2026-04-16 20:24:57.662705: 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 20:24:57.782062 135525424213120 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0416 20:25:55.590708 135525424213120 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 20:25:55.590813 135525424213120 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 20:25:55.590854 135525424213120 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 20:25:55.590879 135525424213120 train_utils.py:364] WARNING: 'dataset_path' might be pointing your local file system I0416 20:25:55.590901 135525424213120 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0416 20:25:55.590973 135525424213120 train.py:811] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 20:25:55.682585 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 20:25:56.079288 135525424213120 maxtext_utils.py:1687] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 20:25:56.178909 135525424213120 checkpointing.py:688] Setting up checkpoint logger... I0416 20:25:56.179009 135525424213120 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 20:25:56.179065 135525424213120 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 20:25:56.179255 135525424213120 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 0x7b41e1dd3410>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 20:25:58.560027 135525424213120 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0416 20:25:58.560234 135525424213120 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 0x7b3b66394350>}, handler_registry=None I0416 20:25:58.560483 135525424213120 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b3b66394350>` 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 20:25:58.560527 135525424213120 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b3b65bac350>` 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 20:25:58.560555 135525424213120 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 0x7b3b66394350>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7b3b66394350>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b3b65bac350>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7b3b65bac350>}). I0416 20:25:58.560873 135525424213120 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 20:25:58.560922 135525424213120 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7b3b65b75760> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 20:25:58.680994 135525424213120 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints I0416 20:25:58.681249 135525424213120 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/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7b3b64b165a0> I0416 20:25:58.681329 135525424213120 checkpointing.py:302] Checkpoint manager created! W0416 20:25:58.697256 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 20:25:58.908167 135525424213120 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0416 20:25:58.908273 135525424213120 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. W0416 20:25:59.018534 4103204 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 20:25:59.105443 135525424213120 maxtext_utils.py:1805] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0416 20:25:59.105546 135525424213120 maxtext_utils.py:1805] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0416 20:25:59.105593 135525424213120 maxtext_utils.py:1805] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0416 20:25:59.105628 135525424213120 maxtext_utils.py:1805] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0416 20:25:59.105657 135525424213120 maxtext_utils.py:1805] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0416 20:25:59.105683 135525424213120 maxtext_utils.py:1805] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0416 20:25:59.105710 135525424213120 maxtext_utils.py:1805] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 20:25:59.105736 135525424213120 maxtext_utils.py:1805] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0416 20:25:59.105763 135525424213120 maxtext_utils.py:1805] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 20:25:59.105789 135525424213120 maxtext_utils.py:1805] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0416 20:25:59.105815 135525424213120 maxtext_utils.py:1805] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0416 20:25:59.105840 135525424213120 maxtext_utils.py:1805] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0416 20:25:59.209214 135525424213120 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 20:25:59.209296 135525424213120 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 20:25:59.214525 135525424213120 nnx_decoders.py:465] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 20:25:59.214570 135525424213120 nnx_decoders.py:465] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 20:25:59.240379 135525424213120 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 20:25:59.240498 135525424213120 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 20:25:59.254330 135525424213120 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 20:25:59.254378 135525424213120 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 20:25:59.276503 135525424213120 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 20:25:59.276562 135525424213120 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 20:25:59.290361 135525424213120 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 20:25:59.290410 135525424213120 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 20:25:59.303947 135525424213120 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 20:25:59.303995 135525424213120 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 20:25:59.331243 135525424213120 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 20:25:59.331300 135525424213120 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 20:25:59.349586 135525424213120 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 20:25:59.349640 135525424213120 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). W0416 20:26:00.090636 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 20:26:00.208793 135525424213120 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 20:26:00.209321 135525424213120 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 20:26:00.210682 135525424213120 metric_logger.py:289] number parameters: 1.104 billion W0416 20:26:00.831524 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 20:26:00.930236 135525424213120 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0416 20:26:01.046165 135525424213120 checkpointing.py:798] Waited 0.11591863632202148 seconds for step 0 to finish before starting checkpointing. I0416 20:26:01.046584 135525424213120 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 20:26:01.046733 135525424213120 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0416 20:26:01.047126 135525424213120 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0. I0416 20:26:01.129330 135525424213120 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 20:26:01.216009 135417974228544 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0 I0416 20:26:01.228799 135525424213120 jax_array_handlers.py:347] Scheduling D2H of 69 prioritized jax.Array. I0416 20:26:01.228890 135525424213120 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 20:26:01.869099 135417963742784 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/items W0416 20:26:01.892073 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.905320 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.916285 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.927941 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.935525 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.946115 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.951896 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:01.956821 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 20:26:02.039507 135417923896896 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776371161775802472, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776371161775802472, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0416 20:26:02.163920 135525424213120 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.936121s I0416 20:26:02.164244 135525424213120 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 11.942 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0416 20:26:02.164293 135525424213120 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.033511s (batch_requests_ready=0.089404s, total_serialization_initiated=0.943892s, others=0.000215s) I0416 20:26:02.164353 135525424213120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.034049s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=1.034029) I0416 20:26:02.165335 135417902925376 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 20:26:02.165423 135525424213120 async_checkpointer.py:561] Finished blocking save. Time taken: 1.118651s. Continuing background save to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0. I0416 20:26:02.165601 135525424213120 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 20:26:02.165735 135417984714304 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 20:26:02.165807 135525424213120 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776371161.046572, 'wait_for_prev_duration_secs': 3.9577484130859375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776371161.0467536, 'checkpointer_blocking_duration_secs': 1.1187775135040283, 'get_old_steps_start_time': 1776371162.1655493, 'get_old_steps_duration_secs': 2.09808349609375e-05, 'checkpoint_manager_blocking_start_time': 1776371161.04649, 'checkpoint_manager_blocking_duration_secs': 1.119290828704834} I0416 20:26:02.165882 135525424213120 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0416 20:26:02.165930 135525424213120 max_utils.py:750] Memstats: After params initialized: I0416 20:26:02.166055 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0416 20:26:02.166084 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 20:26:02.166109 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 20:26:02.166147 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 20:26:02.166166 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 20:26:02.166189 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 20:26:02.166208 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 20:26:02.166230 135525424213120 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 20:26:02.171725 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:02.177204 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 20:26:02.181953 4103204 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 20:26:02.526497 4105545 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 20:26:02.526554 4105545 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 20:26:02.568539 135525424213120 metric_logger.py:185] completed step: 0, seconds: 0.718, TFLOP/s/device: 18.917, Tokens/s/device: 2851.331, total_weights: 16384, loss: 10.881 I0416 20:26:02.570275 135525424213120 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/tensorboard/' I0416 20:26:02.653808 135525424213120 metric_logger.py:185] completed step: 1, seconds: 1.607, TFLOP/s/device: 8.454, Tokens/s/device: 1274.336, total_weights: 16384, loss: 10.881 I0416 20:26:02.773698 135525424213120 metric_logger.py:185] completed step: 2, seconds: 0.047, TFLOP/s/device: 291.694, Tokens/s/device: 43967.368, total_weights: 16384, loss: 9.779 I0416 20:26:02.892568 135525424213120 metric_logger.py:185] completed step: 3, seconds: 0.077, TFLOP/s/device: 176.617, Tokens/s/device: 26621.604, total_weights: 16384, loss: 8.837 I0416 20:26:03.320327 135417974228544 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0 I0416 20:26:22.819321 135525424213120 metric_logger.py:185] completed step: 4, seconds: 0.121, TFLOP/s/device: 112.602, Tokens/s/device: 16972.610, total_weights: 16384, loss: 7.961 I0416 20:26:22.828760 135525424213120 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 114.970, Tokens/s/device: 17329.497, total_weights: 16384, loss: 7.226 I0416 20:26:22.943555 135525424213120 metric_logger.py:185] completed step: 6, seconds: 19.927, TFLOP/s/device: 0.682, Tokens/s/device: 102.774, total_weights: 16384, loss: 6.669 I0416 20:26:23.061814 135525424213120 metric_logger.py:185] completed step: 7, seconds: 0.007, TFLOP/s/device: 2016.792, Tokens/s/device: 303992.875, total_weights: 16384, loss: 6.286 I0416 20:26:23.179785 135525424213120 metric_logger.py:185] completed step: 8, seconds: 0.116, TFLOP/s/device: 117.455, Tokens/s/device: 17704.164, total_weights: 16384, loss: 6.048 I0416 20:26:23.298618 135525424213120 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0416 20:26:23.299647 135525424213120 checkpointing.py:798] Waited 0.0010499954223632812 seconds for step 9 to finish before starting checkpointing. I0416 20:26:23.299935 135525424213120 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 20:26:37.440520 135417913411136 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 34.922279s (commit=34.444151s, array_metadata_write=0.478128s) I0416 20:26:37.441588 135417902925376 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 348.018 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 36 seconds) (per-host) I0416 20:26:37.441652 135417902925376 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.276178s. I0416 20:26:37.673656 135417902925376 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776371161775802472, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0416 20:26:37.843113 135417902925376 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 20:26:38.072190 135417923896896 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776371161775802472, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0416 20:26:38.327471 135417902925376 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 20:26:38.328094 135417902925376 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.617228s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/items I0416 20:26:38.328751 135417902925376 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0/items I0416 20:26:38.586007 135417902925376 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0 I0416 20:26:39.335072 135417902925376 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0`. I0416 20:26:39.335696 135417902925376 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 38.288927s. directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/0 I0416 20:26:39.335777 135417902925376 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 37.170308s. I0416 20:26:39.335897 135417984714304 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 20:26:39.335949 135417984714304 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 20:26:39.336006 135417984714304 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 20:26:39.336062 135417984714304 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 20:26:39.336208 135525424213120 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 20:26:39.336346 135525424213120 checkpoint_manager.py:1441] Waiting for previous save to complete took 16.036429 seconds. If this number is high, consider checkpointing less frequently. I0416 20:26:39.337568 135525424213120 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0416 20:26:39.337890 135525424213120 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9. I0416 20:26:39.512258 135525424213120 jax_array_handlers.py:347] Scheduling D2H of 69 prioritized jax.Array. I0416 20:26:39.512404 135417984714304 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9 I0416 20:26:39.512491 135525424213120 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 20:26:40.232686 135417953257024 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9/items I0416 20:26:50.569731 135525424213120 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.058827s I0416 20:26:50.575604 135525424213120 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.106 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host) I0416 20:26:50.575672 135525424213120 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 11.155930s (batch_requests_ready=0.083664s, total_serialization_initiated=11.066498s, others=0.005768s) I0416 20:26:50.575744 135525424213120 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 11.156450s (all_items=0.000024s, per_item={'items': '0.00002432'}, temp_paths=11.156425) I0416 20:26:50.577092 135417871468096 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 20:26:50.577217 135525424213120 async_checkpointer.py:561] Finished blocking save. Time taken: 11.239587s. Continuing background save to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9. I0416 20:26:50.577481 135525424213120 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 20:26:50.577656 135417902925376 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 20:26:50.577736 135525424213120 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776371183.299881, 'wait_for_prev_duration_secs': 16.036428689956665, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776371199.3376062, 'checkpointer_blocking_duration_secs': 11.239773035049438, 'get_old_steps_start_time': 1776371210.577395, 'get_old_steps_duration_secs': 4.696846008300781e-05, 'checkpoint_manager_blocking_start_time': 1776371183.2998233, 'checkpoint_manager_blocking_duration_secs': 27.277886629104614} I0416 20:26:50.577852 135525424213120 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0416 20:26:50.577891 135525424213120 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 20:26:51.359733 135417974228544 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0 I0416 20:27:34.574027 135417984714304 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.997698s (commit=43.543012s, array_metadata_write=0.454686s) I0416 20:27:34.575097 135417871468096 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 229.113 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 55 seconds) (per-host) I0416 20:27:34.575207 135417871468096 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.997944s. I0416 20:27:35.020453 135417871468096 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 20:27:35.449141 135417871468096 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 20:27:35.449763 135417871468096 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.570129s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9/items I0416 20:27:35.450411 135417871468096 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9/items I0416 20:27:35.712195 135417871468096 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9 I0416 20:27:36.417711 135417871468096 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9`. I0416 20:27:36.418370 135417871468096 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 57.080749s. directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_true/checkpoints/9 I0416 20:27:36.418448 135417871468096 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 45.841188s. I0416 20:27:36.418545 135417902925376 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 20:27:36.418594 135417902925376 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 20:27:36.418650 135417902925376 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 20:27:36.418695 135417902925376 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 20:27:36.418838 135525424213120 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 20:27:36.418980 135525424213120 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 20:27:36.419570 135525424213120 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.993, Tokens/s/device: 17333.017, total_weights: 16384, loss: 5.908 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops