XPK Start: Fri Apr 17 02:46:10 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 02:46:33.789351: 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 02:46:33.966781 132685827929920 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 02:46:43.005475 132685827929920 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 02:46:43.007771 132685827929920 distributed.py:172] Connecting to JAX distributed service on mt-03-dropout-i3hgp-slice-job-0-0.mt-03-dropout-i3hgp:8482 I0417 02:46:44.470717 132685827929920 max_utils.py:284] Jax distributed system initialized! I0417 02:46:50.549579 132685827929920 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 02:46:50.549679 132685827929920 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 02:46:50.549718 132685827929920 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 02:46:50.549750 132685827929920 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 02:46:51.570583 132685827929920 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 02:46:51.571209 132685827929920 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 02:46:51.571410 132685827929920 checkpointing.py:688] Setting up checkpoint logger... I0417 02:46:51.571475 132685827929920 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 02:46:51.571539 132685827929920 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 02:46:51.572008 132685827929920 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 0x78acc4b18f20>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 02:46:55.306802 132685827929920 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 02:46:55.307086 132685827929920 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 0x78ab90129c40>}, handler_registry=None I0417 02:46:55.307325 132685827929920 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78ab90129c40>` 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 02:46:55.307377 132685827929920 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78ab9014a450>` 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 02:46:55.307414 132685827929920 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 0x78ab90129c40>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x78ab90129c40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78ab9014a450>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x78ab9014a450>}). I0417 02:46:55.307755 132685827929920 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 02:46:55.307824 132685827929920 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x78ab902c4ae0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 02:46:56.845223 132685827929920 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints I0417 02:46:57.316795 132685827929920 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/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x78ab90129940> I0417 02:46:57.317009 132685827929920 checkpointing.py:302] Checkpoint manager created! I0417 02:46:57.651587 132685827929920 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 02:46:57.651688 132685827929920 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0417 02:46:58.028207 132685827929920 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 02:46:58.028299 132685827929920 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0417 02:46:58.044331 132685827929920 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,128,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 02:46:58.044388 132685827929920 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,128,2048]....................................... ('fsdp', None, None). I0417 02:46:58.067751 132685827929920 attentions.py:1154] attentions/query Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 02:46:58.067816 132685827929920 attentions.py:1154] attentions/query Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0417 02:46:58.083954 132685827929920 attentions.py:1155] attentions/key Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 02:46:58.084012 132685827929920 attentions.py:1155] attentions/key Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0417 02:46:58.100206 132685827929920 attentions.py:1156] attentions/value Logical: bfloat16[32,128,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 02:46:58.100265 132685827929920 attentions.py:1156] attentions/value Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0417 02:46:58.125442 132685827929920 attentions.py:1197] attentions/out Logical: bfloat16[32,128,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 02:46:58.125507 132685827929920 attentions.py:1197] attentions/out Physical: bfloat16[32,128,16,128]..................................... ('fsdp', None, None, None). I0417 02:46:58.150730 132685827929920 linears.py:525] linears/x Logical: bfloat16[32,128,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 02:46:58.150794 132685827929920 linears.py:525] linears/x Physical: bfloat16[32,128,7168]....................................... ('fsdp', None, None). I0417 02:46:58.375380 132685827929920 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 02:46:58.375485 132685827929920 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 02:46:59.804301 132685827929920 maxtext_utils.py:1796] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 02:46:59.804433 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 02:46:59.804487 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 02:46:59.804546 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0417 02:46:59.804598 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 02:46:59.804637 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 02:46:59.804688 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0417 02:46:59.804740 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0417 02:46:59.804781 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0417 02:46:59.804818 132685827929920 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0417 02:46:59.804888 132685827929920 maxtext_utils.py:1796] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 02:46:59.804947 132685827929920 maxtext_utils.py:1796] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 02:47:00.325884 132685827929920 train.py:157] train/xent Logical: float32[32,128]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 02:47:00.325972 132685827929920 train.py:157] train/xent Physical: float32[32,128]............................................. ('fsdp', None). I0417 02:47:00.341491 132685827929920 train.py:164] train/z_loss Logical: float32[32,128]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 02:47:00.341554 132685827929920 train.py:164] train/z_loss Physical: float32[32,128]............................................. ('fsdp', None). I0417 02:47:03.875602 132685827929920 max_utils.py:791] Total memory size: 0.8 GB, Output size: 0.4 GB, Temp size: 0.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0417 02:47:03.883031 132685827929920 metric_logger.py:289] number parameters: 1.104 billion I0417 02:47:08.229019 132685827929920 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 02:47:08.301597 132685827929920 checkpointing.py:798] Waited 0.07256960868835449 seconds for step 0 to finish before starting checkpointing. I0417 02:47:08.303975 132685827929920 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 02:47:08.306183 132685827929920 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 02:47:08.307750 132685827929920 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0. I0417 02:47:09.010176 132685827929920 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 02:47:09.024030 132685827929920 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 02:47:09.024097 132685827929920 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 02:47:09.273120 132557639243520 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0 I0417 02:47:09.301493 132685827929920 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.278204s I0417 02:47:09.302308 132685827929920 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.427 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2842535972595215 s) (per-host) I0417 02:47:09.302371 132685827929920 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.284333s (batch_requests_ready=0.002016s, total_serialization_initiated=0.281590s, others=0.000727s) I0417 02:47:09.302444 132685827929920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.288172s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.288156) I0417 02:47:09.303207 132685827929920 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.00 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0. I0417 02:47:09.303546 132557647636224 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 03:07:09.303508 I0417 02:47:09.742985 132685827929920 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 02:47:09.743362 132557081401088 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 02:47:09.743510 132685827929920 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776394028.3039556, 'wait_for_prev_duration_secs': 6.413459777832031e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776394028.306222, 'checkpointer_blocking_duration_secs': 0.9974765777587891, 'get_old_steps_start_time': 1776394029.3037233, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776394028.3020685, 'checkpoint_manager_blocking_duration_secs': 1.441399097442627} I0417 02:47:09.743666 132685827929920 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 02:47:09.743719 132685827929920 max_utils.py:750] Memstats: After params initialized: I0417 02:47:09.743776 132685827929920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_0(process=0,(0,0,0,0)) I0417 02:47:09.743809 132685827929920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_1(process=0,(1,0,0,0)) I0417 02:47:09.743846 132685827929920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_4(process=0,(0,1,0,0)) I0417 02:47:09.743873 132685827929920 max_utils.py:756] Using (GB) 0.41 / 31.25 (1.312000%) on TPU_5(process=0,(1,1,0,0)) I0417 02:47:10.078363 132685827929920 metric_logger.py:185] completed step: 0, seconds: 4.151, TFLOP/s/device: 0.193, Tokens/s/device: 30.835, total_weights: 4096, loss: 10.889 I0417 02:47:10.081367 132685827929920 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/tensorboard/' I0417 02:47:10.477351 132685827929920 metric_logger.py:185] completed step: 1, seconds: 1.848, TFLOP/s/device: 0.433, Tokens/s/device: 69.269, total_weights: 4096, loss: 10.902 I0417 02:47:10.545967 132685827929920 metric_logger.py:185] completed step: 2, seconds: 0.400, TFLOP/s/device: 2.004, Tokens/s/device: 320.332, total_weights: 4096, loss: 9.956 I0417 02:47:10.616986 132685827929920 metric_logger.py:185] completed step: 3, seconds: 0.008, TFLOP/s/device: 95.240, Tokens/s/device: 15221.786, total_weights: 4096, loss: 9.139 I0417 02:47:11.895811 132557622458112 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/items I0417 02:47:12.057512 132554395854592 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776394031646959388, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776394031646959388, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0417 02:47:13.449711 2791 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 02:47:15.628413 132557597816576 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/items/array_metadatas/process_0 I0417 02:47:20.989549 132685827929920 metric_logger.py:185] completed step: 4, seconds: 0.070, TFLOP/s/device: 11.517, Tokens/s/device: 1840.667, total_weights: 4096, loss: 8.459 I0417 02:47:20.997319 132685827929920 metric_logger.py:185] completed step: 5, seconds: 0.071, TFLOP/s/device: 11.260, Tokens/s/device: 1799.699, total_weights: 4096, loss: 7.924 I0417 02:47:21.066231 132685827929920 metric_logger.py:185] completed step: 6, seconds: 10.371, TFLOP/s/device: 0.077, Tokens/s/device: 12.342, total_weights: 4096, loss: 7.553 I0417 02:47:21.138317 132685827929920 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 141.373, Tokens/s/device: 22594.881, total_weights: 4096, loss: 7.267 I0417 02:47:21.209790 132685827929920 metric_logger.py:185] completed step: 8, seconds: 0.070, TFLOP/s/device: 11.421, Tokens/s/device: 1825.364, total_weights: 4096, loss: 7.111 I0417 02:47:21.281460 132685827929920 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 02:47:21.282182 132685827929920 checkpointing.py:798] Waited 0.0007398128509521484 seconds for step 9 to finish before starting checkpointing. I0417 02:47:21.284172 132685827929920 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 02:47:45.376078 132557089793792 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 31.927853s (commit=30.715330s, array_metadata_write=1.212523s) I0417 02:47:45.377193 132557647636224 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 43.444 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.35911297798157 s) (per-host) I0417 02:47:45.377306 132557647636224 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.073646s. I0417 02:47:46.937460 132557647636224 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776394031646959388, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0417 02:47:47.672039 132557647636224 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000245s. I0417 02:47:47.693514 132554395854592 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776394031646959388, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0417 02:47:48.307807 132557647636224 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 02:47:49.260698 132557647636224 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.195987s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/items I0417 02:47:49.262467 132557647636224 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0/items I0417 02:47:49.827650 132557647636224 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0 I0417 02:47:51.704702 132557647636224 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0`. I0417 02:47:51.705615 132557647636224 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 43.40 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/0 I0417 02:47:51.707155 132557647636224 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 42.403492s. I0417 02:47:51.707276 132557081401088 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 02:47:51.707345 132557081401088 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 02:47:51.707412 132557081401088 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 02:47:51.709353 132557081401088 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 02:47:51.709476 132685827929920 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 02:47:51.709568 132685827929920 checkpoint_manager.py:1452] Waiting for previous save to complete took 30.425398 seconds. If this number is high, consider checkpointing less frequently. I0417 02:47:51.711475 132685827929920 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 02:47:51.713615 132685827929920 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9. I0417 02:47:52.432494 132685827929920 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 02:47:52.432589 132685827929920 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 02:47:52.464999 132685827929920 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.033282s I0417 02:47:52.465726 132685827929920 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 41.001 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.037622690200805664 s) (per-host) I0417 02:47:52.465796 132685827929920 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037709s (batch_requests_ready=0.001782s, total_serialization_initiated=0.035255s, others=0.000672s) I0417 02:47:52.465875 132685827929920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041803s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.041785) I0417 02:47:52.466506 132685827929920 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9. I0417 02:47:52.466854 132557597816576 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 03:07:52.466802 I0417 02:47:52.476417 132685827929920 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 02:47:52.476680 132557081401088 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 02:47:52.476849 132685827929920 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776394041.284141, 'wait_for_prev_duration_secs': 30.425397872924805, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776394071.7115152, 'checkpointer_blocking_duration_secs': 0.7554843425750732, 'get_old_steps_start_time': 1776394072.467026, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776394041.2824109, 'checkpoint_manager_blocking_duration_secs': 31.1943941116333} I0417 02:47:52.476968 132685827929920 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 02:47:52.477015 132685827929920 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 02:47:52.668481 132557647636224 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9 I0417 02:47:54.413351 132557119158016 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9/items I0417 02:47:57.269300 132557622458112 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9/items/array_metadatas/process_0 I0417 02:48:38.593992 132557614065408 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.066341s (commit=41.449283s, array_metadata_write=1.617058s) I0417 02:48:38.595235 132557597816576 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.215 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.167086601257324 s) (per-host) I0417 02:48:38.595365 132557597816576 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.128392s. I0417 02:48:40.749796 132557597816576 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 02:48:40.904858 132557597816576 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000212s. I0417 02:48:41.825268 132557597816576 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.906773s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9/items I0417 02:48:41.827734 132557597816576 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9/items I0417 02:48:42.383246 132557597816576 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9 I0417 02:48:43.785367 132557597816576 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9`. I0417 02:48:43.786200 132557597816576 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.07 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_03_dropout/checkpoints/9 I0417 02:48:43.787646 132557597816576 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.320671s. I0417 02:48:43.787879 132557081401088 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 02:48:43.788001 132557081401088 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 02:48:43.788063 132557081401088 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 02:48:43.790626 132557081401088 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 02:48:43.790823 132685827929920 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 02:48:43.790980 132685827929920 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 02:48:43.791681 132685827929920 metric_logger.py:185] completed step: 9, seconds: 0.072, TFLOP/s/device: 11.089, Tokens/s/device: 1772.313, total_weights: 4096, loss: 6.981 Per train step: Total TFLOPs: 0.80 split as 99.60% learnable weight flops and 0.40% attention flops XPK End: Fri Apr 17 02:48:56 UTC 2026 EXIT_CODE=0