2026-04-21 15:51:12.202632: 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 00:00:1776786672.213611 1538599 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776786672.216775 1538599 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776786672.227362 1538599 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776786672.227373 1538599 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776786672.227375 1538599 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776786672.227377 1538599 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. 2026-04-21 15:51:15.579101: 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 15:51:15.696870 127065353903232 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0421 15:52:09.970093 127065353903232 max_utils.py:800] System Information: Jax Version: 0.8.1 I0421 15:52:09.970370 127065353903232 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0421 15:52:09.970416 127065353903232 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 15:52:09.970448 127065353903232 train_utils.py:335] WARNING: 'dataset_path' might be pointing your local file system I0421 15:52:09.970470 127065353903232 train_utils.py:348] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 15:52:09.970644 127065353903232 train.py:703] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0421 15:52:10.336618 127065353903232 maxtext_utils.py:1551] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 15:52:10.336989 127065353903232 checkpointing.py:677] Setting up checkpoint logger... I0421 15:52:10.337036 127065353903232 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True I0421 15:52:10.337140 127065353903232 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0421 15:52:10.337981 127065353903232 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 0x73906b9a71d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 15:52:12.775032 127065353903232 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0421 15:52:12.775296 127065353903232 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 0x738ec3776750>}, handler_registry=None I0421 15:52:12.775664 127065353903232 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x738ec3776750>` 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 15:52:12.775708 127065353903232 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x738ec37759a0>` 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 15:52:12.775739 127065353903232 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 0x738ec3776750>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x738ec3776750>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x738ec37759a0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x738ec37759a0>}). I0421 15:52:12.776024 127065353903232 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0421 15:52:12.776095 127065353903232 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x738ec3795260> timeout: 600 secs and primary_host=0 for async checkpoint writes I0421 15:52:12.891076 127065353903232 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints I0421 15:52:12.891321 127065353903232 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_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x738ec3774c80> I0421 15:52:12.891415 127065353903232 checkpointing.py:301] Checkpoint manager created! I0421 15:52:13.523192 127065353903232 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 15:52:13.523278 127065353903232 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0421 15:52:13.799715 127065353903232 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 15:52:13.799788 127065353903232 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0421 15:52:13.811775 127065353903232 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 15:52:13.811822 127065353903232 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0421 15:52:13.830568 127065353903232 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 15:52:13.830625 127065353903232 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:52:13.842602 127065353903232 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 15:52:13.842652 127065353903232 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:52:13.854552 127065353903232 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 15:52:13.854603 127065353903232 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:52:13.873643 127065353903232 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 15:52:13.873697 127065353903232 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0421 15:52:13.888761 127065353903232 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 15:52:13.888809 127065353903232 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0421 15:52:14.042026 127065353903232 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0421 15:52:14.042127 127065353903232 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 15:52:14.602483 127065353903232 maxtext_utils.py:1654] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0421 15:52:14.602588 127065353903232 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 15:52:14.602635 127065353903232 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 15:52:14.602679 127065353903232 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 15:52:14.602718 127065353903232 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 15:52:14.602746 127065353903232 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 15:52:14.602783 127065353903232 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 15:52:14.602822 127065353903232 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 15:52:14.602849 127065353903232 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 15:52:14.602875 127065353903232 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 15:52:14.602909 127065353903232 maxtext_utils.py:1654] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 15:52:14.602941 127065353903232 maxtext_utils.py:1654] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 15:52:15.002059 127065353903232 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0421 15:52:15.002138 127065353903232 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0421 15:52:15.013379 127065353903232 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0421 15:52:15.013428 127065353903232 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0421 15:52:15.740795 127065353903232 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 15:52:15.741338 127065353903232 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0421 15:52:15.741485 127065353903232 metric_logger.py:301] number parameters: 1.104 billion I0421 15:52:17.536108 127065353903232 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0421 15:52:17.653968 127065353903232 checkpointing.py:776] Waited 0.11784696578979492 seconds for step 0 to finish before starting checkpointing. I0421 15:52:17.654420 127065353903232 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 15:52:17.654564 127065353903232 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0421 15:52:17.654928 127065353903232 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_false/checkpoints/0. I0421 15:52:17.743305 127065353903232 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0421 15:52:17.834171 127065353903232 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0421 15:52:17.834306 126981190452800 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0 I0421 15:52:17.834675 127065353903232 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 15:52:18.621129 126981179967040 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/items I0421 15:52:18.753119 126981144315456 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776786738515153531, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776786738515153531, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0421 15:52:18.904538 127065353903232 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.071121s I0421 15:52:18.905148 127065353903232 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 10.631 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0421 15:52:18.905208 127065353903232 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.160989s (batch_requests_ready=0.085863s, total_serialization_initiated=1.074666s, others=0.000460s) I0421 15:52:18.905279 127065353903232 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.161539s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=1.161520) I0421 15:52:18.906202 126981123343936 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0421 15:52:18.906369 127065353903232 async_checkpointer.py:561] Finished blocking save. Time taken: 1.251762s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0. I0421 15:52:18.906541 127065353903232 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:52:18.906746 126981200938560 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:52:18.906870 127065353903232 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786737.6544075, 'wait_for_prev_duration_secs': 4.029273986816406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776786737.6545844, 'checkpointer_blocking_duration_secs': 1.2518599033355713, 'get_old_steps_start_time': 1776786738.9064696, 'get_old_steps_duration_secs': 2.9087066650390625e-05, 'checkpoint_manager_blocking_start_time': 1776786737.6543279, 'checkpoint_manager_blocking_duration_secs': 1.2525050640106201} I0421 15:52:18.907015 127065353903232 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0421 15:52:18.907080 127065353903232 max_utils.py:750] Memstats: After params initialized: I0421 15:52:18.907159 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0421 15:52:18.907186 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0421 15:52:18.907206 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0421 15:52:18.907225 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0421 15:52:18.907242 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0421 15:52:18.907259 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0421 15:52:18.907276 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0421 15:52:18.907292 127065353903232 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0421 15:52:19.114679 1540657 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0421 15:52:19.114732 1540657 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0421 15:52:19.157621 127065353903232 metric_logger.py:196] completed step: 0, seconds: 1.794, TFLOP/s/device: 7.573, Tokens/s/device: 1141.515, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0421 15:52:19.158515 127065353903232 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/tensorboard/' I0421 15:52:19.274536 127065353903232 metric_logger.py:196] completed step: 1, seconds: 1.620, TFLOP/s/device: 8.389, Tokens/s/device: 1264.507, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0421 15:52:19.393797 127065353903232 metric_logger.py:196] completed step: 2, seconds: 0.006, TFLOP/s/device: 2324.971, Tokens/s/device: 350444.901, total_weights: 16384, loss: 10.276, lm_loss: 10.276, perplexity: 29019.973 I0421 15:52:19.512642 127065353903232 metric_logger.py:196] completed step: 3, seconds: 0.118, TFLOP/s/device: 115.409, Tokens/s/device: 17395.736, total_weights: 16384, loss: 9.282, lm_loss: 9.282, perplexity: 10739.224 I0421 15:52:19.517516 127065353903232 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0421 15:52:19.909774 127065353903232 checkpointing.py:776] Waited 0.39223694801330566 seconds for step 5 to finish before starting checkpointing. I0421 15:52:19.912060 127065353903232 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:52:20.161396 126981169481280 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_false/checkpoints/0/items/array_metadatas/process_0 I0421 15:52:26.147539 126981190452800 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 7.032977s (commit=6.615249s, array_metadata_write=0.417728s) I0421 15:52:26.148473 126981123343936 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.468 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host) I0421 15:52:26.148575 126981123343936 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 7.242256s. I0421 15:52:26.371109 126981123343936 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776786738515153531, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA W0421 15:52:26.461672 126981123343936 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_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/items, I0421 15:52:26.553309 126981123343936 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 15:52:26.553414 126981123343936 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.146651s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/items I0421 15:52:26.554090 126981123343936 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/items I0421 15:52:26.747678 126981144315456 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776786738515153531, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA I0421 15:52:26.788421 126981123343936 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0 I0421 15:52:27.447159 126981123343936 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_false/checkpoints/0`. I0421 15:52:27.447786 126981123343936 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 9.793190s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/0 I0421 15:52:27.447862 126981123343936 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.541545s. I0421 15:52:27.448061 126981200938560 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:52:27.448164 126981200938560 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:52:27.448230 126981200938560 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:52:27.448280 126981200938560 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 15:52:27.448386 127065353903232 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 15:52:27.448450 127065353903232 checkpoint_manager.py:1441] Waiting for previous save to complete took 7.536414 seconds. If this number is high, consider checkpointing less frequently. I0421 15:52:27.449415 127065353903232 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0421 15:52:27.449686 127065353903232 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_false/checkpoints/5. I0421 15:52:27.607401 127065353903232 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0421 15:52:27.607490 127065353903232 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 15:52:27.650779 126981200938560 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5 I0421 15:52:28.372293 126981190452800 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5/items I0421 15:52:39.757433 127065353903232 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.150432s I0421 15:52:39.762246 127065353903232 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.009 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0421 15:52:39.762307 127065353903232 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.236128s (batch_requests_ready=0.077739s, total_serialization_initiated=12.153688s, others=0.004700s) I0421 15:52:39.762380 127065353903232 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.236658s (all_items=0.000021s, per_item={'items': '0.00002122'}, temp_paths=12.236637) I0421 15:52:39.763564 126981200938560 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0421 15:52:39.763677 127065353903232 async_checkpointer.py:561] Finished blocking save. Time taken: 12.314219s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5. I0421 15:52:39.763876 127065353903232 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:52:39.764070 126979376416320 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:52:39.764169 127065353903232 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786739.9120224, 'wait_for_prev_duration_secs': 7.53641414642334, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776786747.4494386, 'checkpointer_blocking_duration_secs': 12.314347743988037, 'get_old_steps_start_time': 1776786759.763804, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776786739.911955, 'checkpoint_manager_blocking_duration_secs': 19.85218906402588} I0421 15:52:39.764327 127065353903232 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0421 15:52:39.765025 127065353903232 metric_logger.py:196] completed step: 4, seconds: 0.120, TFLOP/s/device: 113.687, Tokens/s/device: 17136.068, total_weights: 16384, loss: 8.327, lm_loss: 8.327, perplexity: 4134.812 I0421 15:52:39.773471 127065353903232 metric_logger.py:196] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.638, Tokens/s/device: 17279.492, total_weights: 16384, loss: 7.367, lm_loss: 7.367, perplexity: 1583.522 I0421 15:52:39.890125 127065353903232 metric_logger.py:196] completed step: 6, seconds: 20.254, TFLOP/s/device: 0.671, Tokens/s/device: 101.116, total_weights: 16384, loss: 6.459, lm_loss: 6.459, perplexity: 638.363 I0421 15:52:40.009757 127065353903232 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2187.943, Tokens/s/device: 329790.660, total_weights: 16384, loss: 5.622, lm_loss: 5.622, perplexity: 276.577 I0421 15:52:40.128094 127065353903232 metric_logger.py:196] completed step: 8, seconds: 0.117, TFLOP/s/device: 116.053, Tokens/s/device: 17492.761, total_weights: 16384, loss: 4.864, lm_loss: 4.864, perplexity: 129.523 I0421 15:52:40.247237 127065353903232 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0421 15:52:40.248534 127065353903232 checkpointing.py:776] Waited 0.0012967586517333984 seconds for step 9 to finish before starting checkpointing. I0421 15:52:40.248976 127065353903232 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:52:40.529032 126981169481280 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_false/checkpoints/5/items/array_metadatas/process_0 I0421 15:52:48.319612 126981190452800 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 8.556664s (commit=8.131439s, array_metadata_write=0.425225s) I0421 15:52:48.320761 126981200938560 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 607.698 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 20 seconds) (per-host) I0421 15:52:48.320869 126981200938560 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 8.557147s. W0421 15:52:48.642176 126981200938560 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_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5/items, I0421 15:52:48.736656 126981200938560 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 15:52:48.736759 126981200938560 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.160314s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5/items I0421 15:52:48.737404 126981200938560 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5/items I0421 15:52:48.963323 126981200938560 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5 I0421 15:52:49.655283 126981200938560 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_false/checkpoints/5`. I0421 15:52:49.655915 126981200938560 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 22.206464s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/5 I0421 15:52:49.655993 126981200938560 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 9.892274s. I0421 15:52:49.656224 126979376416320 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:52:49.656335 126979376416320 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:52:49.656399 126979376416320 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:52:49.656443 126979376416320 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 15:52:49.656556 127065353903232 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 15:52:49.656629 127065353903232 checkpoint_manager.py:1441] Waiting for previous save to complete took 9.407665 seconds. If this number is high, consider checkpointing less frequently. I0421 15:52:49.656831 127065353903232 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0421 15:52:49.657070 127065353903232 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_false/checkpoints/9. I0421 15:52:49.816947 126981200938560 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9 I0421 15:52:49.837075 127065353903232 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0421 15:52:49.837170 127065353903232 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 15:52:50.505001 126981190452800 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9/items I0421 15:53:03.460809 127065353903232 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 13.624114s I0421 15:53:03.464859 127065353903232 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 920.561 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 13 seconds) (per-host) I0421 15:53:03.464920 127065353903232 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 13.727364s (batch_requests_ready=0.095520s, total_serialization_initiated=13.627900s, others=0.003944s) I0421 15:53:03.464996 127065353903232 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 13.727876s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=13.727861) I0421 15:53:03.466398 126981200938560 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0421 15:53:03.466491 127065353903232 async_checkpointer.py:561] Finished blocking save. Time taken: 13.809617s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9. I0421 15:53:03.466681 127065353903232 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 15:53:03.466873 126981169481280 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0421 15:53:03.466957 127065353903232 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776786760.2489488, 'wait_for_prev_duration_secs': 9.407665014266968, 'time_between_consecutive_saves_sec': 12.800652265548706, 'checkpointer_blocking_start_time': 1776786769.656856, 'checkpointer_blocking_duration_secs': 13.809753179550171, 'get_old_steps_start_time': 1776786783.4666245, 'get_old_steps_duration_secs': 2.193450927734375e-05, 'checkpoint_manager_blocking_start_time': 1776786760.2489126, 'checkpoint_manager_blocking_duration_secs': 23.21802020072937} I0421 15:53:03.467087 127065353903232 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0421 15:53:03.467118 127065353903232 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 15:53:04.270509 126979428845120 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_false/checkpoints/9/items/array_metadatas/process_0 I0421 15:53:10.275119 126980550821440 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 6.809548s (commit=6.410593s, array_metadata_write=0.398955s) I0421 15:53:10.276061 126981200938560 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 615.275 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 20 seconds) (per-host) I0421 15:53:10.276174 126981200938560 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 6.809638s. W0421 15:53:10.594072 126981200938560 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_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9/items, I0421 15:53:10.685419 126981200938560 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 15:53:10.685523 126981200938560 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.151666s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9/items I0421 15:53:10.686180 126981200938560 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9/items I0421 15:53:10.923979 126981200938560 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9 I0421 15:53:11.599241 126981200938560 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_false/checkpoints/9`. I0421 15:53:11.599850 126981200938560 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 21.942982s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260421_144657/linen_main_20260421_144657_15_ocdbt_false/checkpoints/9 I0421 15:53:11.599911 126981200938560 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 8.133377s. I0421 15:53:11.600019 126981169481280 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 15:53:11.600078 126981169481280 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0421 15:53:11.600116 126981169481280 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 15:53:11.600150 126981169481280 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 15:53:11.600275 127065353903232 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 15:53:11.600414 127065353903232 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 15:53:11.601251 127065353903232 metric_logger.py:196] completed step: 9, seconds: 0.129, TFLOP/s/device: 105.085, Tokens/s/device: 15839.624, 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