2026-04-23 04:08:26.087197: 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 00:00:1776917306.100807 2411274 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776917306.104813 2411274 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776917306.117333 2411274 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776917306.117350 2411274 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776917306.117352 2411274 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776917306.117353 2411274 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 04:08:29.771917: 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 04:08:29.907389 130783241819968 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0423 04:09:26.025897 130783241819968 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 04:09:26.026149 130783241819968 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 04:09:26.026199 130783241819968 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 04:09:26.026228 130783241819968 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system I0423 04:09:26.026253 130783241819968 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 04:09:26.026428 130783241819968 train.py:672] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0423 04:09:26.688065 130783241819968 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 04:09:26.688314 130783241819968 checkpointing.py:677] Setting up checkpoint logger... I0423 04:09:26.688357 130783241819968 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0423 04:09:26.688392 130783241819968 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0423 04:09:26.688812 130783241819968 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x76f20f1b5370>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 04:09:28.367223 130783241819968 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0423 04:09:28.367580 130783241819968 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 0x76f03a3597f0>}, handler_registry=None I0423 04:09:28.367802 130783241819968 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76f03a3597f0>` 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 04:09:28.367843 130783241819968 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76f03a359c10>` 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 04:09:28.367871 130783241819968 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 0x76f03a3597f0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76f03a3597f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76f03a359c10>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76f03a359c10>}). I0423 04:09:28.368144 130783241819968 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0423 04:09:28.368194 130783241819968 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x76f03a3684a0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 04:09:28.811765 130753591510592 checkpoint.py:228] Read Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776917156924988442, 'commit_timestamp_nsecs': 1776917197167989815, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0423 04:09:28.892071 130753581024832 checkpoint_manager.py:1847] Missing metrics for step 5 I0423 04:09:28.942986 130783241819968 checkpoint_manager.py:1788] Found 3 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints I0423 04:09:28.943485 130783241819968 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_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x76f03a359b20> I0423 04:09:28.943586 130783241819968 checkpointing.py:301] Checkpoint manager created! I0423 04:09:29.330451 130783241819968 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 04:09:29.330542 130783241819968 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 04:09:29.612902 130783241819968 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 04:09:29.612982 130783241819968 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 04:09:29.625478 130783241819968 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 04:09:29.625534 130783241819968 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 04:09:29.644553 130783241819968 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 04:09:29.644613 130783241819968 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 04:09:29.657108 130783241819968 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 04:09:29.657156 130783241819968 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 04:09:29.669468 130783241819968 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 04:09:29.669515 130783241819968 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 04:09:29.689018 130783241819968 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 04:09:29.689088 130783241819968 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 04:09:29.704548 130783241819968 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 04:09:29.704602 130783241819968 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0423 04:09:29.855317 130783241819968 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0423 04:09:29.855407 130783241819968 checkpointing.py:581] restoring from this run's directory step 9 I0423 04:09:30.026357 130783241819968 checkpointer.py:304] Restoring checkpoint from gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/9. I0423 04:09:30.601291 2412844 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0423 04:09:30.601351 2412844 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0423 04:09:37.811033 130783241819968 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/read/gbytes_per_sec: 1.657 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host) I0423 04:09:37.811299 130783241819968 checkpointer.py:318] Finished restoring checkpoint in 7.87 seconds from gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/9. I0423 04:09:37.811369 130783241819968 standard_logger.py:34] {'step': 9, 'event_type': 'restore', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints', 'checkpointer_start_time': 1776917369.9416597, 'checkpointer_duration_secs': 7.869687080383301, 'checkpoint_manager_start_time': 1776917369.8556101, 'checkpoint_manager_duration_secs': 7.955738067626953} [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 04:09:38.171312 130783241819968 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 04:09:38.171421 130783241819968 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 04:09:38.171461 130783241819968 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 04:09:38.171507 130783241819968 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 04:09:38.171549 130783241819968 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 04:09:38.171579 130783241819968 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 04:09:38.171620 130783241819968 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 04:09:38.171661 130783241819968 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 04:09:38.171690 130783241819968 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 04:09:38.171716 130783241819968 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 04:09:38.171751 130783241819968 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0423 04:09:38.171786 130783241819968 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0423 04:09:38.572195 130783241819968 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 04:09:38.572276 130783241819968 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0423 04:09:38.583992 130783241819968 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 04:09:38.584053 130783241819968 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0423 04:09:39.355123 130783241819968 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 04:09:39.355648 130783241819968 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0423 04:09:39.355813 130783241819968 metric_logger.py:301] number parameters: 1.104 billion I0423 04:09:41.278634 130783241819968 checkpointing.py:772] Waiting for step 10 to finish before checkpoint... I0423 04:09:41.395743 130783241819968 checkpointing.py:776] Waited 0.11710047721862793 seconds for step 10 to finish before starting checkpointing. I0423 04:09:41.396326 130783241819968 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 04:09:41.396473 130783241819968 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 10 I0423 04:09:41.396766 130783241819968 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_16_zarr3_false/checkpoints/10. I0423 04:09:41.481460 130783241819968 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0423 04:09:41.569171 130783241819968 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 04:09:41.569258 130783241819968 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 04:09:41.575215 130751578244672 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_16_zarr3_false/checkpoints/10 I0423 04:09:42.323762 130751567758912 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_16_zarr3_false/checkpoints/10/items I0423 04:09:42.487922 130751181882944 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776917382239994313, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776917382239994313, "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_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA I0423 04:09:42.504524 130783241819968 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.935784s I0423 04:09:42.504887 130783241819968 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 12.069 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0423 04:09:42.504937 130783241819968 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.022561s (batch_requests_ready=0.083654s, total_serialization_initiated=0.938604s, others=0.000304s) I0423 04:09:42.504996 130783241819968 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.023083s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=1.023063) I0423 04:09:42.505763 130750961681984 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 04:09:42.505856 130783241819968 async_checkpointer.py:561] Finished blocking save. Time taken: 1.109344s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/10. I0423 04:09:42.506029 130783241819968 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0423 04:09:42.506218 130751588730432 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 04:09:42.506318 130783241819968 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776917381.3963127, 'wait_for_prev_duration_secs': 4.100799560546875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776917381.3964958, 'checkpointer_blocking_duration_secs': 1.1094636917114258, 'get_old_steps_start_time': 1776917382.5059724, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1776917381.396231, 'checkpoint_manager_blocking_duration_secs': 1.110062599182129} I0423 04:09:42.506422 130783241819968 checkpointing.py:408] Started an asynchronous checkpoint save for step 10 I0423 04:09:42.506460 130783241819968 max_utils.py:750] Memstats: After params initialized: I0423 04:09:42.506539 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0423 04:09:42.506565 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0423 04:09:42.506585 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0423 04:09:42.506609 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0423 04:09:42.506626 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0423 04:09:42.506643 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0423 04:09:42.506659 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0423 04:09:42.506675 130783241819968 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0423 04:09:42.753335 130783241819968 metric_logger.py:196] completed step: 10, seconds: 1.922, TFLOP/s/device: 7.068, Tokens/s/device: 1065.392, total_weights: 16384, loss: 3.602, lm_loss: 3.602, perplexity: 36.686 I0423 04:09:42.870342 130783241819968 metric_logger.py:196] completed step: 11, seconds: 1.473, TFLOP/s/device: 9.226, Tokens/s/device: 1390.669, total_weights: 16384, loss: 3.101, lm_loss: 3.101, perplexity: 22.211 I0423 04:09:42.989642 130783241819968 metric_logger.py:196] completed step: 12, seconds: 0.006, TFLOP/s/device: 2214.330, Tokens/s/device: 333767.927, total_weights: 16384, loss: 2.684, lm_loss: 2.684, perplexity: 14.644 I0423 04:09:43.109027 130783241819968 metric_logger.py:196] completed step: 13, seconds: 0.118, TFLOP/s/device: 114.718, Tokens/s/device: 17291.602, total_weights: 16384, loss: 2.348, lm_loss: 2.348, perplexity: 10.461 I0423 04:09:43.114346 130783241819968 checkpointing.py:772] Waiting for step 15 to finish before checkpoint... I0423 04:09:43.347416 130783241819968 checkpointing.py:776] Waited 0.23303747177124023 seconds for step 15 to finish before starting checkpointing. I0423 04:09:43.348492 130783241819968 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 04:09:43.562693 130751578244672 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_16_zarr3_false/checkpoints/10/items/array_metadatas/process_0 I0423 04:10:15.751269 130751418861120 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 32.950990s (commit=32.460490s, array_metadata_write=0.490500s) I0423 04:10:15.752315 130750961681984 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 368.743 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 34 seconds) (per-host) I0423 04:10:15.752376 130750961681984 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 33.246477s. I0423 04:10:16.286150 130750961681984 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 04:10:16.468908 130751181882944 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776917382239994313, '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_16_zarr3_false/checkpoints/10/_CHECKPOINT_METADATA I0423 04:10:16.745365 130750961681984 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.658973s. use_zarr3=False, 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_16_zarr3_false/checkpoints/10/items I0423 04:10:16.746068 130750961681984 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/10/items I0423 04:10:16.995543 130750961681984 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/10 I0423 04:10:17.781743 130750961681984 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_16_zarr3_false/checkpoints/10`. I0423 04:10:17.782379 130750961681984 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 36.385872s. directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/10 I0423 04:10:17.782460 130750961681984 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 35.276562s. I0423 04:10:17.782670 130751588730432 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 04:10:17.782773 130751588730432 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 04:10:17.782839 130751588730432 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0423 04:10:17.782885 130751588730432 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0423 04:10:17.783039 130783241819968 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0423 04:10:17.783182 130783241819968 checkpoint_manager.py:1441] Waiting for previous save to complete took 34.434685 seconds. If this number is high, consider checkpointing less frequently. I0423 04:10:17.784548 130783241819968 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 15 I0423 04:10:17.784839 130783241819968 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_16_zarr3_false/checkpoints/15. I0423 04:10:17.968658 130783241819968 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 04:10:17.968751 130783241819968 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 04:10:17.970099 130750961681984 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_16_zarr3_false/checkpoints/15 I0423 04:10:18.737120 130751578244672 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_16_zarr3_false/checkpoints/15/items I0423 04:10:20.285481 130783241819968 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.317204s I0423 04:10:20.289255 130783241819968 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.123 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0423 04:10:20.289322 130783241819968 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.408834s (batch_requests_ready=0.084364s, total_serialization_initiated=2.320797s, others=0.003672s) I0423 04:10:20.289394 130783241819968 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.409384s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=2.409368) I0423 04:10:20.290938 130751588730432 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 04:10:20.291019 130783241819968 async_checkpointer.py:561] Finished blocking save. Time taken: 2.506427s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/15. I0423 04:10:20.291224 130783241819968 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=15] Starting CheckpointManager Save Finalize thread=save_finalize I0423 04:10:20.291378 130750961681984 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 04:10:20.291479 130783241819968 standard_logger.py:34] {'step': 15, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776917383.3484666, 'wait_for_prev_duration_secs': 34.43468523025513, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776917417.7845726, 'checkpointer_blocking_duration_secs': 2.506561756134033, 'get_old_steps_start_time': 1776917420.2911525, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776917383.3484044, 'checkpoint_manager_blocking_duration_secs': 36.94304823875427} I0423 04:10:20.291614 130783241819968 checkpointing.py:408] Started an asynchronous checkpoint save for step 15 I0423 04:10:20.292271 130783241819968 metric_logger.py:196] completed step: 14, seconds: 0.119, TFLOP/s/device: 114.037, Tokens/s/device: 17188.851, total_weights: 16384, loss: 2.083, lm_loss: 2.083, perplexity: 8.030 I0423 04:10:20.301323 130783241819968 metric_logger.py:196] completed step: 15, seconds: 0.119, TFLOP/s/device: 114.059, Tokens/s/device: 17192.170, total_weights: 16384, loss: 1.880, lm_loss: 1.880, perplexity: 6.551 I0423 04:10:20.417856 130783241819968 metric_logger.py:196] completed step: 16, seconds: 37.185, TFLOP/s/device: 0.365, Tokens/s/device: 55.076, total_weights: 16384, loss: 1.726, lm_loss: 1.726, perplexity: 5.616 I0423 04:10:20.536906 130783241819968 metric_logger.py:196] completed step: 17, seconds: 0.006, TFLOP/s/device: 2170.815, Tokens/s/device: 327208.819, total_weights: 16384, loss: 1.610, lm_loss: 1.610, perplexity: 5.002 I0423 04:10:20.655711 130783241819968 metric_logger.py:196] completed step: 18, seconds: 0.118, TFLOP/s/device: 115.505, Tokens/s/device: 17410.229, total_weights: 16384, loss: 1.521, lm_loss: 1.521, perplexity: 4.577 I0423 04:10:20.773720 130783241819968 checkpointing.py:772] Waiting for step 19 to finish before checkpoint... I0423 04:10:20.774710 130783241819968 checkpointing.py:776] Waited 0.0010178089141845703 seconds for step 19 to finish before starting checkpointing. I0423 04:10:20.774960 130783241819968 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=15][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 04:10:21.076895 130753411155520 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_16_zarr3_false/checkpoints/15/items/array_metadatas/process_0 I0423 04:10:59.882802 130751578244672 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 39.592726s (commit=39.083080s, array_metadata_write=0.509646s) I0423 04:10:59.891434 130751588730432 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 300.798 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 42 seconds) (per-host) I0423 04:10:59.891540 130751588730432 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.600466s. I0423 04:11:00.378121 130751588730432 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 04:11:00.880356 130751588730432 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.662224s. use_zarr3=False, 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_16_zarr3_false/checkpoints/15/items I0423 04:11:00.881155 130751588730432 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/15/items I0423 04:11:01.131231 130751588730432 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/15 I0423 04:11:01.832599 130751588730432 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_16_zarr3_false/checkpoints/15`. I0423 04:11:01.833225 130751588730432 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 44.048640s. directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/15 I0423 04:11:01.833300 130751588730432 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 41.542228s. I0423 04:11:01.833504 130750961681984 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 04:11:01.833607 130750961681984 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 04:11:01.833675 130750961681984 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is syncing with other hosts... I0423 04:11:01.833719 130750961681984 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=15] CheckpointManager Save Finalize is done on all hosts. I0423 04:11:01.833891 130783241819968 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=15][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=15. W0423 04:11:01.834006 130783241819968 checkpoint_manager.py:1441] Waiting for previous save to complete took 41.059056 seconds. If this number is high, consider checkpointing less frequently. I0423 04:11:01.834431 130783241819968 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 19 I0423 04:11:01.834745 130783241819968 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_16_zarr3_false/checkpoints/19. I0423 04:11:02.003472 130783241819968 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 04:11:02.003570 130783241819968 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 04:11:02.004902 130751588730432 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_16_zarr3_false/checkpoints/19 I0423 04:11:02.731554 130753411155520 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_16_zarr3_false/checkpoints/19/items I0423 04:11:06.053763 130783241819968 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 4.050680s I0423 04:11:06.057422 130783241819968 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.980 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host) I0423 04:11:06.057482 130783241819968 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.141109s (batch_requests_ready=0.082907s, total_serialization_initiated=4.054647s, others=0.003555s) I0423 04:11:06.057556 130783241819968 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.141613s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=4.141598) I0423 04:11:06.058677 130751578244672 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 04:11:06.058769 130783241819968 async_checkpointer.py:561] Finished blocking save. Time taken: 4.224295s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/19. I0423 04:11:06.058943 130783241819968 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0423 04:11:06.059182 130753184663104 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 04:11:06.059284 130783241819968 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776917420.774934, 'wait_for_prev_duration_secs': 41.0590558052063, 'time_between_consecutive_saves_sec': 2.9920334815979004, 'checkpointer_blocking_start_time': 1776917461.834458, 'checkpointer_blocking_duration_secs': 4.224412441253662, 'get_old_steps_start_time': 1776917466.0588822, 'get_old_steps_duration_secs': 2.6702880859375e-05, 'checkpoint_manager_blocking_start_time': 1776917420.7749023, 'checkpoint_manager_blocking_duration_secs': 45.28435969352722} I0423 04:11:06.059405 130783241819968 checkpointing.py:408] Started an asynchronous checkpoint save for step 19 I0423 04:11:06.059437 130783241819968 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 04:11:06.818492 130753400669760 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_16_zarr3_false/checkpoints/19/items/array_metadatas/process_0 I0423 04:11:45.557980 130753230800448 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 39.499925s (commit=39.023814s, array_metadata_write=0.476111s) I0423 04:11:45.559005 130751578244672 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 289.552 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 43 seconds) (per-host) I0423 04:11:45.559127 130751578244672 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 39.500316s. I0423 04:11:46.034660 130751578244672 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 04:11:46.582461 130751578244672 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.713016s. use_zarr3=False, 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_16_zarr3_false/checkpoints/19/items I0423 04:11:46.583340 130751578244672 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/19/items I0423 04:11:46.831178 130751578244672 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/19 I0423 04:11:47.557874 130751578244672 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_16_zarr3_false/checkpoints/19`. I0423 04:11:47.558487 130751578244672 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 45.724018s. directory=gs://wanglance-maxtext/linen_ckpt_fix_gate_fp8_nan_sanitization_20260423_024158/linen_fix_gate_fp8_nan_sanitization_20260423_024158_16_zarr3_false/checkpoints/19 I0423 04:11:47.558549 130751578244672 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 41.499739s. I0423 04:11:47.558705 130753184663104 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 04:11:47.558816 130753184663104 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 04:11:47.558878 130753184663104 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0423 04:11:47.558919 130753184663104 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0423 04:11:47.559038 130783241819968 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0423 04:11:47.559168 130783241819968 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 04:11:47.559961 130783241819968 metric_logger.py:196] completed step: 19, seconds: 0.119, TFLOP/s/device: 114.004, Tokens/s/device: 17183.947, total_weights: 16384, loss: 1.449, lm_loss: 1.449, perplexity: 4.260 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops