MaxView

← Back to run

Log Summary

XPK Start: Wed Apr 22 17:32:31 UTC 2026
2026-04-22 17:32:35.463482: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1776879155.476710      10 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776879155.480484      10 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776879155.492071      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776879155.492104      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776879155.492106      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776879155.492108      10 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-04-22 17:32:54.725035: 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)
I0422 17:32:55.251778 135387557844800 max_utils.py:273] Attempting to initialize the jax distributed system...
INFO:2026-04-22 17:33:04,292:jax._src.distributed:140: Starting JAX distributed service on [::]:8482
I0422 17:33:04.292955 135387557844800 distributed.py:140] Starting JAX distributed service on [::]:8482
INFO:2026-04-22 17:33:04,295:jax._src.distributed:157: Connecting to JAX distributed service on mt-01-base-ghqzx-slice-job-0-0.mt-01-base-ghqzx:8482
I0422 17:33:04.295393 135387557844800 distributed.py:157] Connecting to JAX distributed service on mt-01-base-ghqzx-slice-job-0-0.mt-01-base-ghqzx:8482
I0422 17:33:05.281018 135387557844800 max_utils.py:284] Jax distributed system initialized!
I0422 17:33:11.100903 135387557844800 max_utils.py:800] System Information: Jax Version: 0.8.1
I0422 17:33:11.101010 135387557844800 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0422 17:33:11.101052 135387557844800 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709
I0422 17:33:11.101088 135387557844800 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 17:33:11.724749 135387557844800 maxtext_utils.py:1631] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 17:33:11.839450 135387557844800 checkpointing.py:688] Setting up checkpoint logger...
I0422 17:33:11.839568 135387557844800 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0422 17:33:11.839612 135387557844800 pytree_checkpoint_handler.py:589] save_device_host_concurrent_bytes=None
I0422 17:33:11.839812 135387557844800 base_pytree_checkpoint_handler.py:415] 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 0x7b21dd7792b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 17:33:15.560358 135387557844800 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0422 17:33:15.560705 135387557844800 checkpoint_manager.py:709] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b0fa5af1160>}, handler_registry=None
I0422 17:33:15.560938 135387557844800 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b0fa5af1160>` 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`.
I0422 17:33:15.560987 135387557844800 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b0fa422b980>` 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`.
I0422 17:33:15.561023 135387557844800 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 0x7b0fa5af1160>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b0fa5af1160>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b0fa422b980>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b0fa422b980>}).
I0422 17:33:15.561368 135387557844800 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.33
I0422 17:33:15.561445 135387557844800 async_checkpointer.py:177] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7b0fa47d0f40> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0422 17:33:16.965425 135387557844800 checkpoint_manager.py:1818] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints
I0422 17:33:16.967648 135387557844800 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created,  primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b0fa5af2330>
I0422 17:33:16.967763 135387557844800 checkpointing.py:302] Checkpoint manager created!
I0422 17:33:17.898231 135387557844800 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 17:33:17.898351 135387557844800 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0422 17:33:19.361805 135387557844800 maxtext_utils.py:1749]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0422 17:33:19.361927 135387557844800 maxtext_utils.py:1749]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 17:33:19.361983 135387557844800 maxtext_utils.py:1749]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 17:33:19.362028 135387557844800 maxtext_utils.py:1749]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0422 17:33:19.362065 135387557844800 maxtext_utils.py:1749]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 17:33:19.362116 135387557844800 maxtext_utils.py:1749]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 17:33:19.362160 135387557844800 maxtext_utils.py:1749]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 17:33:19.362198 135387557844800 maxtext_utils.py:1749]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0422 17:33:19.362234 135387557844800 maxtext_utils.py:1749]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 17:33:19.362267 135387557844800 maxtext_utils.py:1749]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 17:33:19.362300 135387557844800 maxtext_utils.py:1749]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)

I0422 17:33:19.362338 135387557844800 maxtext_utils.py:1749]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0422 17:33:19.756969 135387557844800 nnx_decoders.py:616] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 17:33:19.757129 135387557844800 nnx_decoders.py:616] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 17:33:19.762475 135387557844800 nnx_decoders.py:616] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 17:33:19.762567 135387557844800 nnx_decoders.py:616] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 17:33:19.910608 135387557844800 attentions.py:1084] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 17:33:19.910739 135387557844800 attentions.py:1084] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 17:33:20.065818 135387557844800 attentions.py:1085] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 17:33:20.065957 135387557844800 attentions.py:1085] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 17:33:20.227804 135387557844800 attentions.py:1150] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 17:33:20.227942 135387557844800 attentions.py:1150] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 17:33:20.373043 135387557844800 attentions.py:1151] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 17:33:20.373232 135387557844800 attentions.py:1151] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 17:33:20.516801 135387557844800 attentions.py:1152] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 17:33:20.516936 135387557844800 attentions.py:1152] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 17:33:20.671492 135387557844800 attentions.py:1193] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 17:33:20.671640 135387557844800 attentions.py:1193] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 17:33:20.694138 135387557844800 linears.py:541] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 17:33:20.694233 135387557844800 linears.py:541] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 17:33:31.517740 135387557844800 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.
I0422 17:33:31.520133 135387557844800 metric_logger.py:289] number parameters: 1.104 billion
I0422 17:33:42.450463 135387557844800 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0422 17:33:42.609512 135387557844800 checkpointing.py:798] Waited 0.15903592109680176 seconds for step 0 to finish before starting checkpointing.
I0422 17:33:42.612159 135387557844800 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 17:33:42.613943 135387557844800 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 0
I0422 17:33:42.616624 135387557844800 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints/0.
I0422 17:33:43.198686 135387557844800 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 17:33:43.199728 135387557844800 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0422 17:33:43.199786 135387557844800 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
I0422 17:33:43.449856 135387557844800 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.252301s
I0422 17:33:43.450026 135387557844800 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.005 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.5133066177368164 s) (per-host)
I0422 17:33:43.450075 135387557844800 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.513365s (batch_requests_ready=0.254538s, total_serialization_initiated=0.258760s, others=0.000067s)
I0422 17:33:43.450189 135387557844800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.517452s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.517432)
I0422 17:33:43.451144 135265141872384 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0422 17:33:43.451317 135387557844800 async_checkpointer.py:561] Finished blocking save. Time taken: 0.837306s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints/0.
I0422 17:33:43.497734 135387557844800 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 17:33:43.498054 135264039048960 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 17:33:43.498220 135387557844800 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879222.6121407, 'wait_for_prev_duration_secs': 6.127357482910156e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879222.6139827, 'checkpointer_blocking_duration_secs': 0.8374404907226562, 'get_old_steps_start_time': 1776879223.451444, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776879222.6099112, 'checkpoint_manager_blocking_duration_secs': 0.8882708549499512}
I0422 17:33:43.498334 135387557844800 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0422 17:33:43.498407 135387557844800 max_utils.py:750] 
Memstats: After params initialized:
I0422 17:33:43.498459 135387557844800 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0))
I0422 17:33:43.498492 135387557844800 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0))
I0422 17:33:43.498518 135387557844800 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0))
I0422 17:33:43.498542 135387557844800 max_utils.py:756] 	Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0))
I0422 17:33:43.769733 135387557844800 metric_logger.py:185] completed step: 0, seconds: 10.929, TFLOP/s/device: 1.243, Tokens/s/device: 187.387, total_weights: 65536, loss: 10.874
I0422 17:33:43.934247 135387557844800 metric_logger.py:185] completed step: 1, seconds: 1.317, TFLOP/s/device: 10.313, Tokens/s/device: 1554.547, total_weights: 65536, loss: 10.874
I0422 17:33:44.385371 135387557844800 metric_logger.py:185] completed step: 2, seconds: 0.019, TFLOP/s/device: 708.401, Tokens/s/device: 106777.894, total_weights: 65536, loss: 10.262
I0422 17:33:44.532828 135387557844800 metric_logger.py:185] completed step: 3, seconds: 0.441, TFLOP/s/device: 30.822, Tokens/s/device: 4645.845, total_weights: 65536, loss: 9.732
I0422 17:33:44.828184 135387557844800 metric_logger.py:185] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.893, Tokens/s/device: 12494.509, total_weights: 65536, loss: 9.274
I0422 17:33:44.836293 135387557844800 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.418, Tokens/s/device: 13930.172, total_weights: 65536, loss: 8.888
I0422 17:34:03.860436 135387557844800 metric_logger.py:185] completed step: 6, seconds: 0.296, TFLOP/s/device: 45.884, Tokens/s/device: 6916.209, total_weights: 65536, loss: 8.589
I0422 17:34:04.007846 135387557844800 metric_logger.py:185] completed step: 7, seconds: 18.878, TFLOP/s/device: 0.720, Tokens/s/device: 108.484, total_weights: 65536, loss: 8.380
I0422 17:34:04.155211 135387557844800 metric_logger.py:185] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.582, Tokens/s/device: 13352.023, total_weights: 65536, loss: 8.251
I0422 17:34:04.303636 135387557844800 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0422 17:34:04.304265 135387557844800 checkpointing.py:798] Waited 0.0007033348083496094 seconds for step 9 to finish before starting checkpointing.
I0422 17:34:04.306472 135387557844800 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 17:34:06.123055    2562 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 17:34:08.592587 135265150265088 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints/0/items/array_metadatas/process_6
I0422 17:34:39.152971 135265141872384 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 28.099 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 56.21620965003967 s) (per-host)
I0422 17:34:39.153166 135265141872384 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 55.701834s.
I0422 17:34:47.312187 135265141872384 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 63.860911s.
I0422 17:34:47.312495 135264039048960 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 17:34:47.312624 135264039048960 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 17:34:47.312690 135264039048960 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0422 17:34:47.314076 135264039048960 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0422 17:34:47.314314 135387557844800 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0422 17:34:47.314469 135387557844800 checkpoint_manager.py:1458] Waiting for previous save to complete took 43.007996 seconds. If this number is high, consider checkpointing less frequently.
I0422 17:34:47.316699 135387557844800 checkpoint_manager.py:1518] [process=6] Saving checkpoint at step 9
I0422 17:34:47.320174 135387557844800 async_checkpointer.py:452] [process=6] Started async saving checkpoint to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints/9.
I0422 17:34:48.294056 135387557844800 jax_array_handlers.py:358] Scheduling D2H of 69 prioritized jax.Array.
I0422 17:34:48.294224 135387557844800 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
I0422 17:34:48.337084 135387557844800 base_pytree_checkpoint_handler.py:153] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.044991s
I0422 17:34:48.337267 135387557844800 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.980 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3097364902496338 s) (per-host)
I0422 17:34:48.337321 135387557844800 base_pytree_checkpoint_handler.py:737] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.309801s (batch_requests_ready=0.260073s, total_serialization_initiated=0.049654s, others=0.000074s)
I0422 17:34:48.337421 135387557844800 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.313738s (all_items=0.000016s, per_item={'items': '0.00001574'}, temp_paths=0.313723)
I0422 17:34:48.338316 135264576726784 async_checkpointer.py:79] [process=6][thread=async_save] Background save thread started.
I0422 17:34:48.338436 135387557844800 async_checkpointer.py:561] Finished blocking save. Time taken: 1.021659s. Continuing background save to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints/9.
I0422 17:34:48.347380 135387557844800 checkpoint_manager.py:1566] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 17:34:48.347601 135264039048960 async_checkpointer.py:265] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 17:34:48.347723 135387557844800 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776879244.306444, 'wait_for_prev_duration_secs': 43.00799608230591, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776879287.3167439, 'checkpointer_blocking_duration_secs': 1.0218021869659424, 'get_old_steps_start_time': 1776879288.3385649, 'get_old_steps_duration_secs': 2.574920654296875e-05, 'checkpoint_manager_blocking_start_time': 1776879244.3044882, 'checkpoint_manager_blocking_duration_secs': 44.04320287704468}
I0422 17:34:48.347831 135387557844800 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0422 17:34:48.347874 135387557844800 checkpoint_manager.py:2024] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 17:34:54.583616 135265150265088 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_set_defaults_true_20260422_154647/nnx_xpk_feat_nnx_set_defaults_true_20260422_154647_01_base/checkpoints/9/items/array_metadatas/process_6
I0422 17:35:32.093190 135264576726784 base_pytree_checkpoint_handler.py:129] [process=6] /jax/checkpoint/write/gbytes_per_sec: 35.846 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 44.06562161445618 s) (per-host)
I0422 17:35:32.093310 135264576726784 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.754922s.
I0422 17:35:38.244015 135264576726784 async_checkpointer.py:144] [process=6][thread=async_save] Background save thread done. Time taken: 49.905613s.
I0422 17:35:38.244285 135264039048960 async_checkpointer.py:273] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 17:35:38.244341 135264039048960 async_checkpointer.py:283] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 17:35:38.244385 135264039048960 checkpoint_manager.py:2133] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0422 17:35:38.245950 135264039048960 checkpoint_manager.py:2142] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0422 17:35:38.246119 135387557844800 checkpoint_manager.py:2036] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0422 17:35:38.246287 135387557844800 checkpoint_manager.py:2013] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 17:35:38.246898 135387557844800 metric_logger.py:185] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.381, Tokens/s/device: 13924.679, total_weights: 65536, loss: 8.175
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Wed Apr 22 17:35:47 UTC 2026
EXIT_CODE=0