2026-04-15 23:52:34.053461: 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 23:52:34.168429 130920885755008 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 23:53:28.624748 130920885755008 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 23:53:28.624859 130920885755008 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 23:53:28.624899 130920885755008 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 23:53:28.624925 130920885755008 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 23:53:28.624949 130920885755008 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0415 23:53:28.625077 130920885755008 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 23:53:28.720358 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:53:29.125311 130920885755008 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 23:53:29.125650 130920885755008 checkpointing.py:677] Setting up checkpoint logger... I0415 23:53:29.125701 130920885755008 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0415 23:53:29.125776 130920885755008 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 23:53:29.126465 130920885755008 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 0x7711ce90d8e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 23:53:31.540932 130920885755008 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 23:53:31.541179 130920885755008 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 0x770b525460f0>}, handler_registry=None I0415 23:53:31.541542 130920885755008 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x770b525460f0>` 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 23:53:31.541586 130920885755008 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x770b52545af0>` 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 23:53:31.541617 130920885755008 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 0x770b525460f0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x770b525460f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x770b52545af0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x770b52545af0>}). I0415 23:53:31.541915 130920885755008 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 23:53:31.542009 130920885755008 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x770b5253d4e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 23:53:31.662744 130920885755008 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints I0415 23:53:31.662985 130920885755008 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=10), 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_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x770b52f51b80> I0415 23:53:31.663085 130920885755008 checkpointing.py:301] Checkpoint manager created! W0415 23:53:31.680970 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:53:31.859924 130920885755008 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 23:53:31.860025 130920885755008 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 23:53:31.951343 130920885755008 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 23:53:31.951422 130920885755008 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 23:53:31.965975 130920885755008 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 23:53:31.966024 130920885755008 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 23:53:31.987153 130920885755008 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 23:53:31.987207 130920885755008 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:53:32.001792 130920885755008 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 23:53:32.001842 130920885755008 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:53:32.016474 130920885755008 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 23:53:32.016523 130920885755008 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:53:32.038722 130920885755008 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 23:53:32.038775 130920885755008 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:53:32.057363 130920885755008 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0415 23:53:32.057417 130920885755008 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0415 23:53:32.238399 130920885755008 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 23:53:32.238491 130920885755008 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 23:53:32.288538 3398616 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 23:53:32.666459 130920885755008 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 23:53:32.666563 130920885755008 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0415 23:53:32.666609 130920885755008 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0415 23:53:32.666654 130920885755008 maxtext_utils.py:1623] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0415 23:53:32.666694 130920885755008 maxtext_utils.py:1623] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0415 23:53:32.666725 130920885755008 maxtext_utils.py:1623] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0415 23:53:32.666765 130920885755008 maxtext_utils.py:1623] 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 23:53:32.666802 130920885755008 maxtext_utils.py:1623] 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 23:53:32.666829 130920885755008 maxtext_utils.py:1623] 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 23:53:32.666853 130920885755008 maxtext_utils.py:1623] 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 23:53:32.666886 130920885755008 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0415 23:53:32.666918 130920885755008 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0415 23:53:33.085401 130920885755008 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 23:53:33.085481 130920885755008 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0415 23:53:33.099117 130920885755008 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 23:53:33.099167 130920885755008 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0415 23:53:33.753206 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:53:33.873119 130920885755008 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 23:53:33.873865 130920885755008 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 23:53:33.874073 130920885755008 metric_logger.py:289] number parameters: 1.104 billion W0415 23:53:34.730081 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:53:34.828771 130920885755008 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 23:53:34.945928 130920885755008 checkpointing.py:776] Waited 0.1171426773071289 seconds for step 0 to finish before starting checkpointing. I0415 23:53:34.946503 130920885755008 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 23:53:34.946680 130920885755008 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 23:53:34.947231 130920885755008 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0. I0415 23:53:35.034017 130920885755008 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 23:53:35.121386 130920885755008 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 23:53:35.121475 130920885755008 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 23:53:35.129970 130813335176768 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0 W0415 23:53:35.335584 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.350005 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.364387 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.386201 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.397097 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.403020 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.407991 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:35.412861 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:53:35.775650 130813324691008 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/items I0415 23:53:35.945418 130813293233728 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776297215685103495, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776297215685103495, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0415 23:53:36.264794 3400809 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 23:53:36.264844 3400809 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 23:53:36.348865 130920885755008 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.227937s I0415 23:53:36.353222 130920885755008 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 9.364 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0415 23:53:36.353285 130920885755008 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.317970s (batch_requests_ready=0.081866s, total_serialization_initiated=1.231853s, others=0.004251s) I0415 23:53:36.353359 130920885755008 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.318860s (all_items=0.000029s, per_item={'items': '0.00002861'}, temp_paths=1.318831) I0415 23:53:36.354669 130813219833408 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 23:53:36.354754 130920885755008 async_checkpointer.py:561] Finished blocking save. Time taken: 1.408031s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0. I0415 23:53:36.354963 130920885755008 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 23:53:36.355180 130813345662528 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 23:53:36.355283 130920885755008 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776297214.946486, 'wait_for_prev_duration_secs': 4.792213439941406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776297214.9467018, 'checkpointer_blocking_duration_secs': 1.40816068649292, 'get_old_steps_start_time': 1776297216.3548844, 'get_old_steps_duration_secs': 3.933906555175781e-05, 'checkpoint_manager_blocking_start_time': 1776297214.946358, 'checkpoint_manager_blocking_duration_secs': 1.4089007377624512} I0415 23:53:36.355392 130920885755008 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 23:53:36.355432 130920885755008 max_utils.py:750] Memstats: After params initialized: I0415 23:53:36.355477 130920885755008 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0)) I0415 23:53:36.355502 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0415 23:53:36.355520 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0415 23:53:36.355538 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0415 23:53:36.355555 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0415 23:53:36.355571 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0415 23:53:36.355588 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0415 23:53:36.355606 130920885755008 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0415 23:53:36.361233 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:36.367597 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:53:36.372203 3398616 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:53:36.710868 130920885755008 metric_logger.py:185] completed step: 0, seconds: 0.954, TFLOP/s/device: 14.240, Tokens/s/device: 2146.458, total_weights: 16384, loss: 10.874 I0415 23:53:36.711894 130920885755008 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/tensorboard/' I0415 23:53:36.826712 130920885755008 metric_logger.py:185] completed step: 1, seconds: 1.880, TFLOP/s/device: 7.226, Tokens/s/device: 1089.255, total_weights: 16384, loss: 10.874 I0415 23:53:36.945857 130920885755008 metric_logger.py:185] completed step: 2, seconds: 0.007, TFLOP/s/device: 2009.039, Tokens/s/device: 302824.190, total_weights: 16384, loss: 9.775 I0415 23:53:37.064403 130920885755008 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.455, Tokens/s/device: 17553.333, total_weights: 16384, loss: 8.838 I0415 23:53:37.350598 130813251290688 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0 I0415 23:53:56.045837 130920885755008 metric_logger.py:185] completed step: 4, seconds: 0.119, TFLOP/s/device: 114.089, Tokens/s/device: 17196.789, total_weights: 16384, loss: 7.965 I0415 23:53:56.053139 130920885755008 metric_logger.py:185] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.579, Tokens/s/device: 17270.604, total_weights: 16384, loss: 7.233 I0415 23:53:56.170430 130920885755008 metric_logger.py:185] completed step: 6, seconds: 18.982, TFLOP/s/device: 0.716, Tokens/s/device: 107.891, total_weights: 16384, loss: 6.676 I0415 23:53:56.288989 130920885755008 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2882.905, Tokens/s/device: 434542.754, total_weights: 16384, loss: 6.293 I0415 23:53:56.407469 130920885755008 metric_logger.py:185] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.246, Tokens/s/device: 17371.095, total_weights: 16384, loss: 6.055 I0415 23:53:56.525881 130920885755008 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 23:53:56.526706 130920885755008 checkpointing.py:776] Waited 0.0008432865142822266 seconds for step 9 to finish before starting checkpointing. I0415 23:53:56.526977 130920885755008 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 23:53:59.546230 130813240804928 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 23.192300s (commit=22.726130s, array_metadata_write=0.466170s) I0415 23:53:59.547346 130813219833408 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 515.535 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 24 seconds) (per-host) I0415 23:53:59.547404 130813219833408 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 23.192607s. I0415 23:53:59.808077 130813219833408 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776297215685103495, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0415 23:54:00.016125 130813219833408 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 23:54:00.188593 130813293233728 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776297215685103495, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0415 23:54:00.405371 130813219833408 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 23:54:00.405956 130813219833408 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.556664s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/items I0415 23:54:00.406624 130813219833408 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0/items I0415 23:54:00.641865 130813219833408 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0 I0415 23:54:01.302639 130813219833408 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0`. I0415 23:54:01.303328 130813219833408 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 26.356612s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/0 I0415 23:54:01.303405 130813219833408 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 24.948610s. I0415 23:54:01.303548 130813345662528 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 23:54:01.303656 130813345662528 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 23:54:01.303718 130813345662528 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 23:54:01.303762 130813345662528 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 23:54:01.304000 130920885755008 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 23:54:01.304133 130920885755008 checkpoint_manager.py:1441] Waiting for previous save to complete took 4.777171 seconds. If this number is high, consider checkpointing less frequently. I0415 23:54:01.304847 130920885755008 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 23:54:01.305146 130920885755008 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9. I0415 23:54:01.458625 130813345662528 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9 I0415 23:54:01.480492 130920885755008 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 23:54:01.480583 130920885755008 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 23:54:02.140447 130813303719488 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9/items I0415 23:54:06.368404 130920885755008 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 4.888293s I0415 23:54:06.368927 130920885755008 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.478 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 4 seconds) (per-host) I0415 23:54:06.368979 130920885755008 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 4.980135s (batch_requests_ready=0.088342s, total_serialization_initiated=4.891376s, others=0.000417s) I0415 23:54:06.369054 130920885755008 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 4.980624s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=4.980606) I0415 23:54:06.373700 130813209347648 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 23:54:06.373800 130920885755008 async_checkpointer.py:561] Finished blocking save. Time taken: 5.068910s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9. I0415 23:54:06.374018 130920885755008 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 23:54:06.374246 130813261776448 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 23:54:06.374329 130920885755008 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776297236.5269356, 'wait_for_prev_duration_secs': 4.777170896530151, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776297241.3048694, 'checkpointer_blocking_duration_secs': 5.069065093994141, 'get_old_steps_start_time': 1776297246.373951, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776297236.5268893, 'checkpoint_manager_blocking_duration_secs': 9.847415447235107} I0415 23:54:06.374437 130920885755008 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 23:54:06.374471 130920885755008 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 23:54:07.197605 130813272262208 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0 I0415 23:54:39.530996 130813303719488 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 33.157625s (commit=32.721900s, array_metadata_write=0.435725s) I0415 23:54:39.532243 130813209347648 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 331.297 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 38 seconds) (per-host) I0415 23:54:39.532351 130813209347648 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 33.158507s. I0415 23:54:39.952748 130813209347648 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 23:54:40.363146 130813209347648 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 23:54:40.363360 130813209347648 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.553052s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9/items I0415 23:54:40.364013 130813209347648 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9/items I0415 23:54:40.599224 130813209347648 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9 I0415 23:54:41.262499 130813209347648 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9`. I0415 23:54:41.263157 130813209347648 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 39.958273s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_13_scan_layers_true/checkpoints/9 I0415 23:54:41.263237 130813209347648 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 34.889396s. I0415 23:54:41.263372 130813261776448 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 23:54:41.263424 130813261776448 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 23:54:41.263487 130813261776448 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 23:54:41.263525 130813261776448 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 23:54:41.263670 130920885755008 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 23:54:41.263810 130920885755008 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 23:54:41.264732 130920885755008 metric_logger.py:185] completed step: 9, seconds: 0.119, TFLOP/s/device: 114.486, Tokens/s/device: 17256.633, total_weights: 16384, loss: 5.915 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops