2026-04-15 23:13:15.274763: 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 23:13:15.390276 136505246944384 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 23:14:03.084233 136505246944384 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 23:14:03.084359 136505246944384 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 23:14:03.084395 136505246944384 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 23:14:03.084422 136505246944384 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 23:14:03.084528 136505246944384 train.py:715] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 23:14:03.178160 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:14:03.587754 136505246944384 maxtext_utils.py:1520] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 23:14:03.588083 136505246944384 checkpointing.py:677] Setting up checkpoint logger... I0415 23:14:03.588134 136505246944384 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0415 23:14:03.588184 136505246944384 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 23:14:03.588838 136505246944384 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 0x7c26047fcda0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 23:14:06.039266 136505246944384 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 23:14:06.039482 136505246944384 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 0x7c1f88dd7c50>}, handler_registry=None I0415 23:14:06.039773 136505246944384 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c1f88dd7c50>` 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 23:14:06.039814 136505246944384 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c1f87f3e3f0>` 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 23:14:06.039844 136505246944384 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 0x7c1f88dd7c50>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c1f88dd7c50>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c1f87f3e3f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c1f87f3e3f0>}). I0415 23:14:06.040215 136505246944384 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 23:14:06.040271 136505246944384 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7c1f87f394e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 23:14:06.181328 136505246944384 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_03_dropout/checkpoints I0415 23:14:06.181576 136505246944384 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_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c1f87f3dfd0> I0415 23:14:06.181659 136505246944384 checkpointing.py:301] Checkpoint manager created! I0415 23:14:06.260955 136505246944384 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0415 23:14:06.264061 136505246944384 reader.py:262] Creating a tf.data.Dataset reading 8 files located in folders: tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0. I0415 23:14:06.322398 136505246944384 logging_logger.py:49] Constructing tf.data.Dataset __local_c4_builder for split train, from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0415 23:14:06.352427 136505246944384 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0415 23:14:06.352498 136505246944384 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0415 23:14:07.371445 136505246944384 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 23:14:07.371560 136505246944384 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None). I0415 23:14:07.475809 136505246944384 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 23:14:07.475896 136505246944384 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None). I0415 23:14:07.491271 136505246944384 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 23:14:07.491322 136505246944384 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None). I0415 23:14:07.513236 136505246944384 attentions.py:1154] attentions/query Logical: bfloat16[8,128,16,128]...................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 23:14:07.513295 136505246944384 attentions.py:1154] attentions/query Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 23:14:07.528649 136505246944384 attentions.py:1155] attentions/key Logical: bfloat16[8,128,16,128]...................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 23:14:07.528703 136505246944384 attentions.py:1155] attentions/key Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 23:14:07.544301 136505246944384 attentions.py:1156] attentions/value Logical: bfloat16[8,128,16,128]...................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0415 23:14:07.544353 136505246944384 attentions.py:1156] attentions/value Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 23:14:07.567068 136505246944384 attentions.py:1197] attentions/out Logical: bfloat16[8,128,16,128]...................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 23:14:07.567126 136505246944384 attentions.py:1197] attentions/out Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 23:14:07.591139 136505246944384 linears.py:525] linears/x Logical: bfloat16[8,128,7168]........................................ ('activation_batch', 'activation_length', 'activation_mlp'). I0415 23:14:07.591194 136505246944384 linears.py:525] linears/x Physical: bfloat16[8,128,7168]........................................ ('fsdp', None, None). I0415 23:14:07.789263 136505246944384 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 23:14:07.789360 136505246944384 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 23:14:07.840388 3354671 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 23:14:08.256075 136505246944384 maxtext_utils.py:1623] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 23:14:08.256191 136505246944384 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 23:14:08.256233 136505246944384 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 23:14:08.256278 136505246944384 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 23:14:08.256318 136505246944384 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 23:14:08.256347 136505246944384 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 23:14:08.256386 136505246944384 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 23:14:08.256425 136505246944384 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 23:14:08.256454 136505246944384 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 23:14:08.256479 136505246944384 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 23:14:08.256514 136505246944384 maxtext_utils.py:1623] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed', 'vocab') Physical: ('fsdp', None) I0415 23:14:08.256547 136505246944384 maxtext_utils.py:1623] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed') Physical: (None, 'fsdp') I0415 23:14:08.728078 136505246944384 train.py:155] train/xent Logical: float32[8,128].............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 23:14:08.728167 136505246944384 train.py:155] train/xent Physical: float32[8,128].............................................. ('fsdp', None). I0415 23:14:08.742895 136505246944384 train.py:162] train/z_loss Logical: float32[8,128].............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 23:14:08.742949 136505246944384 train.py:162] train/z_loss Physical: float32[8,128].............................................. ('fsdp', None). W0415 23:14:09.014458 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:14:09.097641 136505246944384 max_utils.py:791] Total memory size: 2.5 GB, Output size: 1.5 GB, Temp size: 1.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB. I0415 23:14:09.098238 136505246944384 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 23:14:09.098382 136505246944384 metric_logger.py:289] number parameters: 1.104 billion W0415 23:14:12.186669 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:14:12.291882 136505246944384 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 23:14:12.357609 136505246944384 checkpointing.py:776] Waited 0.06570744514465332 seconds for step 0 to finish before starting checkpointing. I0415 23:14:12.358078 136505246944384 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 23:14:12.358231 136505246944384 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 23:14:12.358619 136505246944384 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_03_dropout/checkpoints/0. I0415 23:14:12.446909 136505246944384 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 23:14:12.540102 136391956629056 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/0 I0415 23:14:12.560216 136505246944384 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 23:14:12.560320 136505246944384 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 23:14:12.789211 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.802541 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.811644 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.829807 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.840545 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.847234 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.852125 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:12.856785 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:14:13.238432 136391946143296 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/0/items I0415 23:14:13.421397 136391906297408 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776294853156950908, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776294853156950908, "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_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0415 23:14:13.725519 3357352 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 23:14:13.725577 3357352 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 23:14:14.214401 136505246944384 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.654699s I0415 23:14:14.219886 136505246944384 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 6.966 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0415 23:14:14.219955 136505246944384 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.771601s (batch_requests_ready=0.107887s, total_serialization_initiated=1.658377s, others=0.005336s) I0415 23:14:14.220036 136505246944384 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.772557s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=1.772537) I0415 23:14:14.221628 136391967114816 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 23:14:14.221793 136505246944384 async_checkpointer.py:561] Finished blocking save. Time taken: 1.863517s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/0. I0415 23:14:14.222055 136505246944384 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 23:14:14.222241 136391832897088 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 23:14:14.222326 136505246944384 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_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776294852.358065, 'wait_for_prev_duration_secs': 4.1961669921875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776294852.3582513, 'checkpointer_blocking_duration_secs': 1.863691806793213, 'get_old_steps_start_time': 1776294854.2219696, 'get_old_steps_duration_secs': 2.8133392333984375e-05, 'checkpoint_manager_blocking_start_time': 1776294852.3579605, 'checkpoint_manager_blocking_duration_secs': 1.8643333911895752} I0415 23:14:14.222469 136505246944384 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 23:14:14.222518 136505246944384 max_utils.py:750] Memstats: After params initialized: I0415 23:14:14.222567 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_0(process=0,(0,0,0,0)) I0415 23:14:14.222593 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_1(process=0,(1,0,0,0)) I0415 23:14:14.222614 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_2(process=0,(0,1,0,0)) I0415 23:14:14.222633 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_3(process=0,(1,1,0,0)) I0415 23:14:14.222650 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_4(process=0,(0,2,0,0)) I0415 23:14:14.222667 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_5(process=0,(1,2,0,0)) I0415 23:14:14.222684 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_6(process=0,(0,3,0,0)) I0415 23:14:14.222701 136505246944384 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_7(process=0,(1,3,0,0)) W0415 23:14:14.229369 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:14.236322 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 23:14:14.241170 3354671 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 23:14:14.578479 136505246944384 metric_logger.py:185] completed step: 0, seconds: 3.193, TFLOP/s/device: 0.251, Tokens/s/device: 40.088, total_weights: 963, loss: 10.834 I0415 23:14:14.579471 136505246944384 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_03_dropout/tensorboard/' I0415 23:14:14.639873 136505246944384 metric_logger.py:185] completed step: 1, seconds: 2.285, TFLOP/s/device: 0.351, Tokens/s/device: 56.028, total_weights: 945, loss: 10.747 I0415 23:14:14.704004 136505246944384 metric_logger.py:185] completed step: 2, seconds: 0.009, TFLOP/s/device: 84.992, Tokens/s/device: 13583.784, total_weights: 962, loss: 10.246 I0415 23:14:14.767799 136505246944384 metric_logger.py:185] completed step: 3, seconds: 0.062, TFLOP/s/device: 12.891, Tokens/s/device: 2060.362, total_weights: 975, loss: 9.829 I0415 23:14:15.251475 136391925171776 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_03_dropout/checkpoints/0/items/array_metadatas/process_0 I0415 23:14:25.853652 136505246944384 metric_logger.py:185] completed step: 4, seconds: 0.065, TFLOP/s/device: 12.295, Tokens/s/device: 1965.089, total_weights: 962, loss: 9.393 I0415 23:14:25.861922 136505246944384 metric_logger.py:185] completed step: 5, seconds: 0.064, TFLOP/s/device: 12.593, Tokens/s/device: 2012.642, total_weights: 958, loss: 9.248 I0415 23:14:25.924601 136505246944384 metric_logger.py:185] completed step: 6, seconds: 11.084, TFLOP/s/device: 0.072, Tokens/s/device: 11.548, total_weights: 930, loss: 9.424 I0415 23:14:25.988297 136505246944384 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 125.608, Tokens/s/device: 20075.282, total_weights: 935, loss: 9.317 I0415 23:14:26.052064 136505246944384 metric_logger.py:185] completed step: 8, seconds: 0.064, TFLOP/s/device: 12.511, Tokens/s/device: 1999.563, total_weights: 930, loss: 9.478 I0415 23:14:26.115510 136505246944384 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 23:14:26.116351 136505246944384 checkpointing.py:776] Waited 0.0008597373962402344 seconds for step 9 to finish before starting checkpointing. I0415 23:14:26.116629 136505246944384 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 23:14:46.396951 136391853868608 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 32.176193s (commit=31.731903s, array_metadata_write=0.444290s) I0415 23:14:46.398282 136391967114816 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 372.219 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 33 seconds) (per-host) I0415 23:14:46.398407 136391967114816 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.176553s. I0415 23:14:46.682684 136391967114816 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776294853156950908, '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_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0415 23:14:46.863582 136391967114816 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 23:14:47.057176 136391906297408 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776294853156950908, '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_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0415 23:14:47.280364 136391967114816 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 23:14:47.280563 136391967114816 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.558657s. 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_03_dropout/checkpoints/0/items I0415 23:14:47.281329 136391967114816 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/0/items I0415 23:14:47.524904 136391967114816 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/0 I0415 23:14:48.171506 136391967114816 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_03_dropout/checkpoints/0`. I0415 23:14:48.172311 136391967114816 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 35.814048s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/0 I0415 23:14:48.172391 136391967114816 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 33.950541s. I0415 23:14:48.172510 136391832897088 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 23:14:48.172560 136391832897088 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 23:14:48.172616 136391832897088 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 23:14:48.172664 136391832897088 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 23:14:48.172822 136505246944384 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 23:14:48.172985 136505246944384 checkpoint_manager.py:1441] Waiting for previous save to complete took 22.056354 seconds. If this number is high, consider checkpointing less frequently. I0415 23:14:48.173883 136505246944384 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 23:14:48.174228 136505246944384 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_03_dropout/checkpoints/9. I0415 23:14:48.354027 136505246944384 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 23:14:48.354146 136505246944384 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 23:14:48.355216 136391832897088 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/9 I0415 23:14:49.058679 136391822411328 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/9/items I0415 23:14:56.847333 136505246944384 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 8.494185s I0415 23:14:56.848111 136505246944384 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.437 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host) I0415 23:14:56.848176 136505246944384 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 8.589571s (batch_requests_ready=0.091559s, total_serialization_initiated=8.497419s, others=0.000592s) I0415 23:14:56.848251 136505246944384 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 8.590131s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=8.590112) I0415 23:14:56.854337 136392042612288 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 23:14:56.854470 136505246944384 async_checkpointer.py:561] Finished blocking save. Time taken: 8.680536s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/9. I0415 23:14:56.854709 136505246944384 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 23:14:56.854908 136391967114816 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 23:14:56.855025 136505246944384 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_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776294866.1165953, 'wait_for_prev_duration_secs': 22.05635380744934, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776294888.1739082, 'checkpointer_blocking_duration_secs': 8.680684566497803, 'get_old_steps_start_time': 1776294896.8546288, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776294866.1165524, 'checkpoint_manager_blocking_duration_secs': 30.738441705703735} I0415 23:14:56.855185 136505246944384 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 23:14:56.855223 136505246944384 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 23:14:57.626397 136403499353664 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_03_dropout/checkpoints/9/items/array_metadatas/process_0 I0415 23:15:32.104003 136391822411328 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.250379s (commit=34.721133s, array_metadata_write=0.529245s) I0415 23:15:32.105496 136392042612288 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 288.203 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 43 seconds) (per-host) I0415 23:15:32.105617 136392042612288 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.251098s. I0415 23:15:32.543421 136392042612288 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 23:15:32.965383 136392042612288 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 23:15:32.965585 136392042612288 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.571543s. 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_03_dropout/checkpoints/9/items I0415 23:15:32.966344 136392042612288 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/9/items I0415 23:15:33.195058 136392042612288 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/9 I0415 23:15:33.831426 136392042612288 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_03_dropout/checkpoints/9`. I0415 23:15:33.832238 136392042612288 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 45.658316s. directory=gs://wanglance-maxtext/linen_ckpt_feat_migrate_nnx_utils_20260415_230805/linen_feat_migrate_nnx_utils_20260415_230805_03_dropout/checkpoints/9 I0415 23:15:33.832324 136392042612288 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.977807s. I0415 23:15:33.832565 136391967114816 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 23:15:33.832702 136391967114816 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 23:15:33.832774 136391967114816 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 23:15:33.832819 136391967114816 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 23:15:33.833264 136505246944384 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 23:15:33.833435 136505246944384 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 23:15:33.834405 136505246944384 metric_logger.py:185] completed step: 9, seconds: 0.063, TFLOP/s/device: 12.694, Tokens/s/device: 2028.880, total_weights: 948, loss: 9.141 Per train step: Total TFLOPs: 0.80 split as 99.60% learnable weight flops and 0.40% attention flops