MaxView

← Back to run

Log Summary

2026-04-23 03:52:00.144236: 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 00:00:1776916320.158103 2397099 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1776916320.162370 2397099 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1776916320.174686 2397099 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776916320.174700 2397099 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776916320.174702 2397099 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1776916320.174704 2397099 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
`rope_scaling`'s factor field must be a float >= 1, got 40
`rope_scaling`'s beta_fast field must be a float, got 32
`rope_scaling`'s beta_slow field must be a float, got 1
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'}
2026-04-23 03:52:03.847122: 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)
I0423 03:52:03.983557 138804768655168 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
~/maxtext/maxtext_venv/lib/python3.12/site-packages/jax/_src/xla_bridge.py:204: UserWarning: TPU backend initialization is taking more than 60.0 seconds. Did you run your code on all TPU hosts? See https://docs.jax.dev/en/latest/multi_process.html for more information.
  warnings.warn(
I0423 03:53:11.088296 138804768655168 max_utils.py:800] System Information: Jax Version: 0.8.1
I0423 03:53:11.088547 138804768655168 max_utils.py:801] System Information: Jaxlib Version: 0.8.1
I0423 03:53:11.088594 138804768655168 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
I0423 03:53:11.088623 138804768655168 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system
I0423 03:53:11.088646 138804768655168 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0423 03:53:11.088817 138804768655168 train.py:672] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
I0423 03:53:11.743172 138804768655168 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0423 03:53:11.743429 138804768655168 checkpointing.py:677] Setting up checkpoint logger...
I0423 03:53:11.743474 138804768655168 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True
I0423 03:53:11.743509 138804768655168 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0423 03:53:11.743936 138804768655168 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7e3db75b5d90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0423 03:53:14.162284 138804768655168 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0423 03:53:14.162641 138804768655168 checkpoint_manager.py:702] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e3c5bda0440>}, handler_registry=None
I0423 03:53:14.162855 138804768655168 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e3c5bda0440>` 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`.
I0423 03:53:14.162896 138804768655168 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e3bda725850>` 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`.
I0423 03:53:14.162924 138804768655168 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 0x7e3c5bda0440>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e3c5bda0440>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e3bda725850>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e3bda725850>}).
I0423 03:53:14.163216 138804768655168 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0423 03:53:14.163267 138804768655168 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7e3bda754540> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0423 03:53:14.292383 138804768655168 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints
I0423 03:53:14.292594 138804768655168 checkpoint_manager.py:921] [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_hns=False, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=5), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False), root_directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e3bda725f70>
I0423 03:53:14.292874 138804768655168 checkpointing.py:301] Checkpoint manager created!
I0423 03:53:14.684318 138804768655168 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0423 03:53:14.684412 138804768655168 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0423 03:53:14.979399 138804768655168 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 03:53:14.979480 138804768655168 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0423 03:53:14.991647 138804768655168 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0423 03:53:14.991699 138804768655168 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0423 03:53:15.010538 138804768655168 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 03:53:15.010591 138804768655168 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 03:53:15.022932 138804768655168 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 03:53:15.022979 138804768655168 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 03:53:15.035161 138804768655168 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0423 03:53:15.035211 138804768655168 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 03:53:15.054522 138804768655168 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0423 03:53:15.054579 138804768655168 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0423 03:53:15.069897 138804768655168 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0423 03:53:15.069948 138804768655168 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0423 03:53:15.219796 138804768655168 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0423 03:53:15.219890 138804768655168 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
[DECOUPLED NO-OP] gcs_storage: using stubs.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] workload_monitor: using stub.
[DECOUPLED NO-OP] vertex_tensorboard: using stub.
fsdp: 8
I0423 03:53:15.852798 138804768655168 maxtext_utils.py:1707]  params/params/decoder/decoder_norm/scale
    Shape:     float32[2048]
    Logical:   PartitionSpec('norm',)
    Physical:  (None,)
I0423 03:53:15.852907 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wi_0/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 03:53:15.852957 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wi_1/kernel
    Shape:     float32[2048,16,7168]
    Logical:   PartitionSpec('embed', 'layers', 'mlp')
    Physical:  ('fsdp', None, None)
I0423 03:53:15.853004 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/mlp/wo/kernel
    Shape:     float32[7168,16,2048]
    Logical:   PartitionSpec('mlp', 'layers', 'embed')
    Physical:  (None, None, 'fsdp')
I0423 03:53:15.853061 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/post_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0423 03:53:15.853095 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/pre_self_attention_layer_norm/scale
    Shape:     float32[2048,16]
    Logical:   PartitionSpec('norm', 'layers')
    Physical:  (None, None)
I0423 03:53:15.853137 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/key/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0423 03:53:15.853177 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/out/kernel
    Shape:     float32[16,16,128,2048]
    Logical:   PartitionSpec('heads', 'layers', 'kv', 'embed')
    Physical:  (None, None, None, 'fsdp')
I0423 03:53:15.853207 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/query/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'q_heads', 'kv')
    Physical:  ('fsdp', None, None, None)
I0423 03:53:15.853233 138804768655168 maxtext_utils.py:1707]  params/params/decoder/layers/self_attention/value/kernel
    Shape:     float32[2048,16,16,128]
    Logical:   PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim')
    Physical:  ('fsdp', None, None, None)
I0423 03:53:15.853267 138804768655168 maxtext_utils.py:1707]  params/params/decoder/logits_dense/kernel
    Shape:     float32[2048,32000]
    Logical:   PartitionSpec('embed_vocab', 'vocab')
    Physical:  ('fsdp', None)
I0423 03:53:15.853302 138804768655168 maxtext_utils.py:1707]  params/params/token_embedder/embedding
    Shape:     float32[32000,2048]
    Logical:   PartitionSpec('vocab', 'embed_vocab')
    Physical:  (None, 'fsdp')

I0423 03:53:16.260149 138804768655168 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0423 03:53:16.260235 138804768655168 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0423 03:53:16.272003 138804768655168 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0423 03:53:16.272063 138804768655168 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
I0423 03:53:17.030765 138804768655168 max_utils.py:791] Total memory size: 3.6 GB, Output size: 1.5 GB, Temp size: 2.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB.
I0423 03:53:17.031339 138804768655168 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0423 03:53:17.031486 138804768655168 metric_logger.py:301] number parameters: 1.104 billion
I0423 03:53:19.475008 138804768655168 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0423 03:53:19.591878 138804768655168 checkpointing.py:776] Waited 0.11686325073242188 seconds for step 0 to finish before starting checkpointing.
I0423 03:53:19.592487 138804768655168 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 03:53:19.592634 138804768655168 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0423 03:53:19.593004 138804768655168 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0.
I0423 03:53:19.679846 138804768655168 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0423 03:53:19.771447 138804768655168 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0423 03:53:19.771538 138804768655168 replica_slices.py:410] 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
I0423 03:53:19.777236 138773400651328 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0
I0423 03:53:20.530568 138773390165568 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/items
I0423 03:53:20.684813 138773358708288 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776916400446485558, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776916400446485558, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0423 03:53:21.036325 2399252 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0423 03:53:21.036374 2399252 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0423 03:53:21.478244 138804768655168 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.707351s
I0423 03:53:21.482656 138804768655168 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 6.849 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host)
I0423 03:53:21.482918 138804768655168 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.802121s (batch_requests_ready=0.086277s, total_serialization_initiated=1.711327s, others=0.004517s)
I0423 03:53:21.483001 138804768655168 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.802677s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=1.802659)
I0423 03:53:21.484228 138772559693376 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0423 03:53:21.484308 138804768655168 async_checkpointer.py:561] Finished blocking save. Time taken: 1.891632s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0.
I0423 03:53:21.484490 138804768655168 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 03:53:21.484722 138773411137088 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 03:53:21.484825 138804768655168 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776916399.5924747, 'wait_for_prev_duration_secs': 4.1961669921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776916399.5926561, 'checkpointer_blocking_duration_secs': 1.8917632102966309, 'get_old_steps_start_time': 1776916401.4844358, 'get_old_steps_duration_secs': 1.9788742065429688e-05, 'checkpoint_manager_blocking_start_time': 1776916399.592392, 'checkpoint_manager_blocking_duration_secs': 1.8924064636230469}
I0423 03:53:21.484975 138804768655168 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0423 03:53:21.485023 138804768655168 max_utils.py:750] 
Memstats: After params initialized:
I0423 03:53:21.485082 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0))
I0423 03:53:21.485110 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0423 03:53:21.485131 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0423 03:53:21.485149 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0423 03:53:21.485167 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0423 03:53:21.485185 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0423 03:53:21.485202 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0423 03:53:21.485220 138804768655168 max_utils.py:756] 	Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
I0423 03:53:21.738875 138804768655168 metric_logger.py:196] completed step: 0, seconds: 2.443, TFLOP/s/device: 5.562, Tokens/s/device: 838.313, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941
I0423 03:53:21.739848 138804768655168 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/tensorboard/'
I0423 03:53:21.855262 138804768655168 metric_logger.py:196] completed step: 1, seconds: 2.262, TFLOP/s/device: 6.008, Tokens/s/device: 905.523, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941
I0423 03:53:21.974688 138804768655168 metric_logger.py:196] completed step: 2, seconds: 0.007, TFLOP/s/device: 2049.341, Tokens/s/device: 308898.944, total_weights: 16384, loss: 10.276, lm_loss: 10.276, perplexity: 29019.973
I0423 03:53:22.093779 138804768655168 metric_logger.py:196] completed step: 3, seconds: 0.118, TFLOP/s/device: 114.818, Tokens/s/device: 17306.652, total_weights: 16384, loss: 9.282, lm_loss: 9.282, perplexity: 10739.224
I0423 03:53:22.103377 138804768655168 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0423 03:53:22.256872 138773369194048 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_0
I0423 03:53:22.332995 138804768655168 checkpointing.py:776] Waited 0.22955083847045898 seconds for step 5 to finish before starting checkpointing.
I0423 03:53:22.336531 138804768655168 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 03:53:28.692228 138772761019968 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 7.208851s (commit=6.753195s, array_metadata_write=0.455656s)
I0423 03:53:28.693244 138772559693376 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.369 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 9 seconds) (per-host)
I0423 03:53:28.693345 138772559693376 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 7.208990s.
I0423 03:53:28.960907 138772559693376 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776916400446485558, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
W0423 03:53:29.055379 138772559693376 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/items, 
I0423 03:53:29.193643 138772559693376 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0423 03:53:29.193756 138772559693376 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.193743s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/items
I0423 03:53:29.194374 138772559693376 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/items
I0423 03:53:29.368373 138773358708288 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776916400446485558, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0423 03:53:29.428758 138772559693376 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0
I0423 03:53:30.115231 138772559693376 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0`.
I0423 03:53:30.115840 138772559693376 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 10.523172s. directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/0
I0423 03:53:30.115926 138772559693376 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.631573s.
I0423 03:53:30.116088 138773411137088 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 03:53:30.116196 138773411137088 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 03:53:30.116261 138773411137088 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0423 03:53:30.116307 138773411137088 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0423 03:53:30.116528 138804768655168 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0423 03:53:30.116646 138804768655168 checkpoint_manager.py:1441] Waiting for previous save to complete took 7.780138 seconds. If this number is high, consider checkpointing less frequently.
I0423 03:53:30.117901 138804768655168 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5
I0423 03:53:30.118205 138804768655168 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5.
I0423 03:53:30.282754 138804768655168 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0423 03:53:30.282845 138804768655168 replica_slices.py:410] 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
I0423 03:53:30.360373 138772559693376 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5
I0423 03:53:31.005511 138773369194048 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5/items
I0423 03:53:34.642763 138804768655168 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 4.360402s
I0423 03:53:34.647417 138804768655168 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.778 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host)
I0423 03:53:34.647483 138804768655168 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.442854s (batch_requests_ready=0.074476s, total_serialization_initiated=4.363851s, others=0.004528s)
I0423 03:53:34.647586 138804768655168 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.443383s (all_items=0.000013s, per_item={'items': '0.00001287'}, temp_paths=4.443371)
I0423 03:53:34.649112 138773411137088 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0423 03:53:34.649268 138804768655168 async_checkpointer.py:561] Finished blocking save. Time taken: 4.531323s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5.
I0423 03:53:34.649480 138804768655168 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 03:53:34.649621 138772559693376 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 03:53:34.649711 138804768655168 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776916402.3364933, 'wait_for_prev_duration_secs': 7.780137538909912, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776916410.117925, 'checkpointer_blocking_duration_secs': 4.531466245651245, 'get_old_steps_start_time': 1776916414.6494114, 'get_old_steps_duration_secs': 2.5987625122070312e-05, 'checkpoint_manager_blocking_start_time': 1776916402.3360257, 'checkpoint_manager_blocking_duration_secs': 12.313659429550171}
I0423 03:53:34.649889 138804768655168 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0423 03:53:34.650523 138804768655168 metric_logger.py:196] completed step: 4, seconds: 0.120, TFLOP/s/device: 113.361, Tokens/s/device: 17087.029, total_weights: 16384, loss: 8.327, lm_loss: 8.327, perplexity: 4134.812
I0423 03:53:34.659247 138804768655168 metric_logger.py:196] completed step: 5, seconds: 0.122, TFLOP/s/device: 111.745, Tokens/s/device: 16843.352, total_weights: 16384, loss: 7.367, lm_loss: 7.367, perplexity: 1583.522
I0423 03:53:34.775398 138804768655168 metric_logger.py:196] completed step: 6, seconds: 12.554, TFLOP/s/device: 1.082, Tokens/s/device: 163.137, total_weights: 16384, loss: 6.459, lm_loss: 6.459, perplexity: 638.363
I0423 03:53:34.911645 138804768655168 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2232.155, Tokens/s/device: 336454.740, total_weights: 16384, loss: 5.622, lm_loss: 5.622, perplexity: 276.577
I0423 03:53:35.014563 138804768655168 metric_logger.py:196] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.609, Tokens/s/device: 17425.783, total_weights: 16384, loss: 4.864, lm_loss: 4.864, perplexity: 129.523
I0423 03:53:35.133101 138804768655168 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0423 03:53:35.134392 138804768655168 checkpointing.py:776] Waited 0.0013418197631835938 seconds for step 9 to finish before starting checkpointing.
I0423 03:53:35.134720 138804768655168 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 03:53:35.437513 138773209810496 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_0
I0423 03:53:41.796704 138773369194048 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 7.148472s (commit=6.714437s, array_metadata_write=0.434035s)
I0423 03:53:41.797698 138773411137088 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.064 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host)
I0423 03:53:41.797812 138773411137088 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 7.148502s.
W0423 03:53:42.117845 138773411137088 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5/items, 
I0423 03:53:42.247168 138773411137088 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0423 03:53:42.247290 138773411137088 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.178378s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5/items
I0423 03:53:42.247936 138773411137088 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5/items
I0423 03:53:42.471103 138773411137088 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5
I0423 03:53:43.145611 138773411137088 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5`.
I0423 03:53:43.146363 138773411137088 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 13.028424s. directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/5
I0423 03:53:43.146430 138773411137088 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.497122s.
I0423 03:53:43.146537 138772559693376 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 03:53:43.146583 138772559693376 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 03:53:43.146644 138772559693376 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0423 03:53:43.146682 138772559693376 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0423 03:53:43.146836 138804768655168 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5.
W0423 03:53:43.146949 138804768655168 checkpoint_manager.py:1441] Waiting for previous save to complete took 8.012248 seconds. If this number is high, consider checkpointing less frequently.
I0423 03:53:43.147246 138804768655168 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0423 03:53:43.147528 138804768655168 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9.
I0423 03:53:43.294709 138773411137088 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9
I0423 03:53:43.302635 138804768655168 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0423 03:53:43.302730 138804768655168 replica_slices.py:410] 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
I0423 03:53:43.938316 138773209810496 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9/items
I0423 03:53:48.528096 138804768655168 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 5.225779s
I0423 03:53:48.532517 138804768655168 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.322 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 5 seconds) (per-host)
I0423 03:53:48.532580 138804768655168 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 5.315144s (batch_requests_ready=0.081985s, total_serialization_initiated=5.228839s, others=0.004320s)
I0423 03:53:48.532651 138804768655168 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 5.315642s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=5.315631)
I0423 03:53:48.534284 138773411137088 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0423 03:53:48.534382 138804768655168 async_checkpointer.py:561] Finished blocking save. Time taken: 5.387091s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9.
I0423 03:53:48.534550 138804768655168 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0423 03:53:48.534734 138773369194048 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0423 03:53:48.534834 138804768655168 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776916415.1346867, 'wait_for_prev_duration_secs': 8.012248277664185, 'time_between_consecutive_saves_sec': 5.018364191055298, 'checkpointer_blocking_start_time': 1776916423.1472714, 'checkpointer_blocking_duration_secs': 5.387220144271851, 'get_old_steps_start_time': 1776916428.5344996, 'get_old_steps_duration_secs': 1.9311904907226562e-05, 'checkpoint_manager_blocking_start_time': 1776916415.1346502, 'checkpoint_manager_blocking_duration_secs': 13.40015959739685}
I0423 03:53:48.534945 138804768655168 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0423 03:53:48.534978 138804768655168 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0423 03:53:49.470388 138773199324736 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_0
I0423 03:53:56.445670 138773209810496 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 7.912449s (commit=7.479539s, array_metadata_write=0.432910s)
I0423 03:53:56.446622 138773411137088 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 955.227 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 13 seconds) (per-host)
I0423 03:53:56.446728 138773411137088 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 7.912302s.
W0423 03:53:56.788509 138773411137088 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9/items, 
I0423 03:53:56.911849 138773411137088 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0423 03:53:56.911962 138773411137088 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.179887s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9/items
I0423 03:53:56.912617 138773411137088 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9/items
I0423 03:53:57.137826 138773411137088 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9
I0423 03:53:57.831681 138773411137088 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9`.
I0423 03:53:57.832367 138773411137088 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 14.685085s. directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_15_ocdbt_false/checkpoints/9
I0423 03:53:57.832430 138773411137088 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 9.298005s.
I0423 03:53:57.832604 138773369194048 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0423 03:53:57.832721 138773369194048 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0423 03:53:57.832763 138773369194048 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0423 03:53:57.832798 138773369194048 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0423 03:53:57.832920 138804768655168 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0423 03:53:57.833057 138804768655168 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0423 03:53:57.833879 138804768655168 metric_logger.py:196] completed step: 9, seconds: 0.151, TFLOP/s/device: 90.023, Tokens/s/device: 13569.294, total_weights: 16384, loss: 4.189, lm_loss: 4.189, perplexity: 65.979
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops