2026-04-15 22:19:22.981723: 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 22:19:23.097428 130577495841920 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 22:20:16.596834 130577495841920 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 22:20:16.596951 130577495841920 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 22:20:16.596986 130577495841920 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 22:20:16.597012 130577495841920 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 22:20:16.597036 130577495841920 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0415 22:20:16.597147 130577495841920 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 22:20:16.690016 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:20:17.095771 130577495841920 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 22:20:17.096094 130577495841920 checkpointing.py:677] Setting up checkpoint logger... I0415 22:20:17.096139 130577495841920 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=False I0415 22:20:17.096172 130577495841920 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 22:20:17.096819 130577495841920 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x76c1da505cd0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 22:20:19.536019 130577495841920 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 22:20:19.536212 130577495841920 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 0x76bb5e735df0>}, handler_registry=None I0415 22:20:19.536454 130577495841920 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76bb5e735df0>` 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 22:20:19.536495 130577495841920 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76bb5e736330>` 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 22:20:19.536524 130577495841920 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 0x76bb5e735df0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76bb5e735df0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76bb5e736330>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76bb5e736330>}). I0415 22:20:19.536813 130577495841920 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 22:20:19.536860 130577495841920 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x76bb5e7313a0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 22:20:19.665967 130577495841920 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints I0415 22:20:19.666218 130577495841920 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_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x76bcd59fc3b0> I0415 22:20:19.666294 130577495841920 checkpointing.py:301] Checkpoint manager created! W0415 22:20:19.682412 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:20:19.861946 130577495841920 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 22:20:19.862168 130577495841920 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:20:19.951884 130577495841920 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 22:20:19.951961 130577495841920 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:20:19.966212 130577495841920 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 22:20:19.966261 130577495841920 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 22:20:19.986758 130577495841920 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 22:20:19.986811 130577495841920 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:20:20.006846 130577495841920 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 22:20:20.006933 130577495841920 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:20:20.021389 130577495841920 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 22:20:20.021440 130577495841920 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:20:20.043283 130577495841920 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 22:20:20.043338 130577495841920 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 22:20:20.061429 130577495841920 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0415 22:20:20.061478 130577495841920 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0415 22:20:20.242425 130577495841920 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 22:20:20.242522 130577495841920 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 22:20:20.292953 3295480 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 22:20:20.670543 130577495841920 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 22:20:20.670651 130577495841920 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 22:20:20.670692 130577495841920 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 22:20:20.670735 130577495841920 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 22:20:20.670773 130577495841920 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 22:20:20.670804 130577495841920 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 22:20:20.670841 130577495841920 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 22:20:20.670879 130577495841920 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 22:20:20.670909 130577495841920 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 22:20:20.670941 130577495841920 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 22:20:20.670977 130577495841920 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0415 22:20:20.671009 130577495841920 maxtext_utils.py:1620] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0415 22:20:21.087213 130577495841920 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 22:20:21.087296 130577495841920 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0415 22:20:21.100718 130577495841920 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 22:20:21.100769 130577495841920 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0415 22:20:21.711760 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:20:21.827300 130577495841920 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 22:20:21.827768 130577495841920 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 22:20:21.827905 130577495841920 metric_logger.py:289] number parameters: 1.104 billion W0415 22:20:22.651420 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:20:22.745609 130577495841920 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 22:20:22.862771 130577495841920 checkpointing.py:776] Waited 0.11714768409729004 seconds for step 0 to finish before starting checkpointing. I0415 22:20:22.863213 130577495841920 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 22:20:22.863365 130577495841920 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 22:20:22.863734 130577495841920 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_false/checkpoints/0. I0415 22:20:22.952428 130577495841920 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 22:20:23.038883 130577495841920 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:20:23.038972 130577495841920 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 22:20:23.045743 130470052365888 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0 W0415 22:20:23.253231 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.264020 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.273275 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.287285 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.297484 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.305989 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.310602 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.315266 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:20:23.500769 130577495841920 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.462331s I0415 22:20:23.501210 130577495841920 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 22.528 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 547 milliseconds) (per-host) I0415 22:20:23.501262 130577495841920 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.547899s (batch_requests_ready=0.081337s, total_serialization_initiated=0.466221s, others=0.000341s) I0415 22:20:23.501322 130577495841920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.548430s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.548411) I0415 22:20:23.502017 130469989451328 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:20:23.502119 130577495841920 async_checkpointer.py:561] Finished blocking save. Time taken: 0.638715s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0. I0415 22:20:23.502291 130577495841920 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:20:23.502434 130470062851648 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:20:23.502501 130577495841920 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776291622.8632, 'wait_for_prev_duration_secs': 4.363059997558594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776291622.8633862, 'checkpointer_blocking_duration_secs': 0.6388323307037354, 'get_old_steps_start_time': 1776291623.5022366, 'get_old_steps_duration_secs': 2.0742416381835938e-05, 'checkpoint_manager_blocking_start_time': 1776291622.8631148, 'checkpoint_manager_blocking_duration_secs': 0.6393570899963379} I0415 22:20:23.502614 130577495841920 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 22:20:23.502653 130577495841920 max_utils.py:750] Memstats: After params initialized: I0415 22:20:23.502724 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0415 22:20:23.502752 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0415 22:20:23.502773 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0415 22:20:23.502794 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0415 22:20:23.502812 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0415 22:20:23.502831 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0415 22:20:23.502849 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0415 22:20:23.502865 130577495841920 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0415 22:20:23.508187 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.514613 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 22:20:23.519556 3295480 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 22:20:23.696882 130577495841920 metric_logger.py:185] completed step: 0, seconds: 0.917, TFLOP/s/device: 14.813, Tokens/s/device: 2232.829, total_weights: 16384, loss: 10.874 I0415 22:20:23.697633 130577495841920 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/tensorboard/' I0415 22:20:23.752449 130470041880128 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0/items I0415 22:20:23.813173 130577495841920 metric_logger.py:185] completed step: 1, seconds: 0.950, TFLOP/s/device: 14.307, Tokens/s/device: 2156.530, total_weights: 16384, loss: 10.874 I0415 22:20:23.917648 130469970576960 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776291623657170830, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776291623657170830, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0415 22:20:23.931509 130577495841920 metric_logger.py:185] completed step: 2, seconds: 0.005, TFLOP/s/device: 2659.450, Tokens/s/device: 400861.225, total_weights: 16384, loss: 10.276 I0415 22:20:24.049685 130577495841920 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.025, Tokens/s/device: 17488.579, total_weights: 16384, loss: 9.282 I0415 22:20:24.054316 130577495841920 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0415 22:20:24.229340 3297651 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 22:20:24.229409 3297651 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 22:20:24.286451 130577495841920 checkpointing.py:776] Waited 0.2321000099182129 seconds for step 5 to finish before starting checkpointing. I0415 22:20:24.286865 130577495841920 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:20:24.989626 130470010422848 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_false/checkpoints/0/items/array_metadatas/process_0 I0415 22:21:00.311204 130469999937088 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.079099s (commit=35.592690s, array_metadata_write=0.486410s) I0415 22:21:00.312227 130469989451328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 338.254 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 37 seconds) (per-host) I0415 22:21:00.312325 130469989451328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.810161s. I0415 22:21:00.540057 130469989451328 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776291623657170830, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0415 22:21:00.722406 130469989451328 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 22:21:00.934008 130469970576960 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776291623657170830, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0/_CHECKPOINT_METADATA I0415 22:21:01.209435 130469989451328 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.630568s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0/items I0415 22:21:01.210162 130469989451328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0/items I0415 22:21:01.449793 130469989451328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0 I0415 22:21:02.178090 130469989451328 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_false/checkpoints/0`. I0415 22:21:02.178692 130469989451328 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 39.315294s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/0 I0415 22:21:02.178763 130469989451328 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 38.676601s. I0415 22:21:02.178876 130470062851648 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:21:02.178928 130470062851648 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:21:02.178987 130470062851648 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:21:02.179026 130470062851648 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 22:21:02.179209 130577495841920 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 22:21:02.179338 130577495841920 checkpoint_manager.py:1441] Waiting for previous save to complete took 37.892478 seconds. If this number is high, consider checkpointing less frequently. I0415 22:21:02.180121 130577495841920 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0415 22:21:02.180403 130577495841920 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_false/checkpoints/5. I0415 22:21:02.363668 130577495841920 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:21:02.363759 130577495841920 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 22:21:02.373699 130470062851648 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5 I0415 22:21:03.086702 130469918148160 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5/items I0415 22:21:14.936290 130577495841920 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.573017s I0415 22:21:14.936823 130577495841920 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 997.640 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0415 22:21:14.936887 130577495841920 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.666770s (batch_requests_ready=0.090195s, total_serialization_initiated=12.576147s, others=0.000427s) I0415 22:21:14.936956 130577495841920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.667270s (all_items=0.000028s, per_item={'items': '0.00002837'}, temp_paths=12.667242) I0415 22:21:14.941943 130469918148160 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:21:14.942056 130577495841920 async_checkpointer.py:561] Finished blocking save. Time taken: 12.761873s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5. I0415 22:21:14.942251 130577495841920 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:21:14.942414 130469989451328 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:21:14.942512 130577495841920 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776291624.2868347, 'wait_for_prev_duration_secs': 37.892478227615356, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776291662.1801481, 'checkpointer_blocking_duration_secs': 12.762015104293823, 'get_old_steps_start_time': 1776291674.9421813, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776291624.2867687, 'checkpoint_manager_blocking_duration_secs': 50.65571713447571} I0415 22:21:14.942647 130577495841920 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0415 22:21:14.948801 130577495841920 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.677, Tokens/s/device: 17285.326, total_weights: 16384, loss: 8.327 I0415 22:21:14.956512 130577495841920 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 114.956, Tokens/s/device: 17327.445, total_weights: 16384, loss: 7.367 I0415 22:21:15.071898 130577495841920 metric_logger.py:185] completed step: 6, seconds: 50.900, TFLOP/s/device: 0.267, Tokens/s/device: 40.236, total_weights: 16384, loss: 6.459 I0415 22:21:15.190438 130577495841920 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 2285.087, Tokens/s/device: 344433.232, total_weights: 16384, loss: 5.622 I0415 22:21:15.308575 130577495841920 metric_logger.py:185] completed step: 8, seconds: 0.116, TFLOP/s/device: 116.903, Tokens/s/device: 17620.842, total_weights: 16384, loss: 4.864 I0415 22:21:15.426181 130577495841920 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 22:21:15.426968 130577495841920 checkpointing.py:776] Waited 0.0008130073547363281 seconds for step 9 to finish before starting checkpointing. I0415 22:21:15.427220 130577495841920 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:21:15.754766 130468806657600 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_false/checkpoints/5/items/array_metadatas/process_0 I0415 22:21:53.504032 130470010422848 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.562739s (commit=38.119188s, array_metadata_write=0.443551s) I0415 22:21:53.504940 130469918148160 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 246.644 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 51 seconds) (per-host) I0415 22:21:53.504991 130469918148160 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.562890s. I0415 22:21:53.901910 130469918148160 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 22:21:54.471923 130469918148160 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.697745s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5/items I0415 22:21:54.472664 130469918148160 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5/items I0415 22:21:54.713366 130469918148160 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5 I0415 22:21:55.395386 130469918148160 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_false/checkpoints/5`. I0415 22:21:55.395992 130469918148160 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 53.215830s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/5 I0415 22:21:55.396078 130469918148160 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 40.453977s. I0415 22:21:55.396219 130469989451328 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:21:55.396316 130469989451328 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:21:55.396365 130469989451328 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:21:55.396404 130469989451328 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0415 22:21:55.396532 130577495841920 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 22:21:55.396641 130577495841920 checkpoint_manager.py:1441] Waiting for previous save to complete took 39.969431 seconds. If this number is high, consider checkpointing less frequently. I0415 22:21:55.396840 130577495841920 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 22:21:55.397123 130577495841920 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_false/checkpoints/9. I0415 22:21:55.565801 130469918148160 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9 I0415 22:21:55.571609 130577495841920 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 22:21:55.571696 130577495841920 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 22:21:56.292486 130469949605440 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9/items I0415 22:22:03.066149 130577495841920 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 7.494876s I0415 22:22:03.066678 130577495841920 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.627 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host) I0415 22:22:03.066727 130577495841920 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 7.585992s (batch_requests_ready=0.087627s, total_serialization_initiated=7.497948s, others=0.000417s) I0415 22:22:03.066788 130577495841920 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 7.586454s (all_items=0.000010s, per_item={'items': '0.00001025'}, temp_paths=7.586444) I0415 22:22:03.071322 130470062851648 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 22:22:03.071404 130577495841920 async_checkpointer.py:561] Finished blocking save. Time taken: 7.674516s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9. I0415 22:22:03.071565 130577495841920 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 22:22:03.071731 130469989451328 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 22:22:03.071840 130577495841920 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776291675.4271946, 'wait_for_prev_duration_secs': 39.96943140029907, 'time_between_consecutive_saves_sec': 13.24814224243164, 'checkpointer_blocking_start_time': 1776291715.3968706, 'checkpointer_blocking_duration_secs': 7.674637317657471, 'get_old_steps_start_time': 1776291723.0715187, 'get_old_steps_duration_secs': 1.6689300537109375e-05, 'checkpoint_manager_blocking_start_time': 1776291675.4271622, 'checkpoint_manager_blocking_duration_secs': 47.64465308189392} I0415 22:22:03.071955 130577495841920 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 22:22:03.071986 130577495841920 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 22:22:03.776664 130469999937088 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_false/checkpoints/9/items/array_metadatas/process_0 I0415 22:22:32.890653 130469949605440 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.819959s (commit=29.380553s, array_metadata_write=0.439406s) I0415 22:22:32.891759 130470062851648 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 337.783 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 37 seconds) (per-host) I0415 22:22:32.891814 130470062851648 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 29.820368s. I0415 22:22:33.286471 130470062851648 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 22:22:33.844639 130470062851648 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.697816s. use_zarr3=False, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9/items I0415 22:22:33.845307 130470062851648 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9/items I0415 22:22:34.099190 130470062851648 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9 I0415 22:22:34.768970 130470062851648 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_false/checkpoints/9`. I0415 22:22:34.769600 130470062851648 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 39.372719s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_16_zarr3_false/checkpoints/9 I0415 22:22:34.769671 130470062851648 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 31.698226s. I0415 22:22:34.769847 130469989451328 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 22:22:34.769948 130469989451328 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 22:22:34.770004 130469989451328 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 22:22:34.770064 130469989451328 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 22:22:34.770219 130577495841920 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 22:22:34.770372 130577495841920 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 22:22:34.771323 130577495841920 metric_logger.py:185] completed step: 9, seconds: 0.119, TFLOP/s/device: 114.533, Tokens/s/device: 17263.616, 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