996 139164146304832 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.245234s I0423 18:48:49.828188 139164146304832 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.884 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5349252223968506 s) (per-host) I0423 18:48:49.828247 139164146304832 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.534994s (batch_requests_ready=0.283104s, total_serialization_initiated=0.251813s, others=0.000077s) I0423 18:48:49.828346 139164146304832 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.539179s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=0.539159) I0423 18:48:49.829418 139037302843136 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:48:49.829548 139164146304832 async_checkpointer.py:561] Finished blocking save. Time taken: 0.860003s. 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_16_zarr3_true/checkpoints/0. I0423 18:48:49.861854 139164146304832 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:48:49.862178 139037847299840 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:48:49.862341 139164146304832 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970128.967661, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776970128.969513, 'checkpointer_blocking_duration_secs': 0.8602337837219238, 'get_old_steps_start_time': 1776970129.8297677, 'get_old_steps_duration_secs': 2.4318695068359375e-05, 'checkpoint_manager_blocking_start_time': 1776970128.9657683, 'checkpoint_manager_blocking_duration_secs': 0.8965344429016113} I0423 18:48:49.862447 139164146304832 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0423 18:48:49.862520 139164146304832 max_utils.py:750] Memstats: After params initialized: I0423 18:48:49.862571 139164146304832 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0423 18:48:49.862610 139164146304832 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0423 18:48:49.862637 139164146304832 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0423 18:48:49.862662 139164146304832 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0423 18:48:50.132778 139164146304832 metric_logger.py:185] completed step: 0, seconds: 11.225, TFLOP/s/device: 1.210, Tokens/s/device: 182.442, total_weights: 65536, loss: 10.874 I0423 18:48:50.298602 139164146304832 metric_logger.py:185] completed step: 1, seconds: 1.317, TFLOP/s/device: 10.320, Tokens/s/device: 1555.527, total_weights: 65536, loss: 10.874 I0423 18:48:50.712413 139164146304832 metric_logger.py:185] completed step: 2, seconds: 0.021, TFLOP/s/device: 661.947, Tokens/s/device: 99775.894, total_weights: 65536, loss: 10.560 I0423 18:48:50.859982 139164146304832 metric_logger.py:185] completed step: 3, seconds: 0.404, TFLOP/s/device: 33.637, Tokens/s/device: 5070.160, total_weights: 65536, loss: 9.979 I0423 18:48:50.865797 139164146304832 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0423 18:48:51.154625 139164146304832 checkpointing.py:798] Waited 0.2887897491455078 seconds for step 5 to finish before starting checkpointing. I0423 18:48:51.157787 139164146304832 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:48:52.536242 2659 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0423 18:48:54.821928 139037352392448 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/0/items/array_metadatas/process_6 I0423 18:49:27.462944 139037302843136 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 41.384 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.169639587402344 s) (per-host) I0423 18:49:27.463063 139037302843136 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.633425s. I0423 18:49:36.735188 139037302843136 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 46.905533s. I0423 18:49:36.735523 139037847299840 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:49:36.735649 139037847299840 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:49:36.735701 139037847299840 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:49:36.737465 139037847299840 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 18:49:36.737640 139164146304832 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. W0423 18:49:36.737798 139164146304832 checkpoint_manager.py:1458] Waiting for previous save to complete took 45.580014 seconds. If this number is high, consider checkpointing less frequently. I0423 18:49:36.739934 139164146304832 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 5 I0423 18:49:36.743905 139164146304832 async_checkpointer.py:452] [process=6] 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_16_zarr3_true/checkpoints/5. I0423 18:49:37.279850 139164146304832 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:49:37.280002 139164146304832 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 18:49:37.321091 139164146304832 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.043262s I0423 18:49:37.321267 139164146304832 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.107 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30206727981567383 s) (per-host) I0423 18:49:37.321335 139164146304832 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.302141s (batch_requests_ready=0.253511s, total_serialization_initiated=0.048548s, others=0.000082s) I0423 18:49:37.321450 139164146304832 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.307039s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.307023) I0423 18:49:37.322299 139035188123392 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:49:37.322458 139164146304832 async_checkpointer.py:561] Finished blocking save. Time taken: 0.582454s. 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_16_zarr3_true/checkpoints/5. I0423 18:49:37.332915 139164146304832 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:49:37.333232 139037847299840 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:49:37.333401 139164146304832 standard_logger.py:34] {'step': 5, '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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970131.1577585, 'wait_for_prev_duration_secs': 45.58001375198364, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776970176.7399733, 'checkpointer_blocking_duration_secs': 0.582592248916626, 'get_old_steps_start_time': 1776970177.322584, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776970131.154969, 'checkpoint_manager_blocking_duration_secs': 46.1783983707428} I0423 18:49:37.333507 139164146304832 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0423 18:49:37.334264 139164146304832 metric_logger.py:185] completed step: 4, seconds: 0.163, TFLOP/s/device: 83.203, Tokens/s/device: 12541.258, total_weights: 65536, loss: 9.457 I0423 18:49:37.341962 139164146304832 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.374, Tokens/s/device: 13923.543, total_weights: 65536, loss: 8.955 I0423 18:49:57.001335 139164146304832 metric_logger.py:185] completed step: 6, seconds: 46.475, TFLOP/s/device: 0.292, Tokens/s/device: 44.067, total_weights: 65536, loss: 8.465 I0423 18:49:57.148790 139164146304832 metric_logger.py:185] completed step: 7, seconds: 19.514, TFLOP/s/device: 0.696, Tokens/s/device: 104.950, total_weights: 65536, loss: 7.998 I0423 18:49:57.296228 139164146304832 metric_logger.py:185] completed step: 8, seconds: 0.154, TFLOP/s/device: 88.467, Tokens/s/device: 13334.722, total_weights: 65536, loss: 7.567 I0423 18:49:57.444525 139164146304832 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0423 18:49:57.445148 139164146304832 checkpointing.py:798] Waited 0.0006341934204101562 seconds for step 9 to finish before starting checkpointing. I0423 18:49:57.447241 139164146304832 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:50:01.458305 139037352392448 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/5/items/array_metadatas/process_6 I0423 18:50:38.560968 139035188123392 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 25.667 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.541725397109985 s) (per-host) I0423 18:50:38.561091 139035188123392 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.238681s. I0423 18:50:45.688720 139035188123392 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 68.366294s. I0423 18:50:45.689017 139037847299840 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:50:45.689187 139037847299840 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:50:45.689248 139037847299840 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:50:45.691751 139037847299840 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0423 18:50:45.691938 139164146304832 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0423 18:50:45.692073 139164146304832 checkpoint_manager.py:1458] Waiting for previous save to complete took 48.244842 seconds. If this number is high, consider checkpointing less frequently. I0423 18:50:45.693512 139164146304832 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9 I0423 18:50:45.696776 139164146304832 async_checkpointer.py:452] [process=6] 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_16_zarr3_true/checkpoints/9. I0423 18:50:46.261185 139164146304832 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0423 18:50:46.261286 139164146304832 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 18:50:46.303826 139164146304832 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044653s I0423 18:50:46.303995 139164146304832 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.032 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.30655622482299805 s) (per-host) I0423 18:50:46.304046 139164146304832 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.306617s (batch_requests_ready=0.257210s, total_serialization_initiated=0.049341s, others=0.000067s) I0423 18:50:46.304160 139164146304832 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.310785s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.310774) I0423 18:50:46.305063 139037302843136 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started. I0423 18:50:46.305241 139164146304832 async_checkpointer.py:561] Finished blocking save. Time taken: 0.611658s. 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_16_zarr3_true/checkpoints/9. I0423 18:50:46.327327 139164146304832 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 18:50:46.327593 139037847299840 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0423 18:50:46.327747 139164146304832 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776970197.4472148, 'wait_for_prev_duration_secs': 48.24484181404114, 'time_between_consecutive_saves_sec': 20.70970320701599, 'checkpointer_blocking_start_time': 1776970245.6935494, 'checkpointer_blocking_duration_secs': 0.6118073463439941, 'get_old_steps_start_time': 1776970246.3053708, 'get_old_steps_duration_secs': 1.9311904907226562e-05, 'checkpoint_manager_blocking_start_time': 1776970197.4453719, 'checkpoint_manager_blocking_duration_secs': 48.882343769073486} I0423 18:50:46.327862 139164146304832 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0423 18:50:46.327907 139164146304832 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 18:50:51.608896 139037352392448 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_20260423_155242/nnx_xpk_feat_nnx_set_defaults_true_20260423_155242_16_zarr3_true/checkpoints/9/items/array_metadatas/process_6 I0423 18:51:27.914437 139037302843136 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 37.684 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.916953802108765 s) (per-host) I0423 18:51:27.914559 139037302843136 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.609373s. I0423 18:51:36.462690 139037302843136 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 50.157490s. I0423 18:51:36.462965 139037847299840 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 18:51:36.463085 139037847299840 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0423 18:51:36.463154 139037847299840 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 18:51:36.464734 139037847299840 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 18:51:36.464912 139164146304832 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. I0423 18:51:36.465077 139164146304832 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 18:51:36.465804 139164146304832 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.323, Tokens/s/device: 13915.880, total_weights: 65536, loss: 7.176 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Thu Apr 23 18:51:47 UTC 2026 EXIT_CODE=0