MaxView

← Back to run

Log Summary

XPK Start: Mon Apr 20 17:35:08 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-20 17:35:32.980885: 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)
I0420 17:35:33.160125 137969087797056 max_utils.py:273] Attempting to initialize the jax distributed system...
I0420 17:35:42.202534 137969087797056 distributed.py:149] Starting JAX distributed service on [::]:8482
I0420 17:35:42.204884 137969087797056 distributed.py:172] Connecting to JAX distributed service on mt-01-base-7ql7s-slice-job-0-0.mt-01-base-7ql7s:8482
I0420 17:35:43.254761 137969087797056 max_utils.py:284] Jax distributed system initialized!
I0420 17:35:49.277896 137969087797056 max_utils.py:800] System Information: Jax Version: 0.9.2
I0420 17:35:49.277995 137969087797056 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0420 17:35:49.278035 137969087797056 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
I0420 17:35:49.278069 137969087797056 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 17:35:49.962877 137969087797056 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 17:35:50.077117 137969087797056 checkpointing.py:688] Setting up checkpoint logger...
I0420 17:35:50.077229 137969087797056 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 17:35:50.077274 137969087797056 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0420 17:35:50.077467 137969087797056 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 0x7d7ab6b50e30>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 17:35:53.964388 137969087797056 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 17:35:53.964621 137969087797056 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 0x7d79aa5fa000>}, handler_registry=None
I0420 17:35:53.964860 137969087797056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d79aa5fa000>` 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`.
I0420 17:35:53.964908 137969087797056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d656c425550>` 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`.
I0420 17:35:53.964944 137969087797056 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 0x7d79aa5fa000>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d79aa5fa000>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d656c425550>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d656c425550>}).
I0420 17:35:53.965272 137969087797056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0420 17:35:53.965343 137969087797056 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d656c144fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0420 17:35:55.096050 137969087797056 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints
I0420 17:35:55.473138 137969087797056 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d65486421b0>
I0420 17:35:55.473299 137969087797056 checkpointing.py:302] Checkpoint manager created!
I0420 17:35:56.455499 137969087797056 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 17:35:56.455610 137969087797056 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0420 17:35:57.809866 137969087797056 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0420 17:35:57.809978 137969087797056 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 17:35:57.810028 137969087797056 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0420 17:35:57.810101 137969087797056 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0420 17:35:57.810143 137969087797056 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 17:35:57.810179 137969087797056 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0420 17:35:57.810227 137969087797056 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 17:35:57.810277 137969087797056 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0420 17:35:57.810324 137969087797056 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 17:35:57.810360 137969087797056 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0420 17:35:57.810392 137969087797056 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0420 17:35:57.810424 137969087797056 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0420 17:35:58.217347 137969087797056 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 17:35:58.217441 137969087797056 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 17:35:58.222730 137969087797056 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 17:35:58.222788 137969087797056 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 17:35:58.239041 137969087797056 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 17:35:58.239115 137969087797056 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 17:35:58.254591 137969087797056 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 17:35:58.254648 137969087797056 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 17:35:58.279405 137969087797056 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 17:35:58.279472 137969087797056 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 17:35:58.295224 137969087797056 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 17:35:58.295282 137969087797056 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 17:35:58.310877 137969087797056 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 17:35:58.310933 137969087797056 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 17:35:58.340853 137969087797056 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 17:35:58.340922 137969087797056 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 17:35:58.360636 137969087797056 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 17:35:58.360699 137969087797056 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 17:36:09.669820 137969087797056 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.
I0420 17:36:09.679851 137969087797056 metric_logger.py:301] number parameters: 1.104 billion
I0420 17:36:21.178706 137969087797056 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 17:36:21.329730 137969087797056 checkpointing.py:798] Waited 0.15101194381713867 seconds for step 0 to finish before starting checkpointing.
I0420 17:36:21.331975 137969087797056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 17:36:21.333755 137969087797056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0420 17:36:21.335153 137969087797056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0.
I0420 17:36:21.648708 137969087797056 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 17:36:21.667818 137969087797056 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 17:36:21.667893 137969087797056 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
I0420 17:36:21.913842 137839242323712 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0
I0420 17:36:21.940882 137969087797056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.274477s
I0420 17:36:21.941653 137969087797056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.405 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28540802001953125 s) (per-host)
I0420 17:36:21.941715 137969087797056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285488s (batch_requests_ready=0.003220s, total_serialization_initiated=0.281585s, others=0.000684s)
I0420 17:36:21.941794 137969087797056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289357s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.289340)
I0420 17:36:21.942580 137969087797056 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0.
I0420 17:36:21.942857 137839250716416 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:56:21.942828
I0420 17:36:21.951918 137969087797056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 17:36:21.952155 137837601396480 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 17:36:21.952276 137969087797056 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776706581.3319566, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776706581.3337939, 'checkpointer_blocking_duration_secs': 0.609158992767334, 'get_old_steps_start_time': 1776706581.9429772, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776706581.3301303, 'checkpoint_manager_blocking_duration_secs': 0.6221096515655518}
I0420 17:36:21.952381 137969087797056 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 17:36:21.952445 137969087797056 max_utils.py:750] 
Memstats: After params initialized:
I0420 17:36:21.952494 137969087797056 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0420 17:36:21.952526 137969087797056 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0420 17:36:21.952552 137969087797056 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0420 17:36:21.952581 137969087797056 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0420 17:36:22.296418 137969087797056 metric_logger.py:196] completed step: 0, seconds: 11.191, TFLOP/s/device: 1.214, Tokens/s/device: 182.998, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0420 17:36:22.299347 137969087797056 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/tensorboard/'
I0420 17:36:22.762094 137969087797056 metric_logger.py:196] completed step: 1, seconds: 1.114, TFLOP/s/device: 12.197, Tokens/s/device: 1838.400, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0420 17:36:22.904475 137969087797056 metric_logger.py:196] completed step: 2, seconds: 0.467, TFLOP/s/device: 29.111, Tokens/s/device: 4387.901, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133
I0420 17:36:23.051796 137969087797056 metric_logger.py:196] completed step: 3, seconds: 0.012, TFLOP/s/device: 1131.883, Tokens/s/device: 170609.797, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789
I0420 17:36:23.909368 137839215879936 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/items
I0420 17:36:24.232495 137836036617984 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776706583637545347, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776706583637545347, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/_CHECKPOINT_METADATA
I0420 17:36:25.055352    2869 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 17:36:26.830450 137837646108416 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/items/array_metadatas/process_0
I0420 17:36:45.375096 137969087797056 metric_logger.py:196] completed step: 4, seconds: 0.144, TFLOP/s/device: 94.439, Tokens/s/device: 14234.875, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594
I0420 17:36:45.386674 137969087797056 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.616, Tokens/s/device: 13960.083, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978
I0420 17:36:45.530171 137969087797056 metric_logger.py:196] completed step: 6, seconds: 22.323, TFLOP/s/device: 0.609, Tokens/s/device: 91.743, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156
I0420 17:36:45.677641 137969087797056 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1690.572, Tokens/s/device: 254821.451, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465
I0420 17:36:45.825287 137969087797056 metric_logger.py:196] completed step: 8, seconds: 0.145, TFLOP/s/device: 93.733, Tokens/s/device: 14128.523, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256
I0420 17:36:45.973380 137969087797056 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 17:36:45.974047 137969087797056 checkpointing.py:798] Waited 0.0006821155548095703 seconds for step 9 to finish before starting checkpointing.
I0420 17:36:45.976044 137969087797056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 17:37:02.230148 137837637715712 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.182835s (commit=35.952419s, array_metadata_write=1.230416s)
I0420 17:37:02.231366 137839250716416 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 38.930 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 40.575093507766724 s) (per-host)
I0420 17:37:02.231496 137839250716416 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.288567s.
I0420 17:37:05.231951 137839250716416 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776706583637545347, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/_CHECKPOINT_METADATA
I0420 17:37:05.978014 137836036617984 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776706583637545347, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/_CHECKPOINT_METADATA
I0420 17:37:06.422019 137839250716416 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000404s.
I0420 17:37:06.633574 137839250716416 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 17:37:07.572643 137839250716416 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.215700s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/items
I0420 17:37:07.574412 137839250716416 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0/items
I0420 17:37:08.174736 137839250716416 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0
I0420 17:37:09.594823 137839250716416 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0`.
I0420 17:37:09.595716 137839250716416 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.26 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/0
I0420 17:37:09.597232 137839250716416 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.654300s.
I0420 17:37:09.597445 137837601396480 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 17:37:09.597560 137837601396480 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 17:37:09.597621 137837601396480 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 17:37:09.600304 137837601396480 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 17:37:09.600480 137969087797056 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0420 17:37:09.600610 137969087797056 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.624563 seconds. If this number is high, consider checkpointing less frequently.
I0420 17:37:09.602463 137969087797056 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0420 17:37:09.604480 137969087797056 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9.
I0420 17:37:09.920798 137969087797056 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0420 17:37:09.920897 137969087797056 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
I0420 17:37:09.948198 137969087797056 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.028848s
I0420 17:37:09.949021 137969087797056 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 40.821 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03778886795043945 s) (per-host)
I0420 17:37:09.949109 137969087797056 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.037894s (batch_requests_ready=0.002783s, total_serialization_initiated=0.034357s, others=0.000754s)
I0420 17:37:09.949200 137969087797056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.041950s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.041932)
I0420 17:37:09.949977 137969087797056 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9.
I0420 17:37:09.950285 137837601396480 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 17:57:09.950254
I0420 17:37:09.956567 137969087797056 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 17:37:09.956813 137837101967104 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 17:37:09.956943 137969087797056 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776706605.9760141, 'wait_for_prev_duration_secs': 23.624563455581665, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776706629.6025002, 'checkpointer_blocking_duration_secs': 0.3479187488555908, 'get_old_steps_start_time': 1776706629.9504442, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776706605.9742732, 'checkpoint_manager_blocking_duration_secs': 23.98263454437256}
I0420 17:37:09.957063 137969087797056 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 17:37:09.957127 137969087797056 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 17:37:10.158771 137839250716416 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9
I0420 17:37:11.924141 137815450949376 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9/items
I0420 17:37:15.490639 137837640496896 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9/items/array_metadatas/process_0
I0420 17:37:55.467184 137837110359808 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.429506s (commit=41.325204s, array_metadata_write=1.104302s)
I0420 17:37:55.468407 137837601396480 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.673 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.55714750289917 s) (per-host)
I0420 17:37:55.468539 137837601396480 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.518143s.
I0420 17:37:56.976633 137837601396480 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000426s.
I0420 17:37:57.666776 137837601396480 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 17:37:58.594644 137837601396480 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.325893s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9/items
I0420 17:37:58.596347 137837601396480 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9/items
I0420 17:37:59.126694 137837601396480 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9
I0420 17:38:00.502424 137837601396480 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9`.
I0420 17:38:00.503326 137837601396480 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 50.90 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_01_base/checkpoints/9
I0420 17:38:00.504657 137837601396480 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 50.554258s.
I0420 17:38:00.504855 137837101967104 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 17:38:00.504979 137837101967104 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 17:38:00.505041 137837101967104 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 17:38:00.507820 137837101967104 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 17:38:00.507981 137969087797056 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.
I0420 17:38:00.508108 137969087797056 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 17:38:00.509224 137969087797056 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.278, Tokens/s/device: 13909.169, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3551.438
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Mon Apr 20 17:38:09 UTC 2026
EXIT_CODE=0