XPK Start: Fri Apr 17 04:33:29 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 04:33:53.339592: 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 04:33:53.872444 133246996551488 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 04:34:02.911871 133246996551488 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 04:34:02.914148 133246996551488 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-rs-6rcly-slice-job-0-0.mt-16-zarr3-false-rs-6rcly:8482 I0417 04:34:04.045967 133246996551488 max_utils.py:284] Jax distributed system initialized! I0417 04:34:10.048001 133246996551488 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 04:34:10.048103 133246996551488 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 04:34:10.048145 133246996551488 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 04:34:10.048180 133246996551488 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 04:34:10.727883 133246996551488 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 04:34:10.728455 133246996551488 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 04:34:10.728639 133246996551488 checkpointing.py:688] Setting up checkpoint logger... I0417 04:34:10.728690 133246996551488 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0417 04:34:10.728734 133246996551488 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 04:34:10.729089 133246996551488 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x792f69f1c110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 04:34:12.467293 133246996551488 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 04:34:12.467520 133246996551488 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x791b046f8e00>}, handler_registry=None I0417 04:34:12.467780 133246996551488 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x791b046f8e00>` 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 04:34:12.467865 133246996551488 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x791b046f8440>` 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 04:34:12.467921 133246996551488 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 0x791b046f8e00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x791b046f8e00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x791b046f8440>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x791b046f8440>}). I0417 04:34:12.468262 133246996551488 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 04:34:12.468335 133246996551488 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x791b04468680> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 04:34:15.145578 133125497665280 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776400222785553612, 'commit_timestamp_nsecs': 1776400271083280669, '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_16_zarr3_false/checkpoints/5/_CHECKPOINT_METADATA I0417 04:34:15.402172 133125506057984 checkpoint_manager.py:1871] Missing metrics for step 0 I0417 04:34:15.597781 133246996551488 checkpoint_manager.py:1812] Found 3 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_16_zarr3_false/checkpoints I0417 04:34:16.678630 133246996551488 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x791b046f9e20> I0417 04:34:16.678792 133246996551488 checkpointing.py:302] Checkpoint manager created! I0417 04:34:17.620562 133246996551488 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 04:34:17.620664 133246996551488 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:34:18.003902 133246996551488 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 04:34:18.003994 133246996551488 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:34:18.020223 133246996551488 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 04:34:18.020281 133246996551488 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:34:18.043842 133246996551488 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 04:34:18.043913 133246996551488 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:34:18.060395 133246996551488 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 04:34:18.060456 133246996551488 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:34:18.076714 133246996551488 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 04:34:18.076781 133246996551488 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:34:18.102067 133246996551488 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 04:34:18.102134 133246996551488 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:34:18.122713 133246996551488 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 04:34:18.122788 133246996551488 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 04:34:18.335097 133246996551488 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 04:34:18.335205 133246996551488 checkpointing.py:582] restoring from this run's directory step 9 I0417 04:34:18.578971 133246996551488 event_tracking.py:70] [process=6] [sync] Started load checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9. I0417 04:34:18.828402 133246996551488 checkpointer.py:307] Restoring checkpoint from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9. I0417 04:34:18.828863 133246996551488 event_tracking.py:125] [process=6] [sync] Finished blocking load in 0.25 seconds. Continuing load @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9. I0417 04:34:20.270549 1453 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 04:34:26.553503 133246996551488 jax_array_handlers.py:843] [process=6] /jax/orbax/read/worker/io/requested throughput: 251.286 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.288735866546631 s) (per-host) I0417 04:34:26.566722 133246996551488 base_pytree_checkpoint_handler.py:130] [process=6] /jax/checkpoint/read/gbytes_per_sec: 234.879 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 6.737908363342285 s) (per-host) I0417 04:34:29.803192 133246996551488 event_tracking.py:138] [process=6] [sync] Finished load in 11.22 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints/9 I0417 04:34:29.803351 133246996551488 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776400458.5788708, 'checkpointer_duration_secs': 11.22445034980774, 'checkpoint_manager_start_time': 1776400458.3354514, 'checkpoint_manager_duration_secs': 11.46787142753601} fsdp: 32 I0417 04:34:30.212901 133246996551488 maxtext_utils.py:1796] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 04:34:30.213027 133246996551488 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 04:34:30.213080 133246996551488 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 04:34:30.213135 133246996551488 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 04:34:30.213193 133246996551488 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 04:34:30.213233 133246996551488 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 04:34:30.213285 133246996551488 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 04:34:30.213336 133246996551488 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 04:34:30.213375 133246996551488 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 04:34:30.213410 133246996551488 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 04:34:30.213455 133246996551488 maxtext_utils.py:1796] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 04:34:30.213500 133246996551488 maxtext_utils.py:1796] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 04:34:30.697011 133246996551488 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 04:34:30.697105 133246996551488 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0417 04:34:30.712932 133246996551488 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 04:34:30.712990 133246996551488 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0417 04:34:41.731517 133246996551488 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 04:34:41.732315 133246996551488 metric_logger.py:289] number parameters: 1.104 billion I0417 04:34:52.905142 133246996551488 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0417 04:34:53.060781 133246996551488 checkpointing.py:798] Waited 0.15561914443969727 seconds for step 10 to finish before starting checkpointing. I0417 04:34:53.063425 133246996551488 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 04:34:53.065224 133246996551488 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 10 I0417 04:34:53.066616 133246996551488 event_tracking.py:70] [process=6] [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_16_zarr3_false/checkpoints/10. I0417 04:34:53.830407 133246996551488 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 04:34:53.831479 133246996551488 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:34:53.831538 133246996551488 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 04:34:54.114551 133246996551488 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.284222s I0417 04:34:54.114729 133246996551488 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.355 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28806638717651367 s) (per-host) I0417 04:34:54.114783 133246996551488 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.288131s (batch_requests_ready=0.002134s, total_serialization_initiated=0.285924s, others=0.000073s) I0417 04:34:54.114892 133246996551488 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.292479s (all_items=0.000025s, per_item={'items': '0.00002503'}, temp_paths=0.292454) I0417 04:34:54.115622 133246996551488 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.05 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_16_zarr3_false/checkpoints/10. I0417 04:34:54.115986 133124918863616 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:54:54.115947 I0417 04:34:54.119630 133246996551488 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:34:54.119956 133121680852736 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:34:54.120110 133246996551488 standard_logger.py:34] {'step': 10, '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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776400493.0634062, 'wait_for_prev_duration_secs': 6.246566772460938e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776400493.0652628, 'checkpointer_blocking_duration_secs': 1.050872802734375, 'get_old_steps_start_time': 1776400494.1161585, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776400493.0613155, 'checkpoint_manager_blocking_duration_secs': 1.0587570667266846} I0417 04:34:54.120218 133246996551488 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0417 04:34:54.120265 133246996551488 max_utils.py:750] Memstats: After params initialized: I0417 04:34:54.120314 133246996551488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0417 04:34:54.120344 133246996551488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0417 04:34:54.120369 133246996551488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0417 04:34:54.120393 133246996551488 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0417 04:34:54.434999 133246996551488 metric_logger.py:185] completed step: 10, seconds: 11.173, TFLOP/s/device: 1.216, Tokens/s/device: 183.304, total_weights: 65536, loss: 6.833 I0417 04:34:54.601851 133246996551488 metric_logger.py:185] completed step: 11, seconds: 1.529, TFLOP/s/device: 8.889, Tokens/s/device: 1339.798, total_weights: 65536, loss: 6.527 I0417 04:34:54.749196 133246996551488 metric_logger.py:185] completed step: 12, seconds: 0.020, TFLOP/s/device: 695.029, Tokens/s/device: 104762.392, total_weights: 65536, loss: 6.265 I0417 04:34:54.896528 133246996551488 metric_logger.py:185] completed step: 13, seconds: 0.153, TFLOP/s/device: 88.724, Tokens/s/device: 13373.384, total_weights: 65536, loss: 6.045 I0417 04:34:54.901194 133246996551488 checkpointing.py:794] Waiting for step 15 to finish before checkpoint... I0417 04:34:55.191333 133246996551488 checkpointing.py:798] Waited 0.2901039123535156 seconds for step 15 to finish before starting checkpointing. I0417 04:34:55.194312 133246996551488 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:34:59.421968 133121689245440 array_metadata_store.py:203] [process=6][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_16_zarr3_false/checkpoints/10/items/array_metadatas/process_6 I0417 04:35:35.886440 133124918863616 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.556 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.05973410606384 s) (per-host) I0417 04:35:35.886558 133124918863616 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.770458s. I0417 04:35:44.513566 133124918863616 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.397451s. I0417 04:35:44.513883 133121680852736 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:35:44.514001 133121680852736 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:35:44.514053 133121680852736 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:35:44.515603 133121680852736 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0417 04:35:44.515766 133246996551488 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0417 04:35:44.515907 133246996551488 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.321594 seconds. If this number is high, consider checkpointing less frequently. I0417 04:35:44.517677 133246996551488 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 15 I0417 04:35:44.520009 133246996551488 event_tracking.py:70] [process=6] [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_16_zarr3_false/checkpoints/15. I0417 04:35:45.634740 133246996551488 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:35:45.634845 133246996551488 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 04:35:45.660649 133246996551488 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026929s I0417 04:35:45.660790 133246996551488 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 50.776 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0303802490234375 s) (per-host) I0417 04:35:45.660853 133246996551488 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030450s (batch_requests_ready=0.001802s, total_serialization_initiated=0.028572s, others=0.000076s) I0417 04:35:45.660939 133246996551488 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035334s (all_items=0.000014s, per_item={'items': '0.00001383'}, temp_paths=0.035320) I0417 04:35:45.661584 133246996551488 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.14 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_16_zarr3_false/checkpoints/15. I0417 04:35:45.661919 133121689245440 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:55:45.661881 I0417 04:35:45.673322 133246996551488 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:35:45.673574 133121680852736 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:35:45.673718 133246996551488 standard_logger.py:34] {'step': 15, '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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776400495.1942842, 'wait_for_prev_duration_secs': 49.32159447669983, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776400544.5177174, 'checkpointer_blocking_duration_secs': 1.1443531513214111, 'get_old_steps_start_time': 1776400545.6620948, 'get_old_steps_duration_secs': 3.218650817871094e-05, 'checkpoint_manager_blocking_start_time': 1776400495.1919742, 'checkpoint_manager_blocking_duration_secs': 50.481711626052856} I0417 04:35:45.673849 133246996551488 checkpointing.py:409] Started an asynchronous checkpoint save for step 15 I0417 04:35:45.674439 133246996551488 metric_logger.py:185] completed step: 14, seconds: 0.147, TFLOP/s/device: 92.136, Tokens/s/device: 13887.759, total_weights: 65536, loss: 5.864 I0417 04:35:45.680361 133246996551488 metric_logger.py:185] completed step: 15, seconds: 0.147, TFLOP/s/device: 92.295, Tokens/s/device: 13911.720, total_weights: 65536, loss: 5.719 I0417 04:35:54.601654 133118979712768 array_metadata_store.py:203] [process=6][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_16_zarr3_false/checkpoints/15/items/array_metadatas/process_6 I0417 04:36:07.544856 133246996551488 metric_logger.py:185] completed step: 16, seconds: 50.778, TFLOP/s/device: 0.268, Tokens/s/device: 40.333, total_weights: 65536, loss: 5.605 I0417 04:36:07.692296 133246996551488 metric_logger.py:185] completed step: 17, seconds: 21.719, TFLOP/s/device: 0.626, Tokens/s/device: 94.297, total_weights: 65536, loss: 5.517 I0417 04:36:07.839568 133246996551488 metric_logger.py:185] completed step: 18, seconds: 0.152, TFLOP/s/device: 89.496, Tokens/s/device: 13489.748, total_weights: 65536, loss: 5.448 I0417 04:36:07.986551 133246996551488 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0417 04:36:07.987275 133246996551488 checkpointing.py:798] Waited 0.0007457733154296875 seconds for step 19 to finish before starting checkpointing. I0417 04:36:07.989396 133246996551488 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:36:29.636364 133121689245440 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 35.895 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.00593137741089 s) (per-host) I0417 04:36:29.636471 133121689245440 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.974432s. I0417 04:36:39.424992 133121689245440 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 53.762937s. I0417 04:36:39.425284 133121680852736 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:36:39.425408 133121680852736 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:36:39.425457 133121680852736 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:36:39.430113 133121680852736 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0417 04:36:39.430295 133246996551488 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0417 04:36:39.430422 133246996551488 checkpoint_manager.py:1452] Waiting for previous save to complete took 31.441030 seconds. If this number is high, consider checkpointing less frequently. I0417 04:36:39.431844 133246996551488 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 19 I0417 04:36:39.433887 133246996551488 event_tracking.py:70] [process=6] [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_16_zarr3_false/checkpoints/19. I0417 04:36:40.150915 133246996551488 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:36:40.151010 133246996551488 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 04:36:40.176456 133246996551488 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.026448s I0417 04:36:40.176633 133246996551488 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 51.542 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.02992868423461914 s) (per-host) I0417 04:36:40.176685 133246996551488 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.029991s (batch_requests_ready=0.001667s, total_serialization_initiated=0.028257s, others=0.000067s) I0417 04:36:40.176797 133246996551488 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.033969s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.033958) I0417 04:36:40.177613 133246996551488 event_tracking.py:125] [process=6] [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_16_zarr3_false/checkpoints/19. I0417 04:36:40.177957 133121680852736 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:56:40.177919 I0417 04:36:40.182047 133246996551488 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:36:40.182322 133124918863616 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:36:40.182460 133246996551488 standard_logger.py:34] {'step': 19, '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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776400567.9893677, 'wait_for_prev_duration_secs': 31.441030263900757, 'time_between_consecutive_saves_sec': 23.473724126815796, 'checkpointer_blocking_start_time': 1776400599.431884, 'checkpointer_blocking_duration_secs': 0.7462174892425537, 'get_old_steps_start_time': 1776400600.1781204, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776400567.987506, 'checkpoint_manager_blocking_duration_secs': 32.194920778274536} I0417 04:36:40.182580 133246996551488 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0417 04:36:40.182629 133246996551488 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:36:47.027248 133118988105472 array_metadata_store.py:203] [process=6][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_16_zarr3_false/checkpoints/19/items/array_metadatas/process_6 I0417 04:37:22.244906 133121680852736 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.522 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.09816002845764 s) (per-host) I0417 04:37:22.245033 133121680852736 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.066962s. I0417 04:37:30.735922 133121680852736 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.557834s. I0417 04:37:30.736220 133124918863616 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:37:30.736332 133124918863616 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:37:30.736372 133124918863616 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:37:30.737858 133124918863616 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0417 04:37:30.738029 133246996551488 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0417 04:37:30.738174 133246996551488 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 04:37:30.738924 133246996551488 metric_logger.py:185] completed step: 19, seconds: 0.147, TFLOP/s/device: 92.149, Tokens/s/device: 13889.642, total_weights: 65536, loss: 5.391 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Fri Apr 17 04:37:40 UTC 2026 EXIT_CODE=0