XPK Start: Fri Apr 17 05:45:01 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-17 05:45:25.688325: 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) I0417 05:45:26.224859 132834474538816 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 05:45:35.265083 132834474538816 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 05:45:35.267436 132834474538816 distributed.py:172] Connecting to JAX distributed service on mt-13-scan-layers-true-tbzas-slice-job-0-0.mt-13-scan-layers-true-tbzas:8482 I0417 05:45:36.694959 132834474538816 max_utils.py:284] Jax distributed system initialized! I0417 05:45:42.882597 132834474538816 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 05:45:42.882700 132834474538816 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 05:45:42.882741 132834474538816 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0417 05:45:42.882774 132834474538816 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 05:45:43.548117 132834474538816 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 05:45:43.661061 132834474538816 checkpointing.py:688] Setting up checkpoint logger... I0417 05:45:43.661180 132834474538816 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 05:45:43.661224 132834474538816 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 05:45:43.661427 132834474538816 base_pytree_checkpoint_handler.py:441] 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 0x78cf56b182c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 05:45:46.996234 132834474538816 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 05:45:46.996473 132834474538816 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78badc74b6b0>}, handler_registry=None I0417 05:45:46.996709 132834474538816 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78badc74b6b0>` 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`. I0417 05:45:46.996757 132834474538816 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78badc73c830>` 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`. I0417 05:45:46.996793 132834474538816 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 0x78badc74b6b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78badc74b6b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78badc73c830>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78badc73c830>}). I0417 05:45:46.997138 132834474538816 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 05:45:46.997209 132834474538816 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78badc7c1ee0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 05:45:48.933502 132834474538816 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints I0417 05:45:48.949702 132834474538816 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_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78badc3ea0c0> I0417 05:45:48.949831 132834474538816 checkpointing.py:302] Checkpoint manager created! I0417 05:45:49.902307 132834474538816 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 05:45:49.902422 132834474538816 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 05:45:51.285390 132834474538816 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0417 05:45:51.285506 132834474538816 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 05:45:51.285557 132834474538816 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 05:45:51.285598 132834474538816 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0417 05:45:51.285634 132834474538816 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 05:45:51.285668 132834474538816 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 05:45:51.285704 132834474538816 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 05:45:51.285739 132834474538816 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0417 05:45:51.285772 132834474538816 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 05:45:51.285805 132834474538816 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 05:45:51.285854 132834474538816 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0417 05:45:51.285888 132834474538816 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0417 05:45:51.684317 132834474538816 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 05:45:51.684407 132834474538816 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:45:51.689622 132834474538816 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 05:45:51.689679 132834474538816 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:45:51.705895 132834474538816 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 05:45:51.705963 132834474538816 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:45:51.721528 132834474538816 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 05:45:51.721582 132834474538816 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 05:45:51.745323 132834474538816 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 05:45:51.745388 132834474538816 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:45:51.761069 132834474538816 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 05:45:51.761130 132834474538816 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:45:51.776702 132834474538816 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 05:45:51.776763 132834474538816 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:45:51.806014 132834474538816 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 05:45:51.806082 132834474538816 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 05:45:51.825590 132834474538816 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 05:45:51.825652 132834474538816 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 05:46:03.179885 132834474538816 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. I0417 05:46:03.189038 132834474538816 metric_logger.py:289] number parameters: 1.104 billion I0417 05:46:14.409891 132834474538816 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 05:46:14.561737 132834474538816 checkpointing.py:798] Waited 0.15183162689208984 seconds for step 0 to finish before starting checkpointing. I0417 05:46:14.564576 132834474538816 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 05:46:14.566325 132834474538816 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 05:46:14.567740 132834474538816 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0. I0417 05:46:15.286315 132834474538816 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 05:46:15.305392 132834474538816 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 05:46:15.305460 132834474538816 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 I0417 05:46:15.575054 132834474538816 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271076s I0417 05:46:15.575879 132834474538816 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.473 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28186535835266113 s) (per-host) I0417 05:46:15.575943 132834474538816 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281967s (batch_requests_ready=0.003061s, total_serialization_initiated=0.278175s, others=0.000730s) I0417 05:46:15.576019 132834474538816 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285874s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.285857) I0417 05:46:15.576803 132834474538816 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0. I0417 05:46:15.577126 132705653868288 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 06:06:15.577088 I0417 05:46:15.587304 132834474538816 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 05:46:15.587592 132703509112576 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 05:46:15.587753 132834474538816 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776404774.5645585, 'wait_for_prev_duration_secs': 5.9604644775390625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776404774.5663645, 'checkpointer_blocking_duration_secs': 1.0109167098999023, 'get_old_steps_start_time': 1776404775.5773072, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776404774.562159, 'checkpoint_manager_blocking_duration_secs': 1.0255537033081055} I0417 05:46:15.587873 132834474538816 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 05:46:15.587943 132834474538816 max_utils.py:750] Memstats: After params initialized: I0417 05:46:15.587996 132834474538816 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 05:46:15.588030 132834474538816 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 05:46:15.588057 132834474538816 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 05:46:15.588088 132834474538816 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 05:46:15.932138 132834474538816 metric_logger.py:185] completed step: 0, seconds: 10.906, TFLOP/s/device: 1.246, Tokens/s/device: 187.789, total_weights: 65536, loss: 10.874 I0417 05:46:15.935147 132834474538816 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/tensorboard/' I0417 05:46:15.994309 132705645475584 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0 I0417 05:46:16.357139 132834474538816 metric_logger.py:185] completed step: 1, seconds: 1.519, TFLOP/s/device: 8.946, Tokens/s/device: 1348.388, total_weights: 65536, loss: 10.874 I0417 05:46:16.498792 132834474538816 metric_logger.py:185] completed step: 2, seconds: 0.426, TFLOP/s/device: 31.901, Tokens/s/device: 4808.482, total_weights: 65536, loss: 10.262 I0417 05:46:16.646248 132834474538816 metric_logger.py:185] completed step: 3, seconds: 0.012, TFLOP/s/device: 1121.143, Tokens/s/device: 168990.841, total_weights: 65536, loss: 9.732 I0417 05:46:19.050872 132705152919296 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/items I0417 05:46:19.272967 132703494608640 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776404778784087787, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776404778784087787, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0417 05:46:20.574314 2775 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 05:46:23.157182 132705123788544 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_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0 I0417 05:46:39.555540 132834474538816 metric_logger.py:185] completed step: 4, seconds: 0.143, TFLOP/s/device: 95.084, Tokens/s/device: 14332.102, total_weights: 65536, loss: 9.274 I0417 05:46:39.564720 132834474538816 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.347, Tokens/s/device: 13919.568, total_weights: 65536, loss: 8.888 I0417 05:46:39.710815 132834474538816 metric_logger.py:185] completed step: 6, seconds: 22.907, TFLOP/s/device: 0.593, Tokens/s/device: 89.406, total_weights: 65536, loss: 8.589 I0417 05:46:39.858448 132834474538816 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 1750.693, Tokens/s/device: 263883.520, total_weights: 65536, loss: 8.380 I0417 05:46:40.005791 132834474538816 metric_logger.py:185] completed step: 8, seconds: 0.148, TFLOP/s/device: 91.854, Tokens/s/device: 13845.228, total_weights: 65536, loss: 8.251 I0417 05:46:40.154186 132834474538816 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 05:46:40.154953 132834474538816 checkpointing.py:798] Waited 0.0007791519165039062 seconds for step 9 to finish before starting checkpointing. I0417 05:46:40.157040 132834474538816 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 05:46:57.192903 132704058205952 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.631202s (commit=35.384992s, array_metadata_write=1.246210s) I0417 05:46:57.194143 132705653868288 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.699 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.9001088142395 s) (per-host) I0417 05:46:57.194282 132705653868288 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.617048s. I0417 05:46:57.886801 132705653868288 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776404778784087787, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0417 05:46:59.073896 132703494608640 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776404778784087787, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0417 05:46:59.491372 132705653868288 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000430s. I0417 05:46:59.664990 132705653868288 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 05:47:00.605779 132705653868288 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.170200s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/items I0417 05:47:00.607458 132705653868288 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0/items I0417 05:47:01.127096 132705653868288 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0 I0417 05:47:03.383633 132705653868288 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0`. I0417 05:47:03.384514 132705653868288 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.82 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/0 I0417 05:47:03.385870 132705653868288 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.808635s. I0417 05:47:03.386086 132703509112576 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 05:47:03.386212 132703509112576 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 05:47:03.386278 132703509112576 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 05:47:03.389381 132703509112576 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 05:47:03.389549 132834474538816 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0417 05:47:03.389683 132834474538816 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.232641 seconds. If this number is high, consider checkpointing less frequently. I0417 05:47:03.391892 132834474538816 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 05:47:03.393791 132834474538816 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9. I0417 05:47:03.693079 132834474538816 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 05:47:03.693173 132834474538816 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 I0417 05:47:03.721244 132834474538816 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029638s I0417 05:47:03.722063 132834474538816 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.057 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03850913047790527 s) (per-host) I0417 05:47:03.722143 132834474538816 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038608s (batch_requests_ready=0.002642s, total_serialization_initiated=0.035221s, others=0.000745s) I0417 05:47:03.722242 132834474538816 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042747s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.042732) I0417 05:47:03.722949 132834474538816 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.33 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9. I0417 05:47:03.723235 132704079177472 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 06:07:03.723200 I0417 05:47:03.729882 132834474538816 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 05:47:03.730160 132703509112576 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 05:47:03.730321 132834474538816 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776404800.1570103, 'wait_for_prev_duration_secs': 23.232641220092773, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776404823.3919296, 'checkpointer_blocking_duration_secs': 0.331449031829834, 'get_old_steps_start_time': 1776404823.7234023, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776404800.1551745, 'checkpoint_manager_blocking_duration_secs': 23.57511258125305} I0417 05:47:03.730436 132834474538816 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 05:47:03.730479 132834474538816 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 05:47:03.941266 132705653868288 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9 I0417 05:47:05.668262 132705144526592 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9/items I0417 05:47:09.848438 132704576280320 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_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0 I0417 05:47:48.519979 132704087570176 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.297962s (commit=40.112356s, array_metadata_write=1.185606s) I0417 05:47:48.521149 132704079177472 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 35.229 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.83754324913025 s) (per-host) I0417 05:47:48.521279 132704079177472 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 44.797931s. I0417 05:47:50.251544 132704079177472 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 05:47:50.864409 132704079177472 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000385s. I0417 05:47:51.333993 132704079177472 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.917559s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9/items I0417 05:47:51.335657 132704079177472 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9/items I0417 05:47:52.307772 132704079177472 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9 I0417 05:47:54.127979 132704079177472 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9`. I0417 05:47:54.128793 132704079177472 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.74 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/nnx_xpk_feat_nnx_post_train_fixes_20260417_023710_13_scan_layers_true/checkpoints/9 I0417 05:47:54.130213 132704079177472 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.406863s. I0417 05:47:54.130426 132703509112576 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 05:47:54.130540 132703509112576 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 05:47:54.130601 132703509112576 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 05:47:54.133198 132703509112576 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 05:47:54.133367 132834474538816 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0417 05:47:54.133541 132834474538816 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 05:47:54.134142 132834474538816 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.451, Tokens/s/device: 13935.196, 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: Fri Apr 17 05:48:02 UTC 2026 EXIT_CODE=0