2026-04-21 15:43:49.801533: 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 00:00:1776786229.811435 1532966 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776786229.814455 1532966 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776786229.824304 1532966 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776786229.824314 1532966 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776786229.824316 1532966 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776786229.824317 1532966 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-21 15:43:53.218150: 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) I0421 15:43:53.334478 129501715967104 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0421 15:44:20.573007 129501715967104 max_utils.py:800] System Information: Jax Version: 0.8.1 I0421 15:44:20.573228 129501715967104 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0421 15:44:20.573275 129501715967104 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 I0421 15:44:20.573306 129501715967104 train_utils.py:335] WARNING: 'dataset_path' might be pointing your local file system I0421 15:44:20.573328 129501715967104 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 15:44:20.573427 129501715967104 train.py:703] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0421 15:44:20.934522 129501715967104 maxtext_utils.py:1551] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 15:44:20.934890 129501715967104 checkpointing.py:677] Setting up checkpoint logger... I0421 15:44:20.934938 129501715967104 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0421 15:44:20.935025 129501715967104 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0421 15:44:20.935856 129501715967104 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x75c7ab7a7140>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 15:44:23.353953 129501715967104 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 15:44:23.354345 129501715967104 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 0x75c604b0ac00>}, handler_registry=None I0421 15:44:23.354774 129501715967104 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x75c604b0ac00>` 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`. I0421 15:44:23.354817 129501715967104 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x75c603b82330>` 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`. I0421 15:44:23.354852 129501715967104 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 0x75c604b0ac00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x75c604b0ac00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x75c603b82330>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x75c603b82330>}). I0421 15:44:23.355230 129501715967104 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0421 15:44:23.355300 129501715967104 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x75c603ba9260> timeout: 600 secs and primary_host=0 for async checkpoint writes I0421 15:44:23.484125 129501715967104 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints I0421 15:44:23.484396 129501715967104 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_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x75c603b81e50> I0421 15:44:23.484502 129501715967104 checkpointing.py:301] Checkpoint manager created! I0421 15:44:24.118013 129501715967104 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 15:44:24.118120 129501715967104 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0421 15:44:24.398302 129501715967104 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 15:44:24.398386 129501715967104 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0421 15:44:24.410176 129501715967104 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 15:44:24.410228 129501715967104 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0421 15:44:24.428937 129501715967104 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 15:44:24.428998 129501715967104 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:44:24.440893 129501715967104 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 15:44:24.440943 129501715967104 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:44:24.452728 129501715967104 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 15:44:24.452777 129501715967104 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:44:24.471634 129501715967104 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 15:44:24.471688 129501715967104 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:44:24.486702 129501715967104 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 15:44:24.486767 129501715967104 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0421 15:44:24.633439 129501715967104 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 15:44:24.633535 129501715967104 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 I0421 15:44:25.175759 129501715967104 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0421 15:44:25.175867 129501715967104 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 15:44:25.175907 129501715967104 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0421 15:44:25.175950 129501715967104 maxtext_utils.py:1654] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0421 15:44:25.175995 129501715967104 maxtext_utils.py:1654] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0421 15:44:25.176023 129501715967104 maxtext_utils.py:1654] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0421 15:44:25.176076 129501715967104 maxtext_utils.py:1654] 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) I0421 15:44:25.176117 129501715967104 maxtext_utils.py:1654] 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') I0421 15:44:25.176146 129501715967104 maxtext_utils.py:1654] 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) I0421 15:44:25.176172 129501715967104 maxtext_utils.py:1654] 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) I0421 15:44:25.176207 129501715967104 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 15:44:25.176241 129501715967104 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 15:44:25.585277 129501715967104 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0421 15:44:25.585370 129501715967104 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0421 15:44:25.597167 129501715967104 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0421 15:44:25.597215 129501715967104 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0421 15:44:26.333161 129501715967104 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. I0421 15:44:26.333743 129501715967104 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0421 15:44:26.333902 129501715967104 metric_logger.py:301] number parameters: 1.104 billion I0421 15:44:28.404548 129501715967104 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 15:44:28.521701 129501715967104 checkpointing.py:776] Waited 0.11713385581970215 seconds for step 0 to finish before starting checkpointing. I0421 15:44:28.522199 129501715967104 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 15:44:28.522350 129501715967104 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0421 15:44:28.522759 129501715967104 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0. I0421 15:44:28.603392 129501715967104 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0421 15:44:28.697760 129501715967104 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0421 15:44:28.697926 129417592440384 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0 I0421 15:44:28.698309 129501715967104 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 I0421 15:44:29.485581 129417581954624 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/items I0421 15:44:29.649929 129417546303040 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776786269398630722, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776786269398630722, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0421 15:44:29.973085 1534949 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0421 15:44:29.973144 1534949 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0421 15:44:30.785776 129501715967104 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.088828s I0421 15:44:30.790452 129501715967104 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.646 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0421 15:44:30.790520 129501715967104 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.185675s (batch_requests_ready=0.088715s, total_serialization_initiated=2.092395s, others=0.004565s) I0421 15:44:30.790600 129501715967104 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.186666s (all_items=0.000032s, per_item={'items': '0.00003171'}, temp_paths=2.186635) I0421 15:44:30.791975 129417602926144 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0421 15:44:30.792071 129501715967104 async_checkpointer.py:561] Finished blocking save. Time taken: 2.269682s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0. I0421 15:44:30.792252 129501715967104 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:44:30.792452 129417483388480 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:44:30.792566 129501715967104 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786268.5221863, 'wait_for_prev_duration_secs': 4.100799560546875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776786268.522371, 'checkpointer_blocking_duration_secs': 2.269806146621704, 'get_old_steps_start_time': 1776786270.7921963, 'get_old_steps_duration_secs': 2.193450927734375e-05, 'checkpoint_manager_blocking_start_time': 1776786268.5221002, 'checkpoint_manager_blocking_duration_secs': 2.270439863204956} I0421 15:44:30.792677 129501715967104 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 15:44:30.792715 129501715967104 max_utils.py:750] Memstats: After params initialized: I0421 15:44:30.792761 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0421 15:44:30.792785 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0421 15:44:30.792805 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0421 15:44:30.792822 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0421 15:44:30.792838 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0421 15:44:30.792855 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0421 15:44:30.792872 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0421 15:44:30.792889 129501715967104 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0421 15:44:31.077724 129501715967104 metric_logger.py:196] completed step: 0, seconds: 2.070, TFLOP/s/device: 6.564, Tokens/s/device: 989.325, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0421 15:44:31.078543 129501715967104 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/tensorboard/' I0421 15:44:31.194797 129501715967104 metric_logger.py:196] completed step: 1, seconds: 2.671, TFLOP/s/device: 5.086, Tokens/s/device: 766.669, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0421 15:44:31.313866 129501715967104 metric_logger.py:196] completed step: 2, seconds: 0.006, TFLOP/s/device: 2306.030, Tokens/s/device: 347589.952, total_weights: 16384, loss: 10.276, lm_loss: 10.276, perplexity: 29019.973 I0421 15:44:31.432757 129501715967104 metric_logger.py:196] completed step: 3, seconds: 0.118, TFLOP/s/device: 115.401, Tokens/s/device: 17394.554, total_weights: 16384, loss: 9.282, lm_loss: 9.282, perplexity: 10739.224 I0421 15:44:31.437685 129501715967104 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0421 15:44:31.538580 129417560983104 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_0 I0421 15:44:31.670135 129501715967104 checkpointing.py:776] Waited 0.23241472244262695 seconds for step 5 to finish before starting checkpointing. I0421 15:44:31.670532 129501715967104 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:45:00.718449 129417504360000 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.927173s (commit=29.501037s, array_metadata_write=0.426135s) I0421 15:45:00.719665 129417602926144 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 393.488 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 32 seconds) (per-host) I0421 15:45:00.719723 129417602926144 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.927608s. I0421 15:45:00.947009 129417602926144 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776786269398630722, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0421 15:45:01.141750 129417602926144 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0421 15:45:01.315982 129417546303040 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776786269398630722, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/_CHECKPOINT_METADATA I0421 15:45:01.561334 129417602926144 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0421 15:45:01.561933 129417602926144 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.570790s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/items I0421 15:45:01.562687 129417602926144 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0/items I0421 15:45:01.824479 129417602926144 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0 I0421 15:45:02.533225 129417602926144 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0`. I0421 15:45:02.533972 129417602926144 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 34.011589s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/0 I0421 15:45:02.534060 129417602926144 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 31.741932s. I0421 15:45:02.534233 129417483388480 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:45:02.534349 129417483388480 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:45:02.534411 129417483388480 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:45:02.534455 129417483388480 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 15:45:02.534644 129501715967104 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. W0421 15:45:02.534785 129501715967104 checkpoint_manager.py:1441] Waiting for previous save to complete took 30.864251 seconds. If this number is high, consider checkpointing less frequently. I0421 15:45:02.535826 129501715967104 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0421 15:45:02.536170 129501715967104 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5. I0421 15:45:02.704683 129417483388480 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5 I0421 15:45:02.711510 129501715967104 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0421 15:45:02.711620 129501715967104 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 I0421 15:45:03.424416 129417472902720 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5/items I0421 15:45:10.952417 129501715967104 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 8.241375s I0421 15:45:10.956751 129501715967104 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.480 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host) I0421 15:45:10.956813 129501715967104 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 8.336415s (batch_requests_ready=0.087003s, total_serialization_initiated=8.245194s, others=0.004218s) I0421 15:45:10.956891 129501715967104 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 8.337034s (all_items=0.000023s, per_item={'items': '0.00002265'}, temp_paths=8.337011) I0421 15:45:10.958657 129417483388480 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0421 15:45:10.958743 129501715967104 async_checkpointer.py:561] Finished blocking save. Time taken: 8.422872s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5. I0421 15:45:10.958930 129501715967104 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:45:10.959111 129417602926144 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:45:10.959216 129501715967104 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786271.67051, 'wait_for_prev_duration_secs': 30.864250898361206, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776786302.5358508, 'checkpointer_blocking_duration_secs': 8.422996997833252, 'get_old_steps_start_time': 1776786310.9588647, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776786271.6704445, 'checkpoint_manager_blocking_duration_secs': 39.28874659538269} I0421 15:45:10.959365 129501715967104 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0421 15:45:10.960017 129501715967104 metric_logger.py:196] completed step: 4, seconds: 0.119, TFLOP/s/device: 113.841, Tokens/s/device: 17159.327, total_weights: 16384, loss: 8.327, lm_loss: 8.327, perplexity: 4134.812 I0421 15:45:10.968008 129501715967104 metric_logger.py:196] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.443, Tokens/s/device: 17250.093, total_weights: 16384, loss: 7.367, lm_loss: 7.367, perplexity: 1583.522 I0421 15:45:11.084456 129501715967104 metric_logger.py:196] completed step: 6, seconds: 39.528, TFLOP/s/device: 0.344, Tokens/s/device: 51.811, total_weights: 16384, loss: 6.459, lm_loss: 6.459, perplexity: 638.363 I0421 15:45:11.203476 129501715967104 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2229.957, Tokens/s/device: 336123.420, total_weights: 16384, loss: 5.622, lm_loss: 5.622, perplexity: 276.577 I0421 15:45:11.322003 129501715967104 metric_logger.py:196] completed step: 8, seconds: 0.117, TFLOP/s/device: 115.751, Tokens/s/device: 17447.309, total_weights: 16384, loss: 4.864, lm_loss: 4.864, perplexity: 129.523 I0421 15:45:11.440110 129501715967104 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 15:45:11.440927 129501715967104 checkpointing.py:776] Waited 0.0008509159088134766 seconds for step 9 to finish before starting checkpointing. I0421 15:45:11.441196 129501715967104 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:45:11.721037 129417560983104 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_0 I0421 15:45:48.959588 129417525331520 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.001880s (commit=37.530407s, array_metadata_write=0.471474s) I0421 15:45:48.960848 129417483388480 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 272.695 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 46 seconds) (per-host) I0421 15:45:48.960911 129417483388480 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.002124s. I0421 15:45:49.362240 129417483388480 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0421 15:45:49.784363 129417483388480 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0421 15:45:49.785057 129417483388480 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.557560s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5/items I0421 15:45:49.785783 129417483388480 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5/items I0421 15:45:50.037917 129417483388480 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5 I0421 15:45:50.716885 129417483388480 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5`. I0421 15:45:50.717641 129417483388480 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 48.181777s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/5 I0421 15:45:50.717720 129417483388480 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 39.758935s. I0421 15:45:50.717968 129417602926144 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:45:50.718104 129417602926144 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:45:50.718174 129417602926144 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:45:50.718223 129417602926144 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 15:45:50.719142 129501715967104 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. W0421 15:45:50.719271 129501715967104 checkpoint_manager.py:1441] Waiting for previous save to complete took 39.278080 seconds. If this number is high, consider checkpointing less frequently. I0421 15:45:50.719522 129501715967104 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0421 15:45:50.719830 129501715967104 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9. I0421 15:45:50.882579 129417602926144 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9 I0421 15:45:50.899680 129501715967104 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0421 15:45:50.899785 129501715967104 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 I0421 15:45:51.608567 129417514845760 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9/items I0421 15:45:56.941660 129501715967104 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 6.042390s I0421 15:45:56.942289 129501715967104 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.011 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 6 seconds) (per-host) I0421 15:45:56.942341 129501715967104 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 6.136954s (batch_requests_ready=0.090625s, total_serialization_initiated=6.045819s, others=0.000510s) I0421 15:45:56.942408 129501715967104 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 6.137537s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=6.137520) I0421 15:45:56.947876 129417483388480 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0421 15:45:56.947981 129501715967104 async_checkpointer.py:561] Finished blocking save. Time taken: 6.228412s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9. I0421 15:45:56.948240 129501715967104 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:45:56.948514 129417493874240 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:45:56.948597 129501715967104 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786311.4411712, 'wait_for_prev_duration_secs': 39.278080463409424, 'time_between_consecutive_saves_sec': 8.906702518463135, 'checkpointer_blocking_start_time': 1776786350.7195437, 'checkpointer_blocking_duration_secs': 6.228625774383545, 'get_old_steps_start_time': 1776786356.948184, 'get_old_steps_duration_secs': 2.2649765014648438e-05, 'checkpoint_manager_blocking_start_time': 1776786311.4411356, 'checkpoint_manager_blocking_duration_secs': 45.50743842124939} I0421 15:45:56.948696 129501715967104 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 15:45:56.948729 129501715967104 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:45:57.730672 129417560983104 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_0 I0421 15:46:24.420750 129417504360000 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 27.473637s (commit=27.012757s, array_metadata_write=0.460880s) I0421 15:46:24.422121 129417483388480 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 375.908 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 33 seconds) (per-host) I0421 15:46:24.422245 129417483388480 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 27.474214s. I0421 15:46:24.823340 129417483388480 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0421 15:46:25.258123 129417483388480 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0421 15:46:25.258323 129417483388480 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.575451s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9/items I0421 15:46:25.259168 129417483388480 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9/items I0421 15:46:25.505735 129417483388480 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9 I0421 15:46:26.193888 129417483388480 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9`. I0421 15:46:26.194662 129417483388480 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 35.475106s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_true/checkpoints/9 I0421 15:46:26.194730 129417483388480 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 29.246700s. I0421 15:46:26.194859 129417493874240 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:46:26.194911 129417493874240 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:46:26.194963 129417493874240 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:46:26.194999 129417493874240 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 15:46:26.195425 129501715967104 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. I0421 15:46:26.196075 129501715967104 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 15:46:26.198486 129501715967104 metric_logger.py:196] completed step: 9, seconds: 0.119, TFLOP/s/device: 114.100, Tokens/s/device: 17198.378, 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