XPK Start: Fri Apr 17 01:40:43 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 01:41:07.027220: 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 01:41:07.561547 133603099961152 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 01:41:16.601885 133603099961152 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 01:41:16.604188 133603099961152 distributed.py:172] Connecting to JAX distributed service on mt-13-scan-layers-true-lpspk-slice-job-0-0.mt-13-scan-layers-true-lpspk:8482 I0417 01:41:17.634018 133603099961152 max_utils.py:284] Jax distributed system initialized! I0417 01:41:23.745344 133603099961152 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 01:41:23.745446 133603099961152 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 01:41:23.745486 133603099961152 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 01:41:23.745520 133603099961152 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 01:41:24.411550 133603099961152 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 01:41:24.526376 133603099961152 checkpointing.py:688] Setting up checkpoint logger... I0417 01:41:24.526494 133603099961152 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 01:41:24.526539 133603099961152 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 01:41:24.526744 133603099961152 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 0x79822d84bb90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 01:41:28.148330 133603099961152 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 01:41:28.148562 133603099961152 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 0x796cc03a5130>}, handler_registry=None I0417 01:41:28.148797 133603099961152 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796cc03a5130>` 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 01:41:28.148861 133603099961152 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x796cc03a4e60>` 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 01:41:28.148899 133603099961152 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 0x796cc03a5130>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796cc03a5130>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x796cc03a4e60>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x796cc03a4e60>}). I0417 01:41:28.149231 133603099961152 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 01:41:28.149302 133603099961152 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x796cc078dee0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 01:41:30.010522 133603099961152 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints I0417 01:41:30.560003 133603099961152 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_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x796cc03e0530> I0417 01:41:30.560173 133603099961152 checkpointing.py:302] Checkpoint manager created! I0417 01:41:31.509291 133603099961152 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 01:41:31.509412 133603099961152 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 01:41:32.932109 133603099961152 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0417 01:41:32.932221 133603099961152 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 01:41:32.932273 133603099961152 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0417 01:41:32.932315 133603099961152 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0417 01:41:32.932352 133603099961152 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 01:41:32.932387 133603099961152 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0417 01:41:32.932423 133603099961152 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:41:32.932457 133603099961152 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0417 01:41:32.932491 133603099961152 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:41:32.932525 133603099961152 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0417 01:41:32.932558 133603099961152 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0417 01:41:32.932590 133603099961152 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0417 01:41:33.334740 133603099961152 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 01:41:33.334852 133603099961152 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:41:33.340249 133603099961152 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 01:41:33.340303 133603099961152 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:41:33.356469 133603099961152 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 01:41:33.356528 133603099961152 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:41:33.372207 133603099961152 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 01:41:33.372267 133603099961152 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 01:41:33.396238 133603099961152 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 01:41:33.396309 133603099961152 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:41:33.411757 133603099961152 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 01:41:33.411816 133603099961152 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:41:33.427318 133603099961152 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 01:41:33.427375 133603099961152 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:41:33.456946 133603099961152 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 01:41:33.457015 133603099961152 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 01:41:33.476552 133603099961152 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 01:41:33.476621 133603099961152 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 01:41:44.391027 133603099961152 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 01:41:44.400312 133603099961152 metric_logger.py:289] number parameters: 1.104 billion I0417 01:41:55.901243 133603099961152 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 01:41:56.053576 133603099961152 checkpointing.py:798] Waited 0.15231871604919434 seconds for step 0 to finish before starting checkpointing. I0417 01:41:56.055992 133603099961152 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 01:41:56.057665 133603099961152 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0417 01:41:56.059190 133603099961152 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0. I0417 01:41:57.239425 133603099961152 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 01:41:57.259444 133603099961152 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 01:41:57.259514 133603099961152 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 01:41:57.534229 133603099961152 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276407s I0417 01:41:57.535081 133603099961152 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.368 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2873423099517822 s) (per-host) I0417 01:41:57.535154 133603099961152 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.287435s (batch_requests_ready=0.003048s, total_serialization_initiated=0.283615s, others=0.000771s) I0417 01:41:57.535239 133603099961152 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.291301s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.291282) I0417 01:41:57.535996 133603099961152 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.48 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0. I0417 01:41:57.536346 133473367049984 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 02:01:57.536310 I0417 01:41:57.547980 133603099961152 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 01:41:57.548257 133472295302912 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 01:41:57.548388 133603099961152 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776390116.0559726, 'wait_for_prev_duration_secs': 6.0558319091796875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776390116.0577025, 'checkpointer_blocking_duration_secs': 1.4787933826446533, 'get_old_steps_start_time': 1776390117.5365229, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776390116.0539904, 'checkpoint_manager_blocking_duration_secs': 1.4943573474884033} I0417 01:41:57.548496 133603099961152 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 01:41:57.548561 133603099961152 max_utils.py:750] Memstats: After params initialized: I0417 01:41:57.548611 133603099961152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0417 01:41:57.548642 133603099961152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0417 01:41:57.548668 133603099961152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0417 01:41:57.548691 133603099961152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0417 01:41:57.893947 133603099961152 metric_logger.py:185] completed step: 0, seconds: 11.196, TFLOP/s/device: 1.214, Tokens/s/device: 182.923, total_weights: 65536, loss: 10.874 I0417 01:41:57.897053 133603099961152 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/tensorboard/' I0417 01:41:58.293134 133473289369344 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0 I0417 01:41:58.312945 133603099961152 metric_logger.py:185] completed step: 1, seconds: 1.989, TFLOP/s/device: 6.830, Tokens/s/device: 1029.534, total_weights: 65536, loss: 10.874 I0417 01:41:58.455090 133603099961152 metric_logger.py:185] completed step: 2, seconds: 0.420, TFLOP/s/device: 32.349, Tokens/s/device: 4875.970, total_weights: 65536, loss: 10.262 I0417 01:41:58.602560 133603099961152 metric_logger.py:185] completed step: 3, seconds: 0.012, TFLOP/s/device: 1141.488, Tokens/s/device: 172057.465, total_weights: 65536, loss: 9.732 I0417 01:42:00.888933 133473272583936 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/items I0417 01:42:01.051809 133467488642816 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776390120640424698, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776390120640424698, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0417 01:42:03.294198 2814 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 01:42:06.741388 133472349845248 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_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0 I0417 01:42:23.188938 133603099961152 metric_logger.py:185] completed step: 4, seconds: 0.144, TFLOP/s/device: 94.551, Tokens/s/device: 14251.715, total_weights: 65536, loss: 9.274 I0417 01:42:23.199154 133603099961152 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.655, Tokens/s/device: 13965.985, total_weights: 65536, loss: 8.888 I0417 01:42:23.343728 133603099961152 metric_logger.py:185] completed step: 6, seconds: 24.585, TFLOP/s/device: 0.553, Tokens/s/device: 83.302, total_weights: 65536, loss: 8.589 I0417 01:42:23.491278 133603099961152 metric_logger.py:185] completed step: 7, seconds: 0.008, TFLOP/s/device: 1749.791, Tokens/s/device: 263747.585, total_weights: 65536, loss: 8.380 I0417 01:42:23.638910 133603099961152 metric_logger.py:185] completed step: 8, seconds: 0.146, TFLOP/s/device: 92.897, Tokens/s/device: 14002.461, total_weights: 65536, loss: 8.251 I0417 01:42:23.788686 133603099961152 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 01:42:23.789421 133603099961152 checkpointing.py:798] Waited 0.0007483959197998047 seconds for step 9 to finish before starting checkpointing. I0417 01:42:23.793168 133603099961152 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 01:42:39.383662 133472303695616 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.103330s (commit=34.860606s, array_metadata_write=1.242724s) I0417 01:42:39.384773 133473367049984 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.487 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.13703179359436 s) (per-host) I0417 01:42:39.384855 133473367049984 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.848392s. I0417 01:42:40.584613 133473367049984 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776390120640424698, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0417 01:42:41.342920 133473367049984 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000455s. I0417 01:42:41.356377 133467488642816 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776390120640424698, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0417 01:42:41.958206 133473367049984 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 01:42:42.902541 133473367049984 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.189025s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/items I0417 01:42:42.904250 133473367049984 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0/items I0417 01:42:43.440562 133473367049984 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0 I0417 01:42:45.280798 133473367049984 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_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0`. I0417 01:42:45.281678 133473367049984 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 49.22 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/0 I0417 01:42:45.283390 133473367049984 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.746924s. I0417 01:42:45.283577 133472295302912 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 01:42:45.283699 133472295302912 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 01:42:45.283763 133472295302912 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 01:42:45.285981 133472295302912 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 01:42:45.286134 133603099961152 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 01:42:45.286231 133603099961152 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.493067 seconds. If this number is high, consider checkpointing less frequently. I0417 01:42:45.288229 133603099961152 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0417 01:42:45.290325 133603099961152 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9. I0417 01:42:45.999682 133603099961152 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0417 01:42:45.999778 133603099961152 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 01:42:46.028079 133603099961152 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.029878s I0417 01:42:46.028831 133603099961152 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 39.901 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03866004943847656 s) (per-host) I0417 01:42:46.028913 133603099961152 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038756s (batch_requests_ready=0.002613s, total_serialization_initiated=0.035429s, others=0.000714s) I0417 01:42:46.028987 133603099961152 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042832s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.042816) I0417 01:42:46.029692 133603099961152 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9. I0417 01:42:46.029971 133472295302912 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 02:02:46.029948 I0417 01:42:46.036242 133603099961152 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 01:42:46.036594 133472316274432 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 01:42:46.036788 133603099961152 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776390143.793134, 'wait_for_prev_duration_secs': 21.493067026138306, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776390165.2882683, 'checkpointer_blocking_duration_secs': 0.7417991161346436, 'get_old_steps_start_time': 1776390166.0300932, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776390143.789658, 'checkpoint_manager_blocking_duration_secs': 22.247085094451904} I0417 01:42:46.036942 133603099961152 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 01:42:46.036989 133603099961152 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 01:42:46.236050 133473367049984 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9 I0417 01:42:48.398641 133473247405824 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9/items I0417 01:42:52.545451 133472349845248 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_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0 I0417 01:43:31.766304 133472324667136 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.831463s (commit=40.228163s, array_metadata_write=1.603300s) I0417 01:43:31.767406 133472295302912 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.506 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.7772331237793 s) (per-host) I0417 01:43:31.767473 133472295302912 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.737426s. I0417 01:43:34.148806 133472295302912 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000389s. I0417 01:43:34.342264 133472295302912 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0417 01:43:35.267255 133472295302912 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.189722s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9/items I0417 01:43:35.268875 133472295302912 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9/items I0417 01:43:35.797363 133472295302912 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9 I0417 01:43:38.517613 133472295302912 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_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9`. I0417 01:43:38.518511 133472295302912 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.23 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260417_003418/nnx_xpk_feat_nnx_post_train_fixes_20260417_003418_13_scan_layers_true/checkpoints/9 I0417 01:43:38.519915 133472295302912 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.489865s. I0417 01:43:38.520120 133472316274432 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 01:43:38.520242 133472316274432 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 01:43:38.520305 133472316274432 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 01:43:38.522575 133472316274432 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 01:43:38.522765 133603099961152 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 01:43:38.522972 133603099961152 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 01:43:38.523598 133603099961152 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.201, Tokens/s/device: 13897.465, 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 01:43:48 UTC 2026 EXIT_CODE=0