2026-04-16 00:14:30.516256: 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 00:14:30.632722 127805476146304 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 00:15:31.346496 127805476146304 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 00:15:31.346607 127805476146304 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 00:15:31.346641 127805476146304 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 00:15:31.346665 127805476146304 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 00:15:31.346689 127805476146304 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 00:15:31.346758 127805476146304 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 00:15:31.437484 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:15:31.843750 127805476146304 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 00:15:31.844018 127805476146304 checkpointing.py:677] Setting up checkpoint logger... I0416 00:15:31.844082 127805476146304 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 00:15:31.844123 127805476146304 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 00:15:31.844567 127805476146304 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 0x743c71b4adb0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 00:15:34.325325 127805476146304 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 00:15:34.325499 127805476146304 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 0x7435f533f110>}, handler_registry=None I0416 00:15:34.325762 127805476146304 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7435f533f110>` 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 00:15:34.325803 127805476146304 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7435f533dac0>` 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 00:15:34.325832 127805476146304 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 0x7435f533f110>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7435f533f110>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7435f533dac0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7435f533dac0>}). I0416 00:15:34.326259 127805476146304 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 00:15:34.326308 127805476146304 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7435f5349580> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 00:15:34.448555 127805476146304 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_16_zarr3_true/checkpoints I0416 00:15:34.448775 127805476146304 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_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7437b65f0d10> I0416 00:15:34.448856 127805476146304 checkpointing.py:301] Checkpoint manager created! W0416 00:15:34.465214 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:15:34.647332 127805476146304 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 00:15:34.647530 127805476146304 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:15:34.744645 127805476146304 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:15:34.744722 127805476146304 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:15:34.758888 127805476146304 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 00:15:34.758941 127805476146304 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 00:15:34.779721 127805476146304 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 00:15:34.779775 127805476146304 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:15:34.793984 127805476146304 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 00:15:34.794033 127805476146304 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:15:34.808228 127805476146304 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 00:15:34.808276 127805476146304 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:15:34.829876 127805476146304 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 00:15:34.829939 127805476146304 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 00:15:34.848128 127805476146304 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 00:15:34.848180 127805476146304 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 00:15:35.028720 127805476146304 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 00:15:35.028835 127805476146304 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0416 00:15:35.079236 3419600 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 00:15:35.458907 127805476146304 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 00:15:35.459014 127805476146304 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) I0416 00:15:35.459068 127805476146304 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) I0416 00:15:35.459114 127805476146304 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') I0416 00:15:35.459153 127805476146304 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) I0416 00:15:35.459180 127805476146304 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) I0416 00:15:35.459218 127805476146304 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) I0416 00:15:35.459256 127805476146304 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') I0416 00:15:35.459285 127805476146304 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) I0416 00:15:35.459311 127805476146304 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) I0416 00:15:35.459344 127805476146304 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0416 00:15:35.459376 127805476146304 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0416 00:15:35.879497 127805476146304 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:15:35.879585 127805476146304 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 00:15:35.893181 127805476146304 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 00:15:35.893232 127805476146304 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 00:15:36.561128 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:15:36.689213 127805476146304 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 00:15:36.689692 127805476146304 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 00:15:36.689829 127805476146304 metric_logger.py:289] number parameters: 1.104 billion W0416 00:15:37.538013 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:15:37.634489 127805476146304 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0416 00:15:37.751456 127805476146304 checkpointing.py:776] Waited 0.11695361137390137 seconds for step 0 to finish before starting checkpointing. I0416 00:15:37.751878 127805476146304 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:15:37.752025 127805476146304 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0416 00:15:37.752396 127805476146304 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_16_zarr3_true/checkpoints/0. I0416 00:15:37.849920 127805476146304 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 00:15:37.941754 127805476146304 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:15:37.941841 127805476146304 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 00:15:37.947449 127697919411776 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/0 W0416 00:15:38.154340 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.168212 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.176456 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.186360 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.199727 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.207720 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.213025 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.217928 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:15:38.659630 127697908926016 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/0/items I0416 00:15:38.790897 127697877468736 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776298538562580504, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776298538562580504, "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_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0416 00:15:38.946770 127805476146304 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.005465s I0416 00:15:38.947359 127805476146304 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 11.256 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0416 00:15:38.947520 127805476146304 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.096651s (batch_requests_ready=0.086830s, total_serialization_initiated=1.009305s, others=0.000516s) I0416 00:15:38.947641 127805476146304 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.097248s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=1.097230) I0416 00:15:38.948497 127697856497216 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:15:38.948594 127805476146304 async_checkpointer.py:561] Finished blocking save. Time taken: 1.196515s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/0. I0416 00:15:38.948797 127805476146304 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:15:38.949072 127697929897536 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:15:38.949179 127805476146304 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776298537.7518654, 'wait_for_prev_duration_secs': 4.029273986816406e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776298537.7520597, 'checkpointer_blocking_duration_secs': 1.1966559886932373, 'get_old_steps_start_time': 1776298538.9487379, 'get_old_steps_duration_secs': 2.2172927856445312e-05, 'checkpoint_manager_blocking_start_time': 1776298537.7517827, 'checkpoint_manager_blocking_duration_secs': 1.1973669528961182} I0416 00:15:38.949295 127805476146304 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0416 00:15:38.949337 127805476146304 max_utils.py:750] Memstats: After params initialized: I0416 00:15:38.949436 127805476146304 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0)) I0416 00:15:38.949463 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 00:15:38.949484 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 00:15:38.949502 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 00:15:38.949520 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 00:15:38.949537 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 00:15:38.949555 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 00:15:38.949572 127805476146304 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 00:15:38.955039 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.960855 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 00:15:38.965913 3419600 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 00:15:39.152615 3422444 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 00:15:39.152689 3422444 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 00:15:39.178512 127805476146304 metric_logger.py:185] completed step: 0, seconds: 0.944, TFLOP/s/device: 14.390, Tokens/s/device: 2169.069, total_weights: 16384, loss: 10.874 I0416 00:15:39.179420 127805476146304 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_16_zarr3_true/tensorboard/' I0416 00:15:39.294523 127805476146304 metric_logger.py:185] completed step: 1, seconds: 1.542, TFLOP/s/device: 8.810, Tokens/s/device: 1327.951, total_weights: 16384, loss: 10.874 I0416 00:15:39.412987 127805476146304 metric_logger.py:185] completed step: 2, seconds: 0.006, TFLOP/s/device: 2417.209, Tokens/s/device: 364347.981, total_weights: 16384, loss: 10.276 I0416 00:15:39.531709 127805476146304 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.316, Tokens/s/device: 17532.445, total_weights: 16384, loss: 9.282 I0416 00:15:39.536129 127805476146304 checkpointing.py:772] Waiting for step 5 to finish before checkpoint... I0416 00:15:39.768290 127805476146304 checkpointing.py:776] Waited 0.23212933540344238 seconds for step 5 to finish before starting checkpointing. I0416 00:15:39.768684 127805476146304 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:15:39.909080 127697866982976 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_16_zarr3_true/checkpoints/0/items/array_metadatas/process_0 I0416 00:16:14.588012 127697898440256 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.433630s (commit=34.995073s, array_metadata_write=0.438557s) I0416 00:16:14.588968 127697856497216 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 343.970 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 36 seconds) (per-host) I0416 00:16:14.589021 127697856497216 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.640382s. I0416 00:16:14.814642 127697856497216 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776298538562580504, '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_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0416 00:16:15.030119 127697856497216 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 00:16:15.194519 127697877468736 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776298538562580504, '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_16_zarr3_true/checkpoints/0/_CHECKPOINT_METADATA I0416 00:16:15.458133 127697856497216 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 00:16:15.458747 127697856497216 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.602387s. 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_16_zarr3_true/checkpoints/0/items I0416 00:16:15.459437 127697856497216 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/0/items I0416 00:16:15.692937 127697856497216 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/0 I0416 00:16:16.379126 127697856497216 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_16_zarr3_true/checkpoints/0`. I0416 00:16:16.379751 127697856497216 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 38.627678s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/0 I0416 00:16:16.379826 127697856497216 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 37.431188s. I0416 00:16:16.380017 127697929897536 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:16:16.380133 127697929897536 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:16:16.380197 127697929897536 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:16:16.380245 127697929897536 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 00:16:16.380671 127805476146304 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 00:16:16.380795 127805476146304 checkpoint_manager.py:1441] Waiting for previous save to complete took 36.612111 seconds. If this number is high, consider checkpointing less frequently. I0416 00:16:16.381550 127805476146304 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5 I0416 00:16:16.381824 127805476146304 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_16_zarr3_true/checkpoints/5. I0416 00:16:16.562214 127805476146304 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:16:16.562352 127697929897536 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/5 I0416 00:16:16.562430 127805476146304 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 00:16:17.265688 127697785194048 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/5/items I0416 00:16:29.038474 127805476146304 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.476988s I0416 00:16:29.042911 127805476146304 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1005.228 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0416 00:16:29.042984 127805476146304 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.571167s (batch_requests_ready=0.086910s, total_serialization_initiated=12.479962s, others=0.004295s) I0416 00:16:29.043084 127805476146304 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.571692s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=12.571676) I0416 00:16:29.044523 127697816651328 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:16:29.044610 127805476146304 async_checkpointer.py:561] Finished blocking save. Time taken: 12.663009s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/5. I0416 00:16:29.044817 127805476146304 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:16:29.045008 127697856497216 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:16:29.045136 127805476146304 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776298539.768661, 'wait_for_prev_duration_secs': 36.61211061477661, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776298576.3815742, 'checkpointer_blocking_duration_secs': 12.66314172744751, 'get_old_steps_start_time': 1776298589.044737, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1776298539.7685962, 'checkpoint_manager_blocking_duration_secs': 49.27651047706604} I0416 00:16:29.045277 127805476146304 checkpointing.py:408] Started an asynchronous checkpoint save for step 5 I0416 00:16:29.045821 127805476146304 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.902, Tokens/s/device: 17319.239, total_weights: 16384, loss: 8.327 I0416 00:16:29.052706 127805476146304 metric_logger.py:185] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.591, Tokens/s/device: 17272.352, total_weights: 16384, loss: 7.367 I0416 00:16:29.168751 127805476146304 metric_logger.py:185] completed step: 6, seconds: 49.515, TFLOP/s/device: 0.274, Tokens/s/device: 41.361, total_weights: 16384, loss: 6.459 I0416 00:16:29.286749 127805476146304 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2548.224, Tokens/s/device: 384096.024, total_weights: 16384, loss: 5.622 I0416 00:16:29.405472 127805476146304 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 116.127, Tokens/s/device: 17503.974, total_weights: 16384, loss: 4.864 I0416 00:16:29.523410 127805476146304 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0416 00:16:29.524251 127805476146304 checkpointing.py:776] Waited 0.0008556842803955078 seconds for step 9 to finish before starting checkpointing. I0416 00:16:29.524521 127805476146304 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:16:29.761250 127697866982976 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_16_zarr3_true/checkpoints/5/items/array_metadatas/process_0 I0416 00:17:03.725283 127697898440256 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 34.681597s (commit=34.154809s, array_metadata_write=0.526788s) I0416 00:17:03.726243 127697816651328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 267.420 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 47 seconds) (per-host) I0416 00:17:03.726300 127697816651328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.681647s. I0416 00:17:04.175693 127697816651328 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 00:17:04.639226 127697816651328 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 00:17:04.639394 127697816651328 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.603233s. 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_16_zarr3_true/checkpoints/5/items I0416 00:17:04.640017 127697816651328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/5/items I0416 00:17:04.893625 127697816651328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/5 I0416 00:17:05.602680 127697816651328 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_16_zarr3_true/checkpoints/5`. I0416 00:17:05.603312 127697816651328 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 49.221724s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/5 I0416 00:17:05.603386 127697816651328 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.558733s. I0416 00:17:05.603527 127697856497216 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:17:05.603627 127697856497216 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:17:05.603689 127697856497216 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:17:05.603738 127697856497216 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0416 00:17:05.603965 127805476146304 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 00:17:05.604467 127805476146304 checkpoint_manager.py:1441] Waiting for previous save to complete took 36.079957 seconds. If this number is high, consider checkpointing less frequently. I0416 00:17:05.604690 127805476146304 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0416 00:17:05.604978 127805476146304 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_16_zarr3_true/checkpoints/9. I0416 00:17:05.794867 127697856497216 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/9 I0416 00:17:05.816221 127805476146304 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 00:17:05.816309 127805476146304 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 00:17:06.516261 127697806165568 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/9/items I0416 00:17:15.809237 127805476146304 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 9.993374s I0416 00:17:15.813486 127805476146304 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.221 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 10 seconds) (per-host) I0416 00:17:15.813543 127805476146304 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 10.110065s (batch_requests_ready=0.109430s, total_serialization_initiated=9.996520s, others=0.004116s) I0416 00:17:15.813614 127805476146304 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 10.110595s (all_items=0.000014s, per_item={'items': '0.00001407'}, temp_paths=10.110581) I0416 00:17:15.815144 127697856497216 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 00:17:15.815226 127805476146304 async_checkpointer.py:561] Finished blocking save. Time taken: 10.210493s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/9. I0416 00:17:15.815389 127805476146304 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 00:17:15.815556 127697816651328 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 00:17:15.815668 127805476146304 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_16_zarr3_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776298589.524494, 'wait_for_prev_duration_secs': 36.07995653152466, 'time_between_consecutive_saves_sec': 13.144232988357544, 'checkpointer_blocking_start_time': 1776298625.6047144, 'checkpointer_blocking_duration_secs': 10.210612773895264, 'get_old_steps_start_time': 1776298635.8153408, 'get_old_steps_duration_secs': 1.7881393432617188e-05, 'checkpoint_manager_blocking_start_time': 1776298589.524461, 'checkpoint_manager_blocking_duration_secs': 46.291181325912476} I0416 00:17:15.815776 127805476146304 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0416 00:17:15.815806 127805476146304 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 00:17:16.570008 127697929897536 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_16_zarr3_true/checkpoints/9/items/array_metadatas/process_0 I0416 00:17:53.634793 127697806165568 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.820541s (commit=37.393974s, array_metadata_write=0.426567s) I0416 00:17:53.635926 127697856497216 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 263.637 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 47 seconds) (per-host) I0416 00:17:53.635981 127697856497216 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.820714s. I0416 00:17:54.056560 127697856497216 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 00:17:54.492002 127697856497216 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 00:17:54.492206 127697856497216 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.586889s. 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_16_zarr3_true/checkpoints/9/items I0416 00:17:54.492835 127697856497216 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/9/items I0416 00:17:54.745534 127697856497216 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/9 I0416 00:17:55.441609 127697856497216 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_16_zarr3_true/checkpoints/9`. I0416 00:17:55.442236 127697856497216 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 49.837511s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_16_zarr3_true/checkpoints/9 I0416 00:17:55.442301 127697856497216 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 39.627034s. I0416 00:17:55.442476 127697816651328 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 00:17:55.442582 127697816651328 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 00:17:55.442643 127697816651328 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 00:17:55.442688 127697816651328 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 00:17:55.443028 127805476146304 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 00:17:55.443547 127805476146304 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 00:17:55.444574 127805476146304 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 115.401, Tokens/s/device: 17394.554, 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