2026-04-15 23:46:51.873422: 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 23:46:51.988541 133210023484544 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 23:47:46.939164 133210023484544 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 23:47:46.939270 133210023484544 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 23:47:46.939304 133210023484544 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 23:47:46.939330 133210023484544 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 23:47:46.939355 133210023484544 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0415 23:47:46.939434 133210023484544 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 23:47:47.031307 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:47:47.426970 133210023484544 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 23:47:47.427302 133210023484544 checkpointing.py:677] Setting up checkpoint logger... I0415 23:47:47.427347 133210023484544 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0415 23:47:47.427385 133210023484544 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 23:47:47.427925 133210023484544 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 0x7926c93cb2f0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 23:47:50.121327 133210023484544 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 23:47:50.121504 133210023484544 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 0x79204d94b6e0>}, handler_registry=None I0415 23:47:50.121750 133210023484544 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79204d94b6e0>` 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 23:47:50.121789 133210023484544 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79204d53a570>` 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 23:47:50.121820 133210023484544 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 0x79204d94b6e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x79204d94b6e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79204d53a570>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79204d53a570>}). I0415 23:47:50.122130 133210023484544 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 23:47:50.122180 133210023484544 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x79204d5314e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 23:47:50.240017 133210023484544 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_11_optimizer_offload_true/checkpoints I0415 23:47:50.240270 133210023484544 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_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x79204d539760> I0415 23:47:50.240350 133210023484544 checkpointing.py:301] Checkpoint manager created! W0415 23:47:50.256605 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:47:50.435690 133210023484544 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 23:47:50.435793 133210023484544 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 23:47:50.524721 133210023484544 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 23:47:50.524801 133210023484544 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 23:47:50.539152 133210023484544 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 23:47:50.539202 133210023484544 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0415 23:47:50.559786 133210023484544 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 23:47:50.559840 133210023484544 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:47:50.586161 133210023484544 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 23:47:50.586287 133210023484544 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:47:50.600897 133210023484544 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 23:47:50.600946 133210023484544 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:47:50.623960 133210023484544 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 23:47:50.624022 133210023484544 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0415 23:47:50.642443 133210023484544 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0415 23:47:50.642493 133210023484544 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0415 23:47:50.827637 133210023484544 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 23:47:50.827747 133210023484544 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 23:47:50.880190 3392714 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 23:47:51.744906 133210023484544 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 23:47:51.745019 133210023484544 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 23:47:51.745093 133210023484544 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 23:47:51.745143 133210023484544 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 23:47:51.745183 133210023484544 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 23:47:51.745213 133210023484544 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 23:47:51.745254 133210023484544 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 23:47:51.745292 133210023484544 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 23:47:51.745321 133210023484544 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 23:47:51.745346 133210023484544 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 23:47:51.745379 133210023484544 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0415 23:47:51.745413 133210023484544 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0415 23:47:52.489054 133210023484544 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 23:47:52.489135 133210023484544 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0415 23:47:52.503240 133210023484544 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 23:47:52.503290 133210023484544 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0415 23:47:53.287461 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:47:53.431878 133210023484544 max_utils.py:791] Total memory size: 2.4 GB, Output size: 0.5 GB, Temp size: 1.9 GB, Argument size: 0.5 GB, Host temp size: 0.0 GB. I0415 23:47:54.848174 133210023484544 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 23:47:54.848454 133210023484544 metric_logger.py:289] number parameters: 1.104 billion W0415 23:47:55.693058 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:47:55.814204 133210023484544 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 23:47:56.001231 133210023484544 checkpointing.py:776] Waited 0.1870594024658203 seconds for step 0 to finish before starting checkpointing. I0415 23:47:56.001920 133210023484544 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 23:47:56.002150 133210023484544 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 23:47:56.002795 133210023484544 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_11_optimizer_offload_true/checkpoints/0. I0415 23:47:56.079180 133210023484544 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 23:47:56.158393 133106946278976 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/0 I0415 23:47:56.180659 133210023484544 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 23:47:56.180749 133210023484544 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 W0415 23:47:56.395965 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.406331 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.420012 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.431916 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.443916 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.450499 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.455003 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.459284 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:47:56.629174 133210023484544 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.448939s I0415 23:47:56.629526 133210023484544 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 22.486 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 548 milliseconds) (per-host) I0415 23:47:56.629580 133210023484544 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.548952s (batch_requests_ready=0.095946s, total_serialization_initiated=0.452745s, others=0.000261s) I0415 23:47:56.629642 133210023484544 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.549972s (all_items=0.000038s, per_item={'items': '0.00003791'}, temp_paths=0.549934) I0415 23:47:56.630446 133106883364416 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 23:47:56.630542 133210023484544 async_checkpointer.py:561] Finished blocking save. Time taken: 0.628345s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/0. I0415 23:47:56.630739 133210023484544 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 23:47:56.630911 133106956764736 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 23:47:56.631026 133210023484544 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_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776296876.0018988, 'wait_for_prev_duration_secs': 6.008148193359375e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776296876.0021741, 'checkpointer_blocking_duration_secs': 0.6284677982330322, 'get_old_steps_start_time': 1776296876.630665, 'get_old_steps_duration_secs': 3.457069396972656e-05, 'checkpoint_manager_blocking_start_time': 1776296876.0017245, 'checkpoint_manager_blocking_duration_secs': 0.6292603015899658} I0415 23:47:56.631155 133210023484544 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 23:47:56.631200 133210023484544 max_utils.py:750] Memstats: After params initialized: I0415 23:47:56.631263 133210023484544 max_utils.py:756] Using (GB) 0.57 / 31.25 (1.824000%) on TPU_0(process=0,(0,0,0,0)) I0415 23:47:56.631289 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_1(process=0,(1,0,0,0)) I0415 23:47:56.631311 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_2(process=0,(0,1,0,0)) I0415 23:47:56.631335 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_3(process=0,(1,1,0,0)) I0415 23:47:56.631354 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_4(process=0,(0,2,0,0)) I0415 23:47:56.631371 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_5(process=0,(1,2,0,0)) I0415 23:47:56.631389 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_6(process=0,(0,3,0,0)) I0415 23:47:56.631405 133210023484544 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_7(process=0,(1,3,0,0)) W0415 23:47:56.637094 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.642822 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:47:56.647625 3392714 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:47:56.964208 133210023484544 metric_logger.py:185] completed step: 0, seconds: 0.965, TFLOP/s/device: 14.078, Tokens/s/device: 2122.047, total_weights: 16384, loss: 10.874 I0415 23:47:56.964897 133210023484544 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_11_optimizer_offload_true/tensorboard/' I0415 23:47:57.039400 133106864490048 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776296876791837579, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776296876791837579, "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_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0415 23:47:57.087219 133106935793216 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/0/items I0415 23:47:57.149356 133210023484544 metric_logger.py:185] completed step: 1, seconds: 1.148, TFLOP/s/device: 11.837, Tokens/s/device: 1784.138, total_weights: 16384, loss: 10.874 I0415 23:47:57.336658 133210023484544 metric_logger.py:185] completed step: 2, seconds: 0.007, TFLOP/s/device: 2083.596, Tokens/s/device: 314062.260, total_weights: 16384, loss: 9.775 I0415 23:47:57.524714 133210023484544 metric_logger.py:185] completed step: 3, seconds: 0.186, TFLOP/s/device: 73.060, Tokens/s/device: 11012.351, total_weights: 16384, loss: 8.838 I0415 23:47:57.563785 3395038 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 23:47:57.563852 3395038 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 23:47:58.323669 133106904335936 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_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_0 I0415 23:48:17.698539 133210023484544 metric_logger.py:185] completed step: 4, seconds: 0.188, TFLOP/s/device: 72.266, Tokens/s/device: 10892.690, total_weights: 16384, loss: 7.965 I0415 23:48:17.706096 133210023484544 metric_logger.py:185] completed step: 5, seconds: 0.188, TFLOP/s/device: 72.402, Tokens/s/device: 10913.179, total_weights: 16384, loss: 7.233 I0415 23:48:17.891503 133210023484544 metric_logger.py:185] completed step: 6, seconds: 20.175, TFLOP/s/device: 0.673, Tokens/s/device: 101.514, total_weights: 16384, loss: 6.677 I0415 23:48:18.078794 133210023484544 metric_logger.py:185] completed step: 7, seconds: 0.004, TFLOP/s/device: 3026.761, Tokens/s/device: 456226.331, total_weights: 16384, loss: 6.293 I0415 23:48:18.266292 133210023484544 metric_logger.py:185] completed step: 8, seconds: 0.186, TFLOP/s/device: 72.919, Tokens/s/device: 10991.193, total_weights: 16384, loss: 6.055 I0415 23:48:18.453230 133210023484544 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 23:48:18.454117 133210023484544 checkpointing.py:776] Waited 0.0009174346923828125 seconds for step 9 to finish before starting checkpointing. I0415 23:48:18.454388 133210023484544 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 23:48:25.067770 133106893850176 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 27.509843s (commit=27.027271s, array_metadata_write=0.482572s) I0415 23:48:25.068904 133106883364416 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 435.928 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 28 seconds) (per-host) I0415 23:48:25.069010 133106883364416 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.438423s. I0415 23:48:25.301959 133106883364416 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776296876791837579, '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_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0415 23:48:25.480011 133106883364416 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 23:48:25.694692 133106864490048 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776296876791837579, '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_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0415 23:48:25.914237 133106883364416 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 23:48:25.914392 133106883364416 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.569589s. 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_11_optimizer_offload_true/checkpoints/0/items I0415 23:48:25.915038 133106883364416 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/0/items I0415 23:48:26.156707 133106883364416 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/0 I0415 23:48:26.823814 133106883364416 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_11_optimizer_offload_true/checkpoints/0`. I0415 23:48:26.824470 133106883364416 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 30.822279s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/0 I0415 23:48:26.824547 133106883364416 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.193962s. I0415 23:48:26.824717 133106956764736 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 23:48:26.824826 133106956764736 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 23:48:26.824889 133106956764736 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 23:48:26.824934 133106956764736 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 23:48:26.825050 133210023484544 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 23:48:26.825123 133210023484544 checkpoint_manager.py:1441] Waiting for previous save to complete took 8.370750 seconds. If this number is high, consider checkpointing less frequently. I0415 23:48:26.826292 133210023484544 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 23:48:26.826568 133210023484544 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_11_optimizer_offload_true/checkpoints/9. I0415 23:48:27.002127 133210023484544 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 23:48:27.002217 133210023484544 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 23:48:27.004399 133106956764736 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/9 I0415 23:48:27.293602 133210023484544 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.291947s I0415 23:48:27.294056 133210023484544 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 32.756 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 376 milliseconds) (per-host) I0415 23:48:27.294107 133210023484544 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.376846s (batch_requests_ready=0.081631s, total_serialization_initiated=0.294862s, others=0.000353s) I0415 23:48:27.294164 133210023484544 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.377360s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.377343) I0415 23:48:27.295109 133106801575488 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 23:48:27.295202 133210023484544 async_checkpointer.py:561] Finished blocking save. Time taken: 0.468868s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/9. I0415 23:48:27.295379 133210023484544 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 23:48:27.295653 133106749146688 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 23:48:27.295774 133210023484544 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_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776296898.4543483, 'wait_for_prev_duration_secs': 8.370750427246094, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776296906.8263147, 'checkpointer_blocking_duration_secs': 0.4689817428588867, 'get_old_steps_start_time': 1776296907.2953124, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1776296898.4543087, 'checkpoint_manager_blocking_duration_secs': 8.841434478759766} I0415 23:48:27.295889 133210023484544 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 23:48:27.295923 133210023484544 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 23:48:27.709835 133106935793216 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/9/items I0415 23:48:28.907241 133106822547008 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_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_0 I0415 23:49:13.915837 133106812061248 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 45.730035s (commit=45.314420s, array_metadata_write=0.415615s) I0415 23:49:13.916912 133106801575488 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 268.870 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 46 seconds) (per-host) I0415 23:49:13.917026 133106801575488 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.621773s. I0415 23:49:14.363418 133106801575488 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 23:49:14.773612 133106801575488 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 23:49:14.774199 133106801575488 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.585994s. 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_11_optimizer_offload_true/checkpoints/9/items I0415 23:49:14.774857 133106801575488 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/9/items I0415 23:49:15.026137 133106801575488 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/9 I0415 23:49:15.697009 133106801575488 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_11_optimizer_offload_true/checkpoints/9`. I0415 23:49:15.697677 133106801575488 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 48.871339s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_11_optimizer_offload_true/checkpoints/9 I0415 23:49:15.697765 133106801575488 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 48.402517s. I0415 23:49:15.697956 133106749146688 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 23:49:15.698081 133106749146688 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 23:49:15.698167 133106749146688 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 23:49:15.698214 133106749146688 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 23:49:15.698373 133210023484544 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 23:49:15.698502 133210023484544 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 23:49:15.699021 133210023484544 metric_logger.py:185] completed step: 9, seconds: 0.187, TFLOP/s/device: 72.553, Tokens/s/device: 10935.906, 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