2026-04-16 12:36:45.480986: 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 12:36:45.599783 129860943047808 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0416 12:37:41.820296 129860943047808 max_utils.py:800] System Information: Jax Version: 0.9.2 I0416 12:37:41.820399 129860943047808 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0416 12:37:41.820433 129860943047808 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 12:37:41.820458 129860943047808 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0416 12:37:41.820539 129860943047808 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0416 12:37:41.909956 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 12:37:42.314064 129860943047808 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0416 12:37:42.314319 129860943047808 checkpointing.py:677] Setting up checkpoint logger... I0416 12:37:42.314363 129860943047808 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0416 12:37:42.314402 129860943047808 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0416 12:37:42.314838 129860943047808 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 0x761b047d30b0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0416 12:37:45.424125 129860943047808 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0416 12:37:45.424296 129860943047808 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 0x76148973c0e0>}, handler_registry=None I0416 12:37:45.424556 129860943047808 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76148973c0e0>` 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 12:37:45.424595 129860943047808 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x761488972630>` 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 12:37:45.424626 129860943047808 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 0x76148973c0e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76148973c0e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x761488972630>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x761488972630>}). I0416 12:37:45.425020 129860943047808 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0416 12:37:45.425088 129860943047808 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7614881419e0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0416 12:37:45.566302 129860943047808 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints I0416 12:37:45.566527 129860943047808 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_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7614881493d0> I0416 12:37:45.566604 129860943047808 checkpointing.py:301] Checkpoint manager created! I0416 12:37:45.647054 129860943047808 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0416 12:37:45.649962 129860943047808 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. I0416 12:37:45.710526 129860943047808 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 I0416 12:37:45.744823 129860943047808 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0416 12:37:45.744896 129860943047808 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0416 12:37:46.354231 129860943047808 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0416 12:37:46.354342 129860943047808 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 12:37:46.458930 129860943047808 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 12:37:46.459008 129860943047808 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 12:37:46.473887 129860943047808 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0416 12:37:46.473940 129860943047808 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0416 12:37:46.495602 129860943047808 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 12:37:46.495659 129860943047808 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 12:37:46.510562 129860943047808 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 12:37:46.510612 129860943047808 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 12:37:46.525513 129860943047808 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 12:37:46.525563 129860943047808 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 12:37:46.547887 129860943047808 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0416 12:37:46.547944 129860943047808 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0416 12:37:46.567089 129860943047808 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0416 12:37:46.567141 129860943047808 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0416 12:37:47.133857 129860943047808 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0416 12:37:47.133951 129860943047808 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0416 12:37:47.184120 3819373 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 12:37:47.574764 129860943047808 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0416 12:37:47.574877 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0416 12:37:47.574920 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0416 12:37:47.574964 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0416 12:37:47.575003 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0416 12:37:47.575031 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0416 12:37:47.575086 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0416 12:37:47.575128 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0416 12:37:47.575157 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0416 12:37:47.575183 129860943047808 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0416 12:37:47.575219 129860943047808 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0416 12:37:47.575253 129860943047808 maxtext_utils.py:1620] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0416 12:37:48.011010 129860943047808 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 12:37:48.011103 129860943047808 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0416 12:37:48.025480 129860943047808 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0416 12:37:48.025530 129860943047808 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). W0416 12:37:48.295538 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 12:37:48.411196 129860943047808 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 12:37:48.411659 129860943047808 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0416 12:37:48.411792 129860943047808 metric_logger.py:289] number parameters: 1.104 billion W0416 12:37:51.456501 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 12:37:51.587808 129860943047808 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0416 12:37:51.707984 129860943047808 checkpointing.py:776] Waited 0.12016534805297852 seconds for step 0 to finish before starting checkpointing. I0416 12:37:51.708406 129860943047808 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 12:37:51.708557 129860943047808 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0416 12:37:51.708908 129860943047808 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0. I0416 12:37:51.808007 129860943047808 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0416 12:37:51.904230 129747763856960 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0 I0416 12:37:51.920809 129860943047808 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 12:37:51.920896 129860943047808 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 W0416 12:37:52.134619 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.143071 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.155465 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.161942 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.172706 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.181991 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.186959 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.191299 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 12:37:52.451598 129860943047808 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.531197s I0416 12:37:52.452027 129860943047808 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 19.200 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 642 milliseconds) (per-host) I0416 12:37:52.452091 129860943047808 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.642840s (batch_requests_ready=0.107173s, total_serialization_initiated=0.535329s, others=0.000337s) I0416 12:37:52.452150 129860943047808 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.643688s (all_items=0.000021s, per_item={'items': '0.00002098'}, temp_paths=0.643667) I0416 12:37:52.452993 129747692553792 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 12:37:52.453103 129860943047808 async_checkpointer.py:561] Finished blocking save. Time taken: 0.744508s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0. I0416 12:37:52.453276 129860943047808 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0416 12:37:52.453474 129747774342720 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 12:37:52.453581 129860943047808 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776343071.7083938, 'wait_for_prev_duration_secs': 3.814697265625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776343071.708577, 'checkpointer_blocking_duration_secs': 0.7446291446685791, 'get_old_steps_start_time': 1776343072.4532235, 'get_old_steps_duration_secs': 2.002716064453125e-05, 'checkpoint_manager_blocking_start_time': 1776343071.708313, 'checkpoint_manager_blocking_duration_secs': 0.7452394962310791} I0416 12:37:52.453693 129860943047808 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0416 12:37:52.453732 129860943047808 max_utils.py:750] Memstats: After params initialized: I0416 12:37:52.453804 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0416 12:37:52.453830 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0416 12:37:52.453850 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0416 12:37:52.453868 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0416 12:37:52.453886 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0416 12:37:52.453902 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0416 12:37:52.453918 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0416 12:37:52.453934 129860943047808 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0416 12:37:52.459771 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.465168 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0416 12:37:52.470289 3819373 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0416 12:37:52.765010 129747753371200 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/items I0416 12:37:52.773498 129860943047808 metric_logger.py:185] completed step: 0, seconds: 3.176, TFLOP/s/device: 4.279, Tokens/s/device: 644.931, total_weights: 13328, loss: 10.828 I0416 12:37:52.774358 129860943047808 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/tensorboard/' I0416 12:37:52.806566 129747673679424 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776343072554906931, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776343072554906931, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/_CHECKPOINT_METADATA I0416 12:37:52.888813 129860943047808 metric_logger.py:185] completed step: 1, seconds: 1.184, TFLOP/s/device: 11.478, Tokens/s/device: 1730.123, total_weights: 12332, loss: 10.828 I0416 12:37:53.006230 129860943047808 metric_logger.py:185] completed step: 2, seconds: 0.009, TFLOP/s/device: 1498.195, Tokens/s/device: 225824.236, total_weights: 15161, loss: 9.823 I0416 12:37:53.124101 129860943047808 metric_logger.py:185] completed step: 3, seconds: 0.116, TFLOP/s/device: 116.809, Tokens/s/device: 17606.754, total_weights: 13327, loss: 9.405 I0416 12:37:53.278503 3822025 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0416 12:37:53.278556 3822025 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0416 12:37:54.277530 129747713525312 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/items/array_metadatas/process_0 I0416 12:38:13.160544 129860943047808 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 115.411, Tokens/s/device: 17396.032, total_weights: 11939, loss: 9.045 I0416 12:38:13.169827 129860943047808 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 115.462, Tokens/s/device: 17403.719, total_weights: 15502, loss: 8.953 I0416 12:38:13.286611 129860943047808 metric_logger.py:185] completed step: 6, seconds: 20.036, TFLOP/s/device: 0.678, Tokens/s/device: 102.214, total_weights: 13864, loss: 8.775 I0416 12:38:13.404253 129860943047808 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 2274.377, Tokens/s/device: 342818.882, total_weights: 12988, loss: 8.688 I0416 12:38:13.521853 129860943047808 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 115.723, Tokens/s/device: 17442.999, total_weights: 13820, loss: 8.740 I0416 12:38:13.638910 129860943047808 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0416 12:38:13.639642 129860943047808 checkpointing.py:776] Waited 0.0007596015930175781 seconds for step 9 to finish before starting checkpointing. I0416 12:38:13.639915 129860943047808 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 12:38:24.835596 129747703039552 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 31.554235s (commit=31.046443s, array_metadata_write=0.507793s) I0416 12:38:24.836684 129747692553792 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 382.616 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 33 seconds) (per-host) I0416 12:38:24.836797 129747692553792 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.383647s. I0416 12:38:25.096494 129747692553792 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776343072554906931, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/_CHECKPOINT_METADATA I0416 12:38:25.277503 129747692553792 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 12:38:25.497770 129747673679424 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776343072554906931, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/_CHECKPOINT_METADATA I0416 12:38:25.747606 129747692553792 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 12:38:25.748227 129747692553792 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.614946s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/items I0416 12:38:25.748849 129747692553792 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0/items I0416 12:38:26.005027 129747692553792 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0 I0416 12:38:26.748026 129747692553792 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0`. I0416 12:38:26.748638 129747692553792 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 35.040049s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/0 I0416 12:38:26.748716 129747692553792 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 34.295568s. I0416 12:38:26.748873 129747774342720 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 12:38:26.748974 129747774342720 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 12:38:26.749033 129747774342720 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0416 12:38:26.749089 129747774342720 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0416 12:38:26.749170 129860943047808 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 12:38:26.749239 129860943047808 checkpoint_manager.py:1441] Waiting for previous save to complete took 13.109345 seconds. If this number is high, consider checkpointing less frequently. I0416 12:38:26.750010 129860943047808 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0416 12:38:26.750290 129860943047808 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9. I0416 12:38:26.934701 129747774342720 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9 I0416 12:38:26.950309 129860943047808 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0416 12:38:26.950404 129860943047808 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 12:38:27.688798 129747537364544 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9/items I0416 12:38:35.881627 129860943047808 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 8.931721s I0416 12:38:35.886175 129860943047808 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.365 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 9 seconds) (per-host) I0416 12:38:35.886253 129860943047808 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 9.041654s (batch_requests_ready=0.102441s, total_serialization_initiated=8.934829s, others=0.004384s) I0416 12:38:35.886348 129860943047808 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 9.042175s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=9.042158) I0416 12:38:35.887890 129747713525312 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0416 12:38:35.888013 129860943047808 async_checkpointer.py:561] Finished blocking save. Time taken: 9.137953s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9. I0416 12:38:35.888403 129860943047808 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0416 12:38:35.888665 129747692553792 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0416 12:38:35.888763 129860943047808 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776343093.6398718, 'wait_for_prev_duration_secs': 13.109345197677612, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776343106.7500334, 'checkpointer_blocking_duration_secs': 9.138143301010132, 'get_old_steps_start_time': 1776343115.8882046, 'get_old_steps_duration_secs': 0.00011277198791503906, 'checkpoint_manager_blocking_start_time': 1776343093.639827, 'checkpoint_manager_blocking_duration_secs': 22.248902559280396} I0416 12:38:35.888874 129860943047808 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0416 12:38:35.888911 129860943047808 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0416 12:38:36.691269 129747742885440 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9/items/array_metadatas/process_0 I0416 12:39:13.957713 129747537364544 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.070483s (commit=37.581737s, array_metadata_write=0.488746s) I0416 12:39:13.958914 129747713525312 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 268.216 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 47 seconds) (per-host) I0416 12:39:13.959028 129747713525312 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.070933s. I0416 12:39:14.402483 129747713525312 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 12:39:14.859874 129747713525312 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0416 12:39:14.860143 129747713525312 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.613648s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9/items I0416 12:39:14.860774 129747713525312 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9/items I0416 12:39:15.109898 129747713525312 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9 I0416 12:39:15.816253 129747713525312 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9`. I0416 12:39:15.816886 129747713525312 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 49.066839s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_01_base/checkpoints/9 I0416 12:39:15.816974 129747713525312 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 39.928881s. I0416 12:39:15.817141 129747692553792 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0416 12:39:15.817192 129747692553792 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0416 12:39:15.817255 129747692553792 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0416 12:39:15.817293 129747692553792 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0416 12:39:15.817642 129860943047808 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 12:39:15.818402 129860943047808 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0416 12:39:15.819079 129860943047808 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 115.480, Tokens/s/device: 17406.381, total_weights: 12300, loss: 8.710 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops