XPK Start: Sun Apr 19 05:11:26 UTC 2026 2026-04-19 05:11:30.263929: 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 00:00:1776575490.276661 10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776575490.280345 10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776575490.291563 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776575490.291581 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776575490.291583 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776575490.291585 10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-19 05:11:49.434186: 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 05:11:49.960494 133254020331328 max_utils.py:273] Attempting to initialize the jax distributed system... INFO:2026-04-19 05:11:59,000:jax._src.distributed:140: Starting JAX distributed service on [::]:8482 I0419 05:11:59.000690 133254020331328 distributed.py:140] Starting JAX distributed service on [::]:8482 INFO:2026-04-19 05:11:59,003:jax._src.distributed:157: Connecting to JAX distributed service on mt-07-eval-a8g3v-slice-job-0-0.mt-07-eval-a8g3v:8482 I0419 05:11:59.003036 133254020331328 distributed.py:157] Connecting to JAX distributed service on mt-07-eval-a8g3v-slice-job-0-0.mt-07-eval-a8g3v:8482 I0419 05:12:00.975685 133254020331328 max_utils.py:284] Jax distributed system initialized! I0419 05:12:08.021232 133254020331328 max_utils.py:800] System Information: Jax Version: 0.8.1 I0419 05:12:08.021336 133254020331328 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0419 05:12:08.021382 133254020331328 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 05:12:08.021427 133254020331328 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 05:12:08.632463 133254020331328 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 05:12:08.746946 133254020331328 checkpointing.py:688] Setting up checkpoint logger... I0419 05:12:08.747064 133254020331328 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 05:12:08.747112 133254020331328 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None I0419 05:12:08.747326 133254020331328 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 0x79311cb54b30>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 05:12:11.877518 133254020331328 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 05:12:11.877881 133254020331328 checkpoint_manager.py:709] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x791d8fe29a90>}, handler_registry=None I0419 05:12:11.878123 133254020331328 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x791d8fe29a90>` 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 05:12:11.878174 133254020331328 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x791d8e746840>` 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 05:12:11.878210 133254020331328 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 0x791d8fe29a90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x791d8fe29a90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x791d8e746840>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x791d8e746840>}). I0419 05:12:11.878533 133254020331328 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33 I0419 05:12:11.878612 133254020331328 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x791d8d280f40> timeout: 600 secs and primary_host=0 for async checkpoint writes I0419 05:12:13.616166 133254020331328 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_07_eval/checkpoints I0419 05:12:14.481699 133254020331328 checkpoint_manager.py:929] [process=0][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_07_eval/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x791d8e745640> I0419 05:12:14.481895 133254020331328 checkpointing.py:302] Checkpoint manager created! I0419 05:12:15.396314 133254020331328 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 05:12:15.396439 133254020331328 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 05:12:16.795139 133254020331328 maxtext_utils.py:1749] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0419 05:12:16.795261 133254020331328 maxtext_utils.py:1749] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 05:12:16.795314 133254020331328 maxtext_utils.py:1749] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0419 05:12:16.795358 133254020331328 maxtext_utils.py:1749] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0419 05:12:16.795394 133254020331328 maxtext_utils.py:1749] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 05:12:16.795430 133254020331328 maxtext_utils.py:1749] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0419 05:12:16.795473 133254020331328 maxtext_utils.py:1749] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:12:16.795510 133254020331328 maxtext_utils.py:1749] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0419 05:12:16.795544 133254020331328 maxtext_utils.py:1749] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:12:16.795578 133254020331328 maxtext_utils.py:1749] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0419 05:12:16.795612 133254020331328 maxtext_utils.py:1749] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0419 05:12:16.795645 133254020331328 maxtext_utils.py:1749] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0419 05:12:17.202548 133254020331328 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 05:12:17.202695 133254020331328 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:12:17.208063 133254020331328 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 05:12:17.208178 133254020331328 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:12:17.354782 133254020331328 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 05:12:17.354932 133254020331328 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:12:17.503343 133254020331328 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 05:12:17.503488 133254020331328 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 05:12:17.664183 133254020331328 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 05:12:17.664321 133254020331328 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:12:17.807992 133254020331328 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 05:12:17.808135 133254020331328 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:12:17.950429 133254020331328 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 05:12:17.950565 133254020331328 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:12:18.101046 133254020331328 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 05:12:18.101178 133254020331328 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 05:12:18.123659 133254020331328 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 05:12:18.123754 133254020331328 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 05:12:29.121871 133254020331328 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 05:12:29.129458 133254020331328 metric_logger.py:289] number parameters: 1.104 billion I0419 05:12:40.603448 133254020331328 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 05:12:40.755389 133254020331328 checkpointing.py:798] Waited 0.15192103385925293 seconds for step 0 to finish before starting checkpointing. I0419 05:12:40.758906 133254020331328 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 05:12:40.760729 133254020331328 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 0 I0419 05:12:40.763471 133254020331328 async_checkpointer.py:452] [process=0] 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_07_eval/checkpoints/0. I0419 05:12:41.503664 133254020331328 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 05:12:42.184481 133125279553280 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0 I0419 05:12:42.200515 133254020331328 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 05:12:42.200612 133254020331328 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 05:12:42.458083 133254020331328 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.259826s I0419 05:12:42.458961 133254020331328 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.628 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.9476804733276367 s) (per-host) I0419 05:12:42.459036 133254020331328 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.947775s (batch_requests_ready=0.679519s, total_serialization_initiated=0.267462s, others=0.000794s) I0419 05:12:42.459124 133254020331328 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.951562s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.951543) I0419 05:12:42.460019 133125849995008 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0419 05:12:42.460172 133254020331328 async_checkpointer.py:561] Finished blocking save. Time taken: 1.699374s. 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_07_eval/checkpoints/0. I0419 05:12:42.870375 133254020331328 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 05:12:42.870755 133124793038592 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 05:12:42.870968 133254020331328 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_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776575560.758887, 'wait_for_prev_duration_secs': 6.437301635742188e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776575560.7607677, 'checkpointer_blocking_duration_secs': 1.6995091438293457, 'get_old_steps_start_time': 1776575562.4602969, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776575560.7557824, 'checkpoint_manager_blocking_duration_secs': 2.115133285522461} I0419 05:12:42.871080 133254020331328 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 05:12:42.871175 133254020331328 max_utils.py:750] Memstats: After params initialized: I0419 05:12:42.871254 133254020331328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0419 05:12:42.871311 133254020331328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0419 05:12:42.871365 133254020331328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0419 05:12:42.871410 133254020331328 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0419 05:12:43.169643 133254020331328 metric_logger.py:185] completed step: 0, seconds: 11.198, TFLOP/s/device: 1.213, Tokens/s/device: 182.885, total_weights: 65536, loss: 10.874 I0419 05:12:43.172536 133254020331328 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/tensorboard/' I0419 05:12:43.595544 133254020331328 metric_logger.py:185] completed step: 1, seconds: 2.563, TFLOP/s/device: 5.302, Tokens/s/device: 799.174, total_weights: 65536, loss: 10.874 I0419 05:12:43.737608 133254020331328 metric_logger.py:185] completed step: 2, seconds: 0.427, TFLOP/s/device: 31.806, Tokens/s/device: 4794.086, total_weights: 65536, loss: 10.262 I0419 05:12:43.885391 133254020331328 metric_logger.py:185] completed step: 3, seconds: 0.011, TFLOP/s/device: 1229.493, Tokens/s/device: 185322.595, total_weights: 65536, loss: 9.732 I0419 05:13:03.071211 133254020331328 metric_logger.py:185] completed step: 4, seconds: 0.144, TFLOP/s/device: 94.628, Tokens/s/device: 14263.428, total_weights: 65536, loss: 9.274 I0419 05:13:03.081194 133254020331328 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.266, Tokens/s/device: 13907.375, total_weights: 65536, loss: 8.888 I0419 05:13:03.225670 133254020331328 metric_logger.py:185] completed step: 6, seconds: 19.185, TFLOP/s/device: 0.708, Tokens/s/device: 106.750, total_weights: 65536, loss: 8.589 I0419 05:13:03.373308 133254020331328 metric_logger.py:185] completed step: 7, seconds: 0.007, TFLOP/s/device: 1851.864, Tokens/s/device: 279133.161, total_weights: 65536, loss: 8.380 I0419 05:13:03.445036 133125866780416 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0/items I0419 05:13:03.520662 133254020331328 metric_logger.py:185] completed step: 8, seconds: 0.146, TFLOP/s/device: 92.945, Tokens/s/device: 14009.741, total_weights: 65536, loss: 8.251 I0419 05:13:03.593334 133125841602304 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776575583196143627, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776575583196143627, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0/_CHECKPOINT_METADATA I0419 05:13:03.669299 133254020331328 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 05:13:03.669947 133254020331328 checkpointing.py:798] Waited 0.0006587505340576172 seconds for step 9 to finish before starting checkpointing. I0419 05:13:03.671762 133254020331328 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 05:13:05.424284 2662 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 05:13:07.060738 133130673440512 array_metadata_store.py:203] [process=0][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_07_eval/checkpoints/0/items/array_metadatas/process_0 I0419 05:13:42.575683 133130665047808 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.157434s (commit=36.103563s, array_metadata_write=1.053871s) I0419 05:13:42.576843 133125849995008 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 25.867 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 61.06552028656006 s) (per-host) I0419 05:13:42.576960 133125849995008 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 60.116832s. I0419 05:13:43.637539 133125849995008 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776575583196143627, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0/_CHECKPOINT_METADATA I0419 05:13:44.787752 133125841602304 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776575583196143627, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0/_CHECKPOINT_METADATA I0419 05:13:45.011642 133125849995008 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 05:13:45.232008 133125849995008 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000448s. I0419 05:13:45.947060 133125849995008 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.188718s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0/items I0419 05:13:45.948999 133125849995008 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0/items I0419 05:13:46.778672 133125849995008 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0 I0419 05:13:48.055966 133125849995008 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0`. I0419 05:13:48.056770 133125849995008 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 67.295977s. directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/0 I0419 05:13:48.058223 133125849995008 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 65.598094s. I0419 05:13:48.058487 133124793038592 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 05:13:48.058628 133124793038592 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 05:13:48.058694 133124793038592 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 05:13:48.060623 133124793038592 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 05:13:48.060764 133254020331328 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 05:13:48.060882 133254020331328 checkpoint_manager.py:1458] Waiting for previous save to complete took 44.389129 seconds. If this number is high, consider checkpointing less frequently. I0419 05:13:48.062891 133254020331328 checkpoint_manager.py:1518] [process=0] Saving checkpoint at step 9 I0419 05:13:48.066389 133254020331328 async_checkpointer.py:452] [process=0] 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_07_eval/checkpoints/9. I0419 05:13:49.444668 133125849995008 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9 I0419 05:13:49.459789 133254020331328 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array. I0419 05:13:49.459904 133254020331328 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 05:13:49.489726 133254020331328 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.031974s I0419 05:13:49.490541 133254020331328 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.327 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.289595365524292 s) (per-host) I0419 05:13:49.490614 133254020331328 base_pytree_checkpoint_handler.py:737] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.289684s (batch_requests_ready=0.251217s, total_serialization_initiated=0.037722s, others=0.000744s) I0419 05:13:49.490692 133254020331328 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.293687s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.293670) I0419 05:13:49.491557 133125321516800 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0419 05:13:49.491708 133254020331328 async_checkpointer.py:561] Finished blocking save. Time taken: 1.428747s. 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_07_eval/checkpoints/9. I0419 05:13:49.932801 133254020331328 checkpoint_manager.py:1566] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 05:13:49.933258 133124793038592 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 05:13:49.933441 133254020331328 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_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776575583.6717198, 'wait_for_prev_duration_secs': 44.38912916183472, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776575628.0629313, 'checkpointer_blocking_duration_secs': 1.4289073944091797, 'get_old_steps_start_time': 1776575629.4918594, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776575583.6701858, 'checkpoint_manager_blocking_duration_secs': 46.26321887969971} I0419 05:13:49.933552 133254020331328 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 05:13:49.933594 133254020331328 checkpoint_manager.py:2024] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 05:13:51.611603 133130673440512 atomicity.py:137] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9/items I0419 05:13:55.666841 133125866780416 array_metadata_store.py:203] [process=0][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_07_eval/checkpoints/9/items/array_metadatas/process_0 I0419 05:14:35.644932 133125329909504 base_pytree_checkpoint_handler.py:1232] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.481543s (commit=41.402672s, array_metadata_write=1.078872s) I0419 05:14:35.646128 133125321516800 base_pytree_checkpoint_handler.py:129] [process=0] /jax/checkpoint/write/gbytes_per_sec: 34.010 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.44515061378479 s) (per-host) I0419 05:14:35.646253 133125321516800 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.154585s. I0419 05:14:37.987300 133125321516800 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000412s. I0419 05:14:38.190912 133125321516800 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 05:14:39.115673 133125321516800 base_pytree_checkpoint_handler.py:1356] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.165645s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9/items I0419 05:14:39.117545 133125321516800 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9/items I0419 05:14:39.563910 133125321516800 atomicity.py:608] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9 I0419 05:14:41.266958 133125321516800 atomicity.py:790] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9`. I0419 05:14:41.267740 133125321516800 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 53.204782s. directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260419_025828/nnx_xpk_feat_nnx_set_defaults_true_20260419_025828_07_eval/checkpoints/9 I0419 05:14:41.269148 133125321516800 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 51.777481s. I0419 05:14:41.269367 133124793038592 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 05:14:41.269486 133124793038592 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 05:14:41.269550 133124793038592 checkpoint_manager.py:2133] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 05:14:41.271093 133124793038592 checkpoint_manager.py:2142] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 05:14:41.271285 133254020331328 checkpoint_manager.py:2036] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 05:14:41.271489 133254020331328 checkpoint_manager.py:2013] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 05:14:41.272258 133254020331328 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.239, Tokens/s/device: 13903.315, 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 05:14:52 UTC 2026 EXIT_CODE=0