2026-04-15 22:12:08.926090: 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) I0415 22:12:09.042631 124662484884608 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 22:13:05.025508 124662484884608 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 22:13:05.025619 124662484884608 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 22:13:05.025653 124662484884608 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Apr 6 2026 20:48:10 (1775533690) cl/895581894 I0415 22:13:05.025680 124662484884608 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 22:13:05.025701 124662484884608 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0415 22:13:05.025780 124662484884608 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 22:13:05.117610 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:13:05.521809 124662484884608 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 22:13:05.522142 124662484884608 checkpointing.py:677] Setting up checkpoint logger... I0415 22:13:05.522186 124662484884608 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0415 22:13:05.522226 124662484884608 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 22:13:05.522790 124662484884608 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 0x7160a83facf0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 22:13:07.984779 124662484884608 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 22:13:07.984951 124662484884608 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 0x715a2c138e90>}, handler_registry=None I0415 22:13:07.985219 124662484884608 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x715a2c138e90>` 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`. I0415 22:13:07.985262 124662484884608 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x715a2c1394f0>` 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`. I0415 22:13:07.985292 124662484884608 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 0x715a2c138e90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x715a2c138e90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x715a2c1394f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x715a2c1394f0>}). I0415 22:13:07.985588 124662484884608 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 22:13:07.985644 124662484884608 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x715a2c12d3a0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 22:13:08.112880 124662484884608 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints I0415 22:13:08.113128 124662484884608 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_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x715a2c139b50> I0415 22:13:08.113208 124662484884608 checkpointing.py:301] Checkpoint manager created! W0415 22:13:08.129344 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:13:08.309924 124662484884608 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 22:13:08.310031 124662484884608 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:13:08.404798 124662484884608 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 22:13:08.404874 124662484884608 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:13:08.419073 124662484884608 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 22:13:08.419126 124662484884608 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:13:08.439363 124662484884608 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 22:13:08.439416 124662484884608 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:13:08.453553 124662484884608 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 22:13:08.453605 124662484884608 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:13:08.467665 124662484884608 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 22:13:08.467711 124662484884608 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:13:08.489228 124662484884608 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 22:13:08.489282 124662484884608 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:13:08.507512 124662484884608 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0415 22:13:08.507562 124662484884608 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0415 22:13:08.689331 124662484884608 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 22:13:08.689421 124662484884608 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 22:13:08.744453 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. [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 I0415 22:13:09.127291 124662484884608 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 22:13:09.127454 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0415 22:13:09.127494 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0415 22:13:09.127537 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0415 22:13:09.127576 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0415 22:13:09.127612 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0415 22:13:09.127655 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0415 22:13:09.127694 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0415 22:13:09.127722 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0415 22:13:09.127748 124662484884608 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0415 22:13:09.127783 124662484884608 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0415 22:13:09.127816 124662484884608 maxtext_utils.py:1620] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0415 22:13:09.547430 124662484884608 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 22:13:09.547513 124662484884608 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0415 22:13:09.560982 124662484884608 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 22:13:09.561030 124662484884608 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0415 22:13:10.200399 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:13:10.325010 124662484884608 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. I0415 22:13:10.325490 124662484884608 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 22:13:10.325636 124662484884608 metric_logger.py:289] number parameters: 1.104 billion W0415 22:13:11.154482 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:13:11.251280 124662484884608 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 22:13:11.368340 124662484884608 checkpointing.py:776] Waited 0.11704826354980469 seconds for step 0 to finish before starting checkpointing. I0415 22:13:11.368763 124662484884608 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 22:13:11.368909 124662484884608 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 22:13:11.369273 124662484884608 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0. I0415 22:13:11.453953 124662484884608 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 22:13:11.544117 124662484884608 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:13:11.544203 124662484884608 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 I0415 22:13:11.549839 124563616695872 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0 W0415 22:13:11.758614 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.771963 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.780605 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.798730 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.805451 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.815649 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.821472 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:11.826686 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:13:12.010071 124662484884608 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.466392s I0415 22:13:12.010503 124662484884608 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 22.212 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 555 milliseconds) (per-host) I0415 22:13:12.010554 124662484884608 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.555692s (batch_requests_ready=0.085124s, total_serialization_initiated=0.470233s, others=0.000335s) I0415 22:13:12.010615 124662484884608 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.556207s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.556189) I0415 22:13:12.011345 124563553781312 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:13:12.011440 124662484884608 async_checkpointer.py:561] Finished blocking save. Time taken: 0.642492s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0. I0415 22:13:12.011619 124662484884608 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:13:12.011776 124563627181632 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:13:12.011885 124662484884608 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776291191.3687499, 'wait_for_prev_duration_secs': 4.124641418457031e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776291191.3689303, 'checkpointer_blocking_duration_secs': 0.6426191329956055, 'get_old_steps_start_time': 1776291192.0115662, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1776291191.368666, 'checkpoint_manager_blocking_duration_secs': 0.6431918144226074} I0415 22:13:12.011993 124662484884608 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 22:13:12.012031 124662484884608 max_utils.py:750] Memstats: After params initialized: I0415 22:13:12.012124 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0415 22:13:12.012153 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0415 22:13:12.012175 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0415 22:13:12.012194 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0415 22:13:12.012213 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0415 22:13:12.012230 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0415 22:13:12.012248 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0415 22:13:12.012265 124662484884608 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0415 22:13:12.018072 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:12.023783 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:13:12.029213 3289279 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:13:12.211951 124563606210112 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/items I0415 22:13:12.227109 124662484884608 metric_logger.py:185] completed step: 0, seconds: 0.925, TFLOP/s/device: 14.686, Tokens/s/device: 2213.654, total_weights: 16384, loss: 10.874 I0415 22:13:12.227999 124662484884608 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/tensorboard/' I0415 22:13:12.343337 124662484884608 metric_logger.py:185] completed step: 1, seconds: 0.974, TFLOP/s/device: 13.949, Tokens/s/device: 2102.501, total_weights: 16384, loss: 10.874 I0415 22:13:12.370946 124563534906944 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776291192117529817, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776291192117529817, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0415 22:13:12.461840 124662484884608 metric_logger.py:185] completed step: 2, seconds: 0.006, TFLOP/s/device: 2464.115, Tokens/s/device: 371418.208, total_weights: 16384, loss: 10.276 I0415 22:13:12.579977 124662484884608 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.057, Tokens/s/device: 17493.359, total_weights: 16384, loss: 9.282 I0415 22:13:12.584383 124662484884608 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0415 22:13:12.689789 3291436 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 22:13:12.689860 3291436 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 22:13:12.816411 124662484884608 checkpointing.py:776] Waited 0.2319962978363037 seconds for step 5 to finish before starting checkpointing. I0415 22:13:12.816798 124662484884608 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:13:13.485884 124563574752832 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/items/array_metadatas/process_0 I0415 22:13:44.166169 124563564267072 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 31.480381s (commit=31.012632s, array_metadata_write=0.467749s) I0415 22:13:44.167175 124563553781312 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 386.301 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 32 seconds) (per-host) I0415 22:13:44.167274 124563553781312 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.155782s. I0415 22:13:44.392179 124563553781312 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776291192117529817, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0415 22:13:44.601537 124563553781312 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0415 22:13:44.777216 124563534906944 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776291192117529817, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0415 22:13:45.046675 124563553781312 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 22:13:45.046853 124563553781312 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.610615s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/items I0415 22:13:45.047523 124563553781312 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0/items I0415 22:13:45.303150 124563553781312 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0 I0415 22:13:46.019797 124563553781312 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0`. I0415 22:13:46.020422 124563553781312 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 34.651480s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/0 I0415 22:13:46.020498 124563553781312 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 34.009008s. I0415 22:13:46.020677 124563627181632 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:13:46.020782 124563627181632 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:13:46.020848 124563627181632 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:13:46.020895 124563627181632 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 22:13:46.021122 124662484884608 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. W0415 22:13:46.021672 124662484884608 checkpoint_manager.py:1441] Waiting for previous save to complete took 33.204874 seconds. If this number is high, consider checkpointing less frequently. I0415 22:13:46.022451 124662484884608 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0415 22:13:46.022724 124662484884608 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5. I0415 22:13:46.203672 124662484884608 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:13:46.203809 124563627181632 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5 I0415 22:13:46.203886 124662484884608 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 I0415 22:13:46.906807 124563606210112 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5/items I0415 22:13:58.782886 124662484884608 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.579865s I0415 22:13:58.787106 124662484884608 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 996.897 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0415 22:13:58.787165 124662484884608 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.676203s (batch_requests_ready=0.089266s, total_serialization_initiated=12.582846s, others=0.004092s) I0415 22:13:58.787256 124662484884608 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.676746s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=12.676729) I0415 22:13:58.788744 124563503449664 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:13:58.788837 124662484884608 async_checkpointer.py:561] Finished blocking save. Time taken: 12.766330s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5. I0415 22:13:58.789116 124662484884608 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:13:58.789295 124563553781312 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:13:58.789393 124662484884608 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776291192.816776, 'wait_for_prev_duration_secs': 33.20487380027771, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776291226.0224743, 'checkpointer_blocking_duration_secs': 12.766508340835571, 'get_old_steps_start_time': 1776291238.7890134, 'get_old_steps_duration_secs': 4.839897155761719e-05, 'checkpoint_manager_blocking_start_time': 1776291192.8167126, 'checkpoint_manager_blocking_duration_secs': 45.97265005111694} I0415 22:13:58.789503 124662484884608 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0415 22:13:58.790135 124662484884608 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.659, Tokens/s/device: 17282.700, total_weights: 16384, loss: 8.327 I0415 22:13:58.797129 124662484884608 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 115.346, Tokens/s/device: 17386.137, total_weights: 16384, loss: 7.367 I0415 22:13:58.913287 124662484884608 metric_logger.py:185] completed step: 6, seconds: 46.211, TFLOP/s/device: 0.294, Tokens/s/device: 44.318, total_weights: 16384, loss: 6.459 I0415 22:13:59.031483 124662484884608 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2505.926, Tokens/s/device: 377720.398, total_weights: 16384, loss: 5.622 I0415 22:13:59.149651 124662484884608 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 116.236, Tokens/s/device: 17520.296, total_weights: 16384, loss: 4.864 I0415 22:13:59.267565 124662484884608 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 22:13:59.268429 124662484884608 checkpointing.py:776] Waited 0.0008897781372070312 seconds for step 9 to finish before starting checkpointing. I0415 22:13:59.268676 124662484884608 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:13:59.576863 124563574752832 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5/items/array_metadatas/process_0 I0415 22:14:27.984922 124563564267072 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.197089s (commit=28.756959s, array_metadata_write=0.440130s) I0415 22:14:27.985979 124563503449664 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 301.774 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 41 seconds) (per-host) I0415 22:14:27.986034 124563503449664 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.197138s. I0415 22:14:28.382470 124563503449664 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0415 22:14:28.803146 124563503449664 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 22:14:28.803913 124563503449664 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.554843s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5/items I0415 22:14:28.804546 124563503449664 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5/items I0415 22:14:29.050820 124563503449664 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5 I0415 22:14:29.740451 124563503449664 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5`. I0415 22:14:29.741090 124563503449664 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 43.718601s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/5 I0415 22:14:29.741168 124563503449664 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.952273s. I0415 22:14:29.741338 124563553781312 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:14:29.741443 124563553781312 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:14:29.741508 124563553781312 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:14:29.741553 124563553781312 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0415 22:14:29.741910 124662484884608 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. W0415 22:14:29.742897 124662484884608 checkpoint_manager.py:1441] Waiting for previous save to complete took 30.474230 seconds. If this number is high, consider checkpointing less frequently. I0415 22:14:29.743482 124662484884608 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 22:14:29.743760 124662484884608 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9. I0415 22:14:29.940275 124662484884608 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:14:29.940415 124563553781312 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9 I0415 22:14:29.940491 124662484884608 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 I0415 22:14:30.674103 124563492963904 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9/items I0415 22:14:39.812674 124662484884608 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 9.873106s I0415 22:14:39.816989 124662484884608 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.238 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 9 seconds) (per-host) I0415 22:14:39.817067 124662484884608 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 9.972151s (batch_requests_ready=0.091400s, total_serialization_initiated=9.876543s, others=0.004208s) I0415 22:14:39.817155 124662484884608 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 9.972629s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=9.972615) I0415 22:14:39.818770 124563553781312 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:14:39.818854 124662484884608 async_checkpointer.py:561] Finished blocking save. Time taken: 10.075326s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9. I0415 22:14:39.819037 124662484884608 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:14:39.819177 124563503449664 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:14:39.819255 124662484884608 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776291239.26865, 'wait_for_prev_duration_secs': 30.47422957420349, 'time_between_consecutive_saves_sec': 13.247738599777222, 'checkpointer_blocking_start_time': 1776291269.7435064, 'checkpointer_blocking_duration_secs': 10.075453996658325, 'get_old_steps_start_time': 1776291279.818979, 'get_old_steps_duration_secs': 2.3126602172851562e-05, 'checkpoint_manager_blocking_start_time': 1776291239.268619, 'checkpoint_manager_blocking_duration_secs': 40.55060791969299} I0415 22:14:39.819380 124662484884608 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 22:14:39.819410 124662484884608 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:14:40.561351 124563627181632 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9/items/array_metadatas/process_0 I0415 22:15:14.657548 124563492963904 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 34.839643s (commit=34.421893s, array_metadata_write=0.417750s) I0415 22:15:14.658498 124563553781312 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 281.986 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 44 seconds) (per-host) I0415 22:15:14.658556 124563553781312 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.839654s. I0415 22:15:15.047530 124563553781312 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0415 22:15:15.469853 124563553781312 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 22:15:15.470024 124563553781312 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.559345s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9/items I0415 22:15:15.470655 124563553781312 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9/items I0415 22:15:15.712641 124563553781312 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9 I0415 22:15:16.388175 124563553781312 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9`. I0415 22:15:16.388797 124563553781312 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 46.645277s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_true/checkpoints/9 I0415 22:15:16.388866 124563553781312 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.569969s. I0415 22:15:16.388970 124563503449664 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:15:16.389026 124563503449664 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:15:16.389105 124563503449664 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:15:16.389151 124563503449664 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 22:15:16.389311 124662484884608 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. I0415 22:15:16.389667 124662484884608 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 22:15:16.390218 124662484884608 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.935, Tokens/s/device: 17324.220, total_weights: 16384, loss: 4.189 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops