2026-04-16 13:42:09.726653: 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) I0416 13:42:09.842892 129942998244480 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. ~/maxtext_venv/lib/python3.12/site-packages/jax/_src/xla_bridge.py:219: UserWarning: TPU backend initialization is taking more than 60.0 seconds. Did you run your code on all TPU hosts? See https://docs.jax.dev/en/latest/multi_process.html for more information. warnings.warn( I0416 13:43:10.999923 129942998244480 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 13:43:11.000032 129942998244480 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 13:43:11.000088 129942998244480 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 I0416 13:43:11.000115 129942998244480 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 13:43:11.000137 129942998244480 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 13:43:11.000219 129942998244480 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 13:43:11.093080 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 13:43:11.497923 129942998244480 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 13:43:11.498263 129942998244480 checkpointing.py:677] Setting up checkpoint logger... I0416 13:43:11.498308 129942998244480 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 13:43:11.498348 129942998244480 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 13:43:11.498875 129942998244480 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 0x762e1f500b90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 13:43:13.838520 129942998244480 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 13:43:13.838696 129942998244480 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 0x7627a3744b00>}, handler_registry=None I0416 13:43:13.838966 129942998244480 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7627a3744b00>` 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`. I0416 13:43:13.839007 129942998244480 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7627a3746b40>` 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`. I0416 13:43:13.839038 129942998244480 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 0x7627a3744b00>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7627a3744b00>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7627a3746b40>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7627a3746b40>}). I0416 13:43:13.839346 129942998244480 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 13:43:13.839397 129942998244480 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7627a374d3a0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 13:43:13.967539 129942998244480 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints I0416 13:43:13.967751 129942998244480 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_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7627a45e3950> I0416 13:43:13.967832 129942998244480 checkpointing.py:301] Checkpoint manager created! W0416 13:43:13.983997 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 13:43:14.164511 129942998244480 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 13:43:14.164625 129942998244480 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 13:43:14.260284 129942998244480 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 13:43:14.260362 129942998244480 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 13:43:14.274966 129942998244480 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 13:43:14.275015 129942998244480 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 13:43:14.295799 129942998244480 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 13:43:14.295852 129942998244480 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 13:43:14.310415 129942998244480 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 13:43:14.310463 129942998244480 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 13:43:14.325121 129942998244480 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0416 13:43:14.325171 129942998244480 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 13:43:14.347006 129942998244480 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 13:43:14.347075 129942998244480 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 13:43:14.365820 129942998244480 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 13:43:14.365876 129942998244480 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 13:43:14.549545 129942998244480 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 13:43:14.549638 129942998244480 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0416 13:43:14.604780 3884910 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 I0416 13:43:14.992670 129942998244480 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 13:43:14.992836 129942998244480 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) I0416 13:43:14.992884 129942998244480 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) I0416 13:43:14.992930 129942998244480 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') I0416 13:43:14.992969 129942998244480 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) I0416 13:43:14.992996 129942998244480 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) I0416 13:43:14.993035 129942998244480 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) I0416 13:43:14.993091 129942998244480 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') I0416 13:43:14.993123 129942998244480 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) I0416 13:43:14.993149 129942998244480 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) I0416 13:43:14.993185 129942998244480 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0416 13:43:14.993219 129942998244480 maxtext_utils.py:1620] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0416 13:43:15.412820 129942998244480 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 13:43:15.412909 129942998244480 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 13:43:15.426561 129942998244480 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 13:43:15.426611 129942998244480 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 13:43:16.067557 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 13:43:16.190888 129942998244480 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. I0416 13:43:16.191398 129942998244480 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 13:43:16.191547 129942998244480 metric_logger.py:289] number parameters: 1.104 billion W0416 13:43:17.016186 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 13:43:17.111276 129942998244480 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0416 13:43:17.228115 129942998244480 checkpointing.py:776] Waited 0.11679744720458984 seconds for step 0 to finish before starting checkpointing. I0416 13:43:17.228559 129942998244480 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 13:43:17.228707 129942998244480 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0416 13:43:17.229095 129942998244480 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0. I0416 13:43:17.308694 129942998244480 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 13:43:17.397413 129942998244480 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 13:43:17.397505 129942998244480 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 I0416 13:43:17.402885 129835512890944 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0 W0416 13:43:17.605638 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.616030 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.631231 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.649947 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.659904 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.665206 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.669785 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:17.674265 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 13:43:18.043107 129835502405184 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/items I0416 13:43:18.205931 129835466753600 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776346997959337047, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776346997959337047, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0416 13:43:18.217496 129942998244480 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.820943s I0416 13:43:18.218137 129942998244480 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 13.591 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 908 milliseconds) (per-host) I0416 13:43:18.218279 129942998244480 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.908337s (batch_requests_ready=0.082912s, total_serialization_initiated=0.824889s, others=0.000536s) I0416 13:43:18.218398 129942998244480 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.909231s (all_items=0.000018s, per_item={'items': '0.00001812'}, temp_paths=0.909213) I0416 13:43:18.219301 129835445782080 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 13:43:18.219396 129942998244480 async_checkpointer.py:561] Finished blocking save. Time taken: 0.990648s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0. I0416 13:43:18.219589 129942998244480 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 13:43:18.219844 129835523376704 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 13:43:18.219953 129942998244480 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776346997.2285464, 'wait_for_prev_duration_secs': 4.029273986816406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776346997.2287285, 'checkpointer_blocking_duration_secs': 0.9907801151275635, 'get_old_steps_start_time': 1776346998.2195292, 'get_old_steps_duration_secs': 2.2649765014648438e-05, 'checkpoint_manager_blocking_start_time': 1776346997.2284613, 'checkpoint_manager_blocking_duration_secs': 0.9914615154266357} I0416 13:43:18.220082 129942998244480 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0416 13:43:18.220124 129942998244480 max_utils.py:750] Memstats: After params initialized: I0416 13:43:18.220218 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0416 13:43:18.220248 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 13:43:18.220270 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 13:43:18.220288 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 13:43:18.220308 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 13:43:18.220326 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 13:43:18.220344 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 13:43:18.220362 129942998244480 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 13:43:18.226157 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:18.235637 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 13:43:18.240763 3884910 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 13:43:18.428977 129942998244480 metric_logger.py:185] completed step: 0, seconds: 0.919, TFLOP/s/device: 14.781, Tokens/s/device: 2227.893, total_weights: 16384, loss: 10.874 I0416 13:43:18.429857 129942998244480 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/tensorboard/' I0416 13:43:18.513774 3887068 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 13:43:18.513831 3887068 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 13:43:18.545220 129942998244480 metric_logger.py:185] completed step: 1, seconds: 1.316, TFLOP/s/device: 10.324, Tokens/s/device: 1556.142, total_weights: 16384, loss: 10.874 I0416 13:43:18.663516 129942998244480 metric_logger.py:185] completed step: 2, seconds: 0.005, TFLOP/s/device: 2556.855, Tokens/s/device: 385397.064, total_weights: 16384, loss: 10.276 I0416 13:43:18.781802 129942998244480 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.067, Tokens/s/device: 17494.853, total_weights: 16384, loss: 9.282 I0416 13:43:18.786304 129942998244480 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0416 13:43:19.018735 129942998244480 checkpointing.py:776] Waited 0.23240041732788086 seconds for step 5 to finish before starting checkpointing. I0416 13:43:19.019124 129942998244480 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 13:43:19.258549 129835512890944 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/items/array_metadatas/process_0 I0416 13:43:44.032553 129835456267840 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 25.515822s (commit=25.088553s, array_metadata_write=0.427269s) I0416 13:43:44.033603 129835445782080 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 472.870 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 26 seconds) (per-host) I0416 13:43:44.033714 129835445782080 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 25.814272s. I0416 13:43:44.255565 129835445782080 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776346997959337047, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0416 13:43:44.444203 129835445782080 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 13:43:44.634859 129835466753600 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776346997959337047, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0416 13:43:44.877500 129835445782080 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 13:43:44.877691 129835445782080 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.583806s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/items I0416 13:43:44.878370 129835445782080 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0/items I0416 13:43:45.125750 129835445782080 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0 I0416 13:43:45.801080 129835445782080 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0`. I0416 13:43:45.801684 129835445782080 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 28.572943s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/0 I0416 13:43:45.801765 129835445782080 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 27.582325s. I0416 13:43:45.801958 129835523376704 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 13:43:45.802084 129835523376704 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 13:43:45.802154 129835523376704 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 13:43:45.802202 129835523376704 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 13:43:45.802422 129942998244480 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. W0416 13:43:45.802543 129942998244480 checkpoint_manager.py:1441] Waiting for previous save to complete took 26.783418 seconds. If this number is high, consider checkpointing less frequently. I0416 13:43:45.803353 129942998244480 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0416 13:43:45.803635 129942998244480 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5. I0416 13:43:45.974288 129835523376704 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5 I0416 13:43:45.979247 129942998244480 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 13:43:45.979340 129942998244480 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 I0416 13:43:46.690143 129835456267840 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5/items I0416 13:43:57.928095 129942998244480 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.949258s I0416 13:43:57.932222 129942998244480 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.025 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0416 13:43:57.932293 129942998244480 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.040721s (batch_requests_ready=0.084150s, total_serialization_initiated=11.952538s, others=0.004033s) I0416 13:43:57.932371 129942998244480 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.041218s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=12.041201) I0416 13:43:57.933859 129835445782080 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 13:43:57.933953 129942998244480 async_checkpointer.py:561] Finished blocking save. Time taken: 12.130545s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5. I0416 13:43:57.934203 129942998244480 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0416 13:43:57.934361 129835374478912 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 13:43:57.934460 129942998244480 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776346999.0191007, 'wait_for_prev_duration_secs': 26.78341841697693, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776347025.8033762, 'checkpointer_blocking_duration_secs': 12.130725622177124, 'get_old_steps_start_time': 1776347037.9341233, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776346999.019024, 'checkpoint_manager_blocking_duration_secs': 38.915409088134766} I0416 13:43:57.934584 129942998244480 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0416 13:43:57.935076 129942998244480 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.929, Tokens/s/device: 17323.341, total_weights: 16384, loss: 8.327 I0416 13:43:57.942105 129942998244480 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 114.929, Tokens/s/device: 17323.341, total_weights: 16384, loss: 7.367 I0416 13:43:58.058314 129942998244480 metric_logger.py:185] completed step: 6, seconds: 39.154, TFLOP/s/device: 0.347, Tokens/s/device: 52.306, total_weights: 16384, loss: 6.459 I0416 13:43:58.176450 129942998244480 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 2418.069, Tokens/s/device: 364477.665, total_weights: 16384, loss: 5.622 I0416 13:43:58.294806 129942998244480 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 116.329, Tokens/s/device: 17534.397, total_weights: 16384, loss: 4.864 I0416 13:43:58.412801 129942998244480 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0416 13:43:58.413635 129942998244480 checkpointing.py:776] Waited 0.0008656978607177734 seconds for step 9 to finish before starting checkpointing. I0416 13:43:58.413874 129942998244480 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 13:43:58.678588 129835512890944 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5/items/array_metadatas/process_0 I0416 13:44:35.000058 129835456267840 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.067097s (commit=36.637255s, array_metadata_write=0.429842s) I0416 13:44:35.000957 129835445782080 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 257.319 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 49 seconds) (per-host) I0416 13:44:35.001012 129835445782080 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.067004s. I0416 13:44:35.411657 129835445782080 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 13:44:35.811553 129835445782080 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 13:44:35.812333 129835445782080 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.541056s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5/items I0416 13:44:35.812973 129835445782080 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5/items I0416 13:44:36.050553 129835445782080 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5 I0416 13:44:36.731683 129835445782080 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5`. I0416 13:44:36.732316 129835445782080 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 50.928925s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/5 I0416 13:44:36.732394 129835445782080 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 38.798387s. I0416 13:44:36.732528 129835374478912 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 13:44:36.732627 129835374478912 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 13:44:36.732678 129835374478912 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0416 13:44:36.732723 129835374478912 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0416 13:44:36.733186 129942998244480 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. W0416 13:44:36.733299 129942998244480 checkpoint_manager.py:1441] Waiting for previous save to complete took 38.319434 seconds. If this number is high, consider checkpointing less frequently. I0416 13:44:36.733496 129942998244480 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0416 13:44:36.733791 129942998244480 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9. I0416 13:44:36.909106 129835374478912 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9 I0416 13:44:36.913650 129942998244480 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 13:44:36.913741 129942998244480 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 I0416 13:44:37.594724 129835523376704 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9/items I0416 13:44:43.316232 129942998244480 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 6.402959s I0416 13:44:43.320222 129942998244480 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.901 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 6 seconds) (per-host) I0416 13:44:43.320280 129942998244480 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 6.492730s (batch_requests_ready=0.082811s, total_serialization_initiated=6.406034s, others=0.003884s) I0416 13:44:43.320349 129942998244480 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 6.493202s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=6.493191) I0416 13:44:43.321938 129835405936192 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 13:44:43.322018 129942998244480 async_checkpointer.py:561] Finished blocking save. Time taken: 6.588481s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9. I0416 13:44:43.322192 129942998244480 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 13:44:43.322356 129835445782080 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 13:44:43.322462 129942998244480 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776347038.4138503, 'wait_for_prev_duration_secs': 38.31943440437317, 'time_between_consecutive_saves_sec': 12.611634016036987, 'checkpointer_blocking_start_time': 1776347076.7335196, 'checkpointer_blocking_duration_secs': 6.588613986968994, 'get_old_steps_start_time': 1776347083.3221452, 'get_old_steps_duration_secs': 1.6450881958007812e-05, 'checkpoint_manager_blocking_start_time': 1776347038.4138188, 'checkpoint_manager_blocking_duration_secs': 44.90862059593201} I0416 13:44:43.322568 129942998244480 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0416 13:44:43.322603 129942998244480 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 13:44:44.056331 129835512890944 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9/items/array_metadatas/process_0 I0416 13:45:12.436125 129835456267840 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.115266s (commit=28.666180s, array_metadata_write=0.449086s) I0416 13:45:12.437270 129835405936192 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 354.869 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 35 seconds) (per-host) I0416 13:45:12.437328 129835405936192 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.115256s. I0416 13:45:12.822536 129835405936192 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0416 13:45:13.257797 129835405936192 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 13:45:13.258451 129835405936192 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.571356s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9/items I0416 13:45:13.259120 129835405936192 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9/items I0416 13:45:13.519792 129835405936192 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9 I0416 13:45:14.184352 129835405936192 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9`. I0416 13:45:14.184986 129835405936192 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 37.451456s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_16_zarr3_true/checkpoints/9 I0416 13:45:14.185072 129835405936192 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.863000s. I0416 13:45:14.185250 129835445782080 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 13:45:14.185358 129835445782080 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 13:45:14.185416 129835445782080 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 13:45:14.185457 129835445782080 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 13:45:14.185853 129942998244480 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. I0416 13:45:14.185984 129942998244480 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 13:45:14.186764 129942998244480 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.814, Tokens/s/device: 17306.067, 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