XPK Start: Sat Apr 18 00:22:23 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-18 00:22:47.619305: 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)
I0418 00:22:48.162910 133755800225600 max_utils.py:273] Attempting to initialize the jax distributed system...
I0418 00:22:57.203809 133755800225600 distributed.py:149] Starting JAX distributed service on [::]:8482
I0418 00:22:57.206246 133755800225600 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-3b0rz-slice-job-0-0.mt-06-grad-accum-3b0rz:8482
I0418 00:22:58.618510 133755800225600 max_utils.py:284] Jax distributed system initialized!
I0418 00:23:04.990087 133755800225600 max_utils.py:800] System Information: Jax Version: 0.9.2
I0418 00:23:04.990185 133755800225600 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0418 00:23:04.990225 133755800225600 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
I0418 00:23:04.990260 133755800225600 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0418 00:23:05.667929 133755800225600 maxtext_utils.py:1551] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0418 00:23:05.668210 133755800225600 checkpointing.py:677] Setting up checkpoint logger...
I0418 00:23:05.668264 133755800225600 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0418 00:23:05.668311 133755800225600 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0418 00:23:05.668641 133755800225600 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 0x79a5d6f14500>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0418 00:23:08.559593 133755800225600 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0418 00:23:08.559822 133755800225600 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 0x79a4ac523e90>}, handler_registry=None
I0418 00:23:08.560078 133755800225600 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a4ac523e90>` 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`.
I0418 00:23:08.560131 133755800225600 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x799180595010>` 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`.
I0418 00:23:08.560168 133755800225600 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 0x79a4ac523e90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79a4ac523e90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x799180595010>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x799180595010>}).
I0418 00:23:08.560492 133755800225600 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0418 00:23:08.560561 133755800225600 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7990946f3c40> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0418 00:23:10.737154 133755800225600 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints
I0418 00:23:11.183186 133755800225600 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_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79918058bc20>
I0418 00:23:11.183350 133755800225600 checkpointing.py:301] Checkpoint manager created!
I0418 00:23:12.863284 133755800225600 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0418 00:23:12.863400 133755800225600 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 00:23:13.241464 133755800225600 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 00:23:13.241554 133755800225600 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 00:23:13.257721 133755800225600 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0418 00:23:13.257782 133755800225600 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0418 00:23:13.281136 133755800225600 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 00:23:13.281200 133755800225600 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:23:13.298067 133755800225600 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 00:23:13.298125 133755800225600 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:23:13.314227 133755800225600 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0418 00:23:13.314285 133755800225600 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:23:13.339497 133755800225600 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0418 00:23:13.339577 133755800225600 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0418 00:23:13.360272 133755800225600 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0418 00:23:13.360332 133755800225600 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0418 00:23:13.572973 133755800225600 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0418 00:23:13.573076 133755800225600 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0418 00:23:14.985681 133755800225600 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0418 00:23:14.985804 133755800225600 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0418 00:23:14.985881 133755800225600 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0418 00:23:14.985946 133755800225600 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0418 00:23:14.986000 133755800225600 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0418 00:23:14.986036 133755800225600 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0418 00:23:14.986085 133755800225600 maxtext_utils.py:1654] 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)
I0418 00:23:14.986135 133755800225600 maxtext_utils.py:1654] 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')
I0418 00:23:14.986175 133755800225600 maxtext_utils.py:1654] 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)
I0418 00:23:14.986212 133755800225600 maxtext_utils.py:1654] 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)
I0418 00:23:14.986256 133755800225600 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0418 00:23:14.986305 133755800225600 maxtext_utils.py:1654] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0418 00:23:15.012083 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0418 00:23:15.012156 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0418 00:23:15.026911 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0418 00:23:15.026968 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0418 00:23:15.056056 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0418 00:23:15.056117 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0418 00:23:15.070619 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0418 00:23:15.070675 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0418 00:23:15.099730 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0418 00:23:15.099799 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0418 00:23:15.114359 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0418 00:23:15.114412 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0418 00:23:15.157888 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0418 00:23:15.157959 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0418 00:23:15.172572 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0418 00:23:15.172628 133755800225600 gradient_accumulation.py:68] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0418 00:23:15.845242 133755800225600 train.py:155] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 00:23:15.845339 133755800225600 train.py:155] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 00:23:15.860769 133755800225600 train.py:162] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0418 00:23:15.860833 133755800225600 train.py:162] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0418 00:23:27.239102 133755800225600 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.
I0418 00:23:27.246863 133755800225600 metric_logger.py:301] number parameters: 1.104 billion
I0418 00:23:29.426914 133755800225600 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0418 00:23:39.913004 133755800225600 checkpointing.py:776] Waited 10.486067056655884 seconds for step 0 to finish before starting checkpointing.
I0418 00:23:39.915343 133755800225600 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:23:39.917140 133755800225600 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0418 00:23:39.918619 133755800225600 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0.
I0418 00:23:40.665384 133755800225600 signaling_client.py:364] Using JaxDistributedSignalingClient
I0418 00:23:40.679661 133755800225600 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 00:23:40.679729 133755800225600 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
I0418 00:23:40.950627 133755800225600 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.271693s
I0418 00:23:40.951492 133755800225600 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.543 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27831578254699707 s) (per-host)
I0418 00:23:40.951603 133755800225600 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.278447s (batch_requests_ready=0.002367s, total_serialization_initiated=0.275274s, others=0.000807s)
I0418 00:23:40.951717 133755800225600 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.282511s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.282493)
I0418 00:23:40.952633 133755800225600 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.04 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0.
I0418 00:23:40.952921 133627681687296 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 00:43:40.952891
I0418 00:23:40.954786 133627673294592 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0
I0418 00:23:40.963529 133755800225600 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 00:23:40.963787 133623904126720 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 00:23:40.963952 133755800225600 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776471819.9153244, 'wait_for_prev_duration_secs': 6.222724914550781e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776471819.9171796, 'checkpointer_blocking_duration_secs': 1.0358870029449463, 'get_old_steps_start_time': 1776471820.9530923, 'get_old_steps_duration_secs': 4.0531158447265625e-05, 'checkpoint_manager_blocking_start_time': 1776471819.9135234, 'checkpoint_manager_blocking_duration_secs': 1.050382137298584}
I0418 00:23:40.964079 133755800225600 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0418 00:23:40.964139 133755800225600 max_utils.py:750]
Memstats: After params initialized:
I0418 00:23:40.964199 133755800225600 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0418 00:23:40.964240 133755800225600 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0418 00:23:40.964275 133755800225600 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0418 00:23:40.964307 133755800225600 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0418 00:23:41.298326 133755800225600 metric_logger.py:196] completed step: 0, seconds: 1.858, TFLOP/s/device: 29.255, Tokens/s/device: 4409.631, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0418 00:23:41.301623 133755800225600 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/tensorboard/'
I0418 00:23:41.874396 133755800225600 metric_logger.py:196] completed step: 1, seconds: 11.870, TFLOP/s/device: 4.579, Tokens/s/device: 690.170, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0418 00:23:42.452961 133755800225600 metric_logger.py:196] completed step: 2, seconds: 0.408, TFLOP/s/device: 133.165, Tokens/s/device: 20072.085, total_weights: 262144, loss: 10.563, lm_loss: 10.563, perplexity: 38662.707
I0418 00:23:43.030935 133755800225600 metric_logger.py:196] completed step: 3, seconds: 0.180, TFLOP/s/device: 301.883, Tokens/s/device: 45503.022, total_weights: 262144, loss: 10.272, lm_loss: 10.272, perplexity: 28909.668
I0418 00:23:44.041898 133627656509184 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/items
I0418 00:23:44.504155 133623892678400 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776471823766969699, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776471823766969699, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:23:45.162867 2618 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0418 00:23:47.778261 133624843650816 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_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/items/array_metadatas/process_0
I0418 00:24:23.568598 133624835258112 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.413037s (commit=36.342292s, array_metadata_write=2.070745s)
I0418 00:24:23.569669 133627681687296 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 36.823 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.89650535583496 s) (per-host)
I0418 00:24:23.569736 133627681687296 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.616705s.
I0418 00:24:24.641929 133627681687296 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776471823766969699, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:24:25.849791 133623892678400 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776471823766969699, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0418 00:24:26.610326 133627681687296 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 00:24:26.625412 133627681687296 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000240s.
I0418 00:24:27.547791 133627681687296 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.334357s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/items
I0418 00:24:27.549563 133627681687296 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0/items
I0418 00:24:28.513025 133627681687296 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0
I0418 00:24:30.223993 133755800225600 metric_logger.py:196] completed step: 4, seconds: 0.578, TFLOP/s/device: 94.066, Tokens/s/device: 14178.677, total_weights: 262144, loss: 10.022, lm_loss: 10.022, perplexity: 22524.992
I0418 00:24:30.233160 133755800225600 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 94.042, Tokens/s/device: 14175.070, total_weights: 262144, loss: 9.820, lm_loss: 9.820, perplexity: 18401.865
I0418 00:24:30.312296 133627681687296 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0`.
I0418 00:24:30.313164 133627681687296 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.40 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/0
I0418 00:24:30.314621 133627681687296 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 49.361587s.
I0418 00:24:30.314854 133623904126720 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 00:24:30.314987 133623904126720 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 00:24:30.315053 133623904126720 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0418 00:24:30.317074 133623904126720 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0418 00:24:30.809980 133755800225600 metric_logger.py:196] completed step: 6, seconds: 47.192, TFLOP/s/device: 1.152, Tokens/s/device: 173.590, total_weights: 262144, loss: 9.667, lm_loss: 9.667, perplexity: 15787.604
I0418 00:24:31.388320 133755800225600 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 8144.540, Tokens/s/device: 1227633.748, total_weights: 262144, loss: 9.561, lm_loss: 9.561, perplexity: 14203.827
I0418 00:24:31.966651 133755800225600 metric_logger.py:196] completed step: 8, seconds: 0.578, TFLOP/s/device: 94.037, Tokens/s/device: 14174.261, total_weights: 262144, loss: 9.496, lm_loss: 9.496, perplexity: 13302.920
I0418 00:24:32.544122 133755800225600 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0418 00:24:32.544881 133755800225600 checkpointing.py:776] Waited 0.0007798671722412109 seconds for step 9 to finish before starting checkpointing.
I0418 00:24:32.546939 133755800225600 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:24:32.548463 133755800225600 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0418 00:24:32.549899 133755800225600 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9.
I0418 00:24:33.257581 133755800225600 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0418 00:24:33.257677 133755800225600 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
I0418 00:24:33.284290 133755800225600 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027790s
I0418 00:24:33.285010 133755800225600 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 47.350 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.032578229904174805 s) (per-host)
I0418 00:24:33.285073 133755800225600 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032657s (batch_requests_ready=0.001957s, total_serialization_initiated=0.030026s, others=0.000674s)
I0418 00:24:33.285139 133755800225600 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036487s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.036471)
I0418 00:24:33.285826 133755800225600 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9.
I0418 00:24:33.286099 133623904126720 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-18 00:44:33.286078
I0418 00:24:33.293148 133755800225600 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0418 00:24:33.293384 133626615215872 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0418 00:24:33.293503 133755800225600 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776471872.546906, 'wait_for_prev_duration_secs': 7.128715515136719e-05, 'time_between_consecutive_saves_sec': 2.229793071746826, 'checkpointer_blocking_start_time': 1776471872.5485024, 'checkpointer_blocking_duration_secs': 0.7376959323883057, 'get_old_steps_start_time': 1776471873.2862222, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776471872.5451071, 'checkpoint_manager_blocking_duration_secs': 0.7483615875244141}
I0418 00:24:33.293607 133755800225600 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0418 00:24:33.293652 133755800225600 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0418 00:24:33.911699 133627681687296 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9
I0418 00:24:36.944391 133631968261888 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9/items
I0418 00:24:43.540977 133626632001280 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_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9/items/array_metadatas/process_0
I0418 00:25:21.085336 133626623608576 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.310073s (commit=40.116788s, array_metadata_write=1.193285s)
I0418 00:25:21.086366 133623904126720 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.023 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 47.83390235900879 s) (per-host)
I0418 00:25:21.086494 133623904126720 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.800316s.
I0418 00:25:23.571336 133623904126720 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0418 00:25:24.639192 133623904126720 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000228s.
I0418 00:25:24.639387 133623904126720 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.323559s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9/items
I0418 00:25:24.644409 133623904126720 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9/items
I0418 00:25:25.192446 133623904126720 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9
I0418 00:25:27.499374 133623904126720 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9`.
I0418 00:25:27.500257 133623904126720 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 54.95 seconds @ gs://lance-maxtext/linen_ckpt_xpk_main_20260417_235819/linen_xpk_main_20260417_235819_06_grad_accum/checkpoints/9
I0418 00:25:27.501508 133623904126720 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 54.215328s.
I0418 00:25:27.501648 133626615215872 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0418 00:25:27.501708 133626615215872 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0418 00:25:27.501783 133626615215872 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0418 00:25:27.504063 133626615215872 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0418 00:25:27.504237 133755800225600 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.
I0418 00:25:27.504386 133755800225600 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0418 00:25:27.505395 133755800225600 metric_logger.py:196] completed step: 9, seconds: 0.578, TFLOP/s/device: 93.995, Tokens/s/device: 14167.912, 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: Sat Apr 18 00:25:40 UTC 2026
EXIT_CODE=0