XPK Start: Fri Apr 17 13:26:34 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 13:26:58.301012: 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 13:26:58.477728 134259751716672 max_utils.py:273] Attempting to initialize the jax distributed system...
I0417 13:27:07.516660 134259751716672 distributed.py:149] Starting JAX distributed service on [::]:8482
I0417 13:27:07.518916 134259751716672 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-fvt38-slice-job-0-0.mt-06-grad-accum-fvt38:8482
I0417 13:27:09.421789 134259751716672 max_utils.py:284] Jax distributed system initialized!
I0417 13:27:15.821562 134259751716672 max_utils.py:800] System Information: Jax Version: 0.9.2
I0417 13:27:15.821661 134259751716672 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0417 13:27:15.821699 134259751716672 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 13:27:15.821734 134259751716672 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0417 13:27:16.847316 134259751716672 maxtext_utils.py:1548] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0417 13:27:16.847604 134259751716672 checkpointing.py:677] Setting up checkpoint logger...
I0417 13:27:16.847670 134259751716672 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0417 13:27:16.847723 134259751716672 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0417 13:27:16.848149 134259751716672 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 0x7a1b2871c110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0417 13:27:24.640459 134259751716672 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0417 13:27:24.640796 134259751716672 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 0x7a1a05673ad0>}, handler_registry=None
I0417 13:27:24.641052 134259751716672 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1a05673ad0>` 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 13:27:24.641101 134259751716672 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1a056bf350>` 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 13:27:24.641134 134259751716672 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 0x7a1a05673ad0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a1a05673ad0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1a056bf350>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a1a056bf350>}).
I0417 13:27:24.641443 134259751716672 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0417 13:27:24.641511 134259751716672 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a061453c180> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0417 13:27:26.767868 134259751716672 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints
I0417 13:27:27.457375 134259751716672 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_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a1a0548ecc0>
I0417 13:27:27.457546 134259751716672 checkpointing.py:301] Checkpoint manager created!
I0417 13:27:29.140869 134259751716672 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0417 13:27:29.140982 134259751716672 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 13:27:29.519831 134259751716672 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 13:27:29.519930 134259751716672 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 13:27:29.536227 134259751716672 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0417 13:27:29.536288 134259751716672 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0417 13:27:29.559919 134259751716672 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 13:27:29.559985 134259751716672 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:27:29.576541 134259751716672 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 13:27:29.576606 134259751716672 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:27:29.592947 134259751716672 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 13:27:29.593007 134259751716672 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:27:29.618402 134259751716672 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0417 13:27:29.618469 134259751716672 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0417 13:27:29.639494 134259751716672 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0417 13:27:29.639568 134259751716672 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0417 13:27:29.863205 134259751716672 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0417 13:27:29.863314 134259751716672 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0417 13:27:31.276186 134259751716672 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0417 13:27:31.276311 134259751716672 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 13:27:31.276363 134259751716672 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0417 13:27:31.276419 134259751716672 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0417 13:27:31.276471 134259751716672 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 13:27:31.276509 134259751716672 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0417 13:27:31.276566 134259751716672 maxtext_utils.py:1651] 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 13:27:31.276617 134259751716672 maxtext_utils.py:1651] 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 13:27:31.276657 134259751716672 maxtext_utils.py:1651] 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 13:27:31.276693 134259751716672 maxtext_utils.py:1651] 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 13:27:31.276739 134259751716672 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0417 13:27:31.276786 134259751716672 maxtext_utils.py:1651] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0417 13:27:31.300131 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0417 13:27:31.300197 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0417 13:27:31.314941 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0417 13:27:31.314997 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0417 13:27:31.344244 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0417 13:27:31.344311 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0417 13:27:31.358951 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0417 13:27:31.359005 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0417 13:27:31.388140 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0417 13:27:31.388205 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0417 13:27:31.402920 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0417 13:27:31.402976 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0417 13:27:31.446794 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0417 13:27:31.446877 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0417 13:27:31.461526 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0417 13:27:31.461580 134259751716672 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0417 13:27:32.126147 134259751716672 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 13:27:32.126242 134259751716672 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 13:27:32.142087 134259751716672 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0417 13:27:32.142149 134259751716672 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0417 13:27:43.695887 134259751716672 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0417 13:27:43.703320 134259751716672 metric_logger.py:301] number parameters: 1.104 billion
I0417 13:27:45.834200 134259751716672 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0417 13:27:56.050436 134259751716672 checkpointing.py:776] Waited 10.216216325759888 seconds for step 0 to finish before starting checkpointing.
I0417 13:27:56.053451 134259751716672 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 13:27:56.055727 134259751716672 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0417 13:27:56.056992 134259751716672 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0.
I0417 13:27:57.219087 134259751716672 signaling_client.py:364] Using JaxDistributedSignalingClient
I0417 13:27:57.233531 134259751716672 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 13:27:57.233597 134259751716672 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 13:27:57.500081 134259751716672 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.267277s
I0417 13:27:57.500946 134259751716672 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.635 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2737388610839844 s) (per-host)
I0417 13:27:57.501007 134259751716672 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.273818s (batch_requests_ready=0.002216s, total_serialization_initiated=0.270832s, others=0.000770s)
I0417 13:27:57.501080 134259751716672 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.277690s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.277672)
I0417 13:27:57.501820 134259751716672 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.45 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0.
I0417 13:27:57.502179 134132845926144 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:47:57.502141
I0417 13:27:57.518662 134259751716672 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 13:27:57.518964 134130167785216 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 13:27:57.519122 134259751716672 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776432476.0534334, 'wait_for_prev_duration_secs': 6.151199340820312e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776432476.0557668, 'checkpointer_blocking_duration_secs': 1.4465601444244385, 'get_old_steps_start_time': 1776432477.5023522, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776432476.0509932, 'checkpoint_manager_blocking_duration_secs': 1.468088150024414}
I0417 13:27:57.519226 134259751716672 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0417 13:27:57.519278 134259751716672 max_utils.py:750]
Memstats: After params initialized:
I0417 13:27:57.519330 134259751716672 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0417 13:27:57.519365 134259751716672 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0417 13:27:57.519394 134259751716672 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0417 13:27:57.519419 134259751716672 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0417 13:27:57.855109 134259751716672 metric_logger.py:196] completed step: 0, seconds: 1.803, TFLOP/s/device: 30.146, Tokens/s/device: 4543.912, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0417 13:27:57.858130 134259751716672 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/tensorboard/'
I0417 13:27:57.915741 134132365522688 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0
I0417 13:27:58.718662 134259751716672 metric_logger.py:196] completed step: 1, seconds: 12.019, TFLOP/s/device: 4.522, Tokens/s/device: 681.585, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0417 13:27:59.295616 134259751716672 metric_logger.py:196] completed step: 2, seconds: 0.864, TFLOP/s/device: 62.881, Tokens/s/device: 9478.125, total_weights: 262144, loss: 10.563, lm_loss: 10.563, perplexity: 38662.707
I0417 13:27:59.873706 134259751716672 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 6080.613, Tokens/s/device: 916536.138, total_weights: 262144, loss: 10.272, lm_loss: 10.272, perplexity: 28909.668
I0417 13:28:01.816865 134132348737280 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/items
I0417 13:28:02.377928 134125923071744 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776432481476583637, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776432481476583637, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0417 13:28:03.635559 2633 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0417 13:28:07.616674 134130700908288 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/items/array_metadatas/process_0
I0417 13:28:43.351766 134130176177920 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 39.716233s (commit=36.284469s, array_metadata_write=3.431763s)
I0417 13:28:43.353042 134132845926144 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.245 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.12580394744873 s) (per-host)
I0417 13:28:43.353174 134132845926144 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.850882s.
I0417 13:28:45.022906 134132845926144 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776432481476583637, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0417 13:28:46.083402 134132845926144 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 13:28:46.649024 134125923071744 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776432481476583637, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0417 13:28:46.694641 134132845926144 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000238s.
I0417 13:28:46.716440 134132845926144 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.557347s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/items
I0417 13:28:46.718111 134132845926144 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0/items
I0417 13:28:47.773746 134132845926144 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0
I0417 13:28:49.667997 134132845926144 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0`.
I0417 13:28:49.668820 134132845926144 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.61 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/0
I0417 13:28:49.670286 134132845926144 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.167990s.
I0417 13:28:49.957035 134130167785216 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 13:28:49.957150 134130167785216 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 13:28:49.957216 134130167785216 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0417 13:28:49.958947 134130167785216 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0417 13:28:50.253416 134259751716672 metric_logger.py:196] completed step: 4, seconds: 0.578, TFLOP/s/device: 93.988, Tokens/s/device: 14166.858, total_weights: 262144, loss: 10.022, lm_loss: 10.022, perplexity: 22524.992
I0417 13:28:50.262035 134259751716672 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.031, Tokens/s/device: 14173.403, total_weights: 262144, loss: 9.820, lm_loss: 9.820, perplexity: 18401.865
I0417 13:28:50.839034 134259751716672 metric_logger.py:196] completed step: 6, seconds: 50.378, TFLOP/s/device: 1.079, Tokens/s/device: 162.610, total_weights: 262144, loss: 9.667, lm_loss: 9.667, perplexity: 15787.604
I0417 13:28:51.417424 134259751716672 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 8462.865, Tokens/s/device: 1275615.073, total_weights: 262144, loss: 9.561, lm_loss: 9.561, perplexity: 14203.827
I0417 13:28:51.995738 134259751716672 metric_logger.py:196] completed step: 8, seconds: 0.578, TFLOP/s/device: 94.042, Tokens/s/device: 14175.095, total_weights: 262144, loss: 9.496, lm_loss: 9.496, perplexity: 13302.920
I0417 13:28:52.573108 134259751716672 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0417 13:28:52.573736 134259751716672 checkpointing.py:776] Waited 0.0006568431854248047 seconds for step 9 to finish before starting checkpointing.
I0417 13:28:52.575749 134259751716672 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 13:28:52.577315 134259751716672 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0417 13:28:52.578676 134259751716672 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9.
I0417 13:28:53.387219 134259751716672 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0417 13:28:53.387308 134259751716672 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 13:28:53.413958 134259751716672 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027405s
I0417 13:28:53.414847 134259751716672 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 47.879 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.032217979431152344 s) (per-host)
I0417 13:28:53.414927 134259751716672 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032336s (batch_requests_ready=0.002148s, total_serialization_initiated=0.029383s, others=0.000805s)
I0417 13:28:53.415012 134259751716672 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036304s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.036287)
I0417 13:28:53.415724 134259751716672 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.84 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9.
I0417 13:28:53.416018 134130167785216 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 13:48:53.415985
I0417 13:28:53.424192 134259751716672 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0417 13:28:53.424451 134131803473664 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0417 13:28:53.424603 134259751716672 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776432532.575716, 'wait_for_prev_duration_secs': 7.390975952148438e-05, 'time_between_consecutive_saves_sec': 2.6167290210723877, 'checkpointer_blocking_start_time': 1776432532.577352, 'checkpointer_blocking_duration_secs': 0.8388054370880127, 'get_old_steps_start_time': 1776432533.416183, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776432532.5739791, 'checkpoint_manager_blocking_duration_secs': 0.8505890369415283}
I0417 13:28:53.424709 134259751716672 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0417 13:28:53.424756 134259751716672 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0417 13:28:53.622988 134132845926144 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9
I0417 13:28:57.991629 134136610133760 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9/items
I0417 13:29:04.261037 134131828651776 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9/items/array_metadatas/process_0
I0417 13:29:43.455022 134131820259072 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.494093s (commit=40.994668s, array_metadata_write=2.499425s)
I0417 13:29:43.456141 134130167785216 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 31.546 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 50.0734965801239 s) (per-host)
I0417 13:29:43.456268 134130167785216 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 50.040136s.
I0417 13:29:46.032482 134130167785216 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0417 13:29:46.947049 134130167785216 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000251s.
I0417 13:29:46.947232 134130167785216 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.750724s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9/items
I0417 13:29:46.948771 134130167785216 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9/items
I0417 13:29:47.985013 134130167785216 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9
I0417 13:29:50.720989 134130167785216 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9`.
I0417 13:29:50.721780 134130167785216 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 58.14 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_125726/linen_xpk_main_20260417_125726_06_grad_accum/checkpoints/9
I0417 13:29:50.723284 134130167785216 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 57.307149s.
I0417 13:29:50.723416 134131803473664 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0417 13:29:50.723474 134131803473664 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0417 13:29:50.723538 134131803473664 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0417 13:29:50.725520 134131803473664 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0417 13:29:50.725672 134259751716672 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 13:29:50.725818 134259751716672 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0417 13:29:50.726881 134259751716672 metric_logger.py:196] completed step: 9, seconds: 0.578, TFLOP/s/device: 93.958, Tokens/s/device: 14162.327, total_weights: 262144, loss: 9.457, lm_loss: 9.457, perplexity: 12802.546
Per train step:
Total TFLOPs: 54.35
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 17 13:30:00 UTC 2026
EXIT_CODE=0