2026-04-17 12:52:47.379738: 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) I0417 12:52:47.497828 135148024736896 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0417 12:53:28.820539 135148024736896 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 12:53:28.820654 135148024736896 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 12:53:28.820697 135148024736896 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 I0417 12:53:28.820723 135148024736896 train_utils.py:335] WARNING: 'dataset_path' might be pointing your local file system I0417 12:53:28.820836 135148024736896 train.py:671] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0417 12:53:28.917648 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 12:53:29.298371 135148024736896 maxtext_utils.py:1548] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 12:53:29.298722 135148024736896 checkpointing.py:677] Setting up checkpoint logger... I0417 12:53:29.298769 135148024736896 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 12:53:29.298848 135148024736896 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0417 12:53:29.299608 135148024736896 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 0x7aea031fc920>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 12:53:32.328489 135148024736896 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0417 12:53:32.328715 135148024736896 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 0x7ae386f557f0>}, handler_registry=None I0417 12:53:32.329105 135148024736896 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae386f557f0>` 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`. I0417 12:53:32.329151 135148024736896 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae386f565d0>` 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`. I0417 12:53:32.329181 135148024736896 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 0x7ae386f557f0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ae386f557f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae386f565d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ae386f565d0>}). I0417 12:53:32.329516 135148024736896 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0417 12:53:32.329572 135148024736896 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7ae386f4e840> timeout: 600 secs and primary_host=0 for async checkpoint writes I0417 12:53:32.455310 135148024736896 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints I0417 12:53:32.455549 135148024736896 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_20260417_125240/linen_main_20260417_125240_01_base/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ae386f56360> I0417 12:53:32.455644 135148024736896 checkpointing.py:301] Checkpoint manager created! I0417 12:53:32.909766 135148024736896 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0417 12:53:32.913320 135148024736896 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. I0417 12:53:32.966335 135148024736896 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 I0417 12:53:32.994033 135148024736896 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0417 12:53:32.994109 135148024736896 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0417 12:53:33.596351 135148024736896 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 12:53:33.596453 135148024736896 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 12:53:33.699599 135148024736896 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 12:53:33.699675 135148024736896 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 12:53:33.714420 135148024736896 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 12:53:33.714468 135148024736896 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 12:53:33.736490 135148024736896 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 12:53:33.736546 135148024736896 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 12:53:33.751369 135148024736896 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 12:53:33.751418 135148024736896 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 12:53:33.766202 135148024736896 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 12:53:33.766248 135148024736896 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 12:53:33.788358 135148024736896 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 12:53:33.788417 135148024736896 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 12:53:33.807370 135148024736896 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 12:53:33.807421 135148024736896 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0417 12:53:34.000763 135148024736896 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0417 12:53:34.000865 135148024736896 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0417 12:53:34.051812 404923 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 I0417 12:53:34.444119 135148024736896 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 12:53:34.444248 135148024736896 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 12:53:34.444290 135148024736896 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 12:53:34.444334 135148024736896 maxtext_utils.py:1651] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0417 12:53:34.444373 135148024736896 maxtext_utils.py:1651] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 12:53:34.444401 135148024736896 maxtext_utils.py:1651] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 12:53:34.444440 135148024736896 maxtext_utils.py:1651] 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) I0417 12:53:34.444478 135148024736896 maxtext_utils.py:1651] 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') I0417 12:53:34.444506 135148024736896 maxtext_utils.py:1651] 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) I0417 12:53:34.444531 135148024736896 maxtext_utils.py:1651] 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) I0417 12:53:34.444565 135148024736896 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 12:53:34.444605 135148024736896 maxtext_utils.py:1651] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 12:53:34.891773 135148024736896 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 12:53:34.891855 135148024736896 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0417 12:53:34.905978 135148024736896 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 12:53:34.906027 135148024736896 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0417 12:53:42.373578 135148024736896 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. I0417 12:53:42.374235 135148024736896 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0417 12:53:42.374407 135148024736896 metric_logger.py:301] number parameters: 1.104 billion I0417 12:53:53.362326 135148024736896 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0417 12:53:53.492028 135148024736896 checkpointing.py:776] Waited 0.12969017028808594 seconds for step 0 to finish before starting checkpointing. I0417 12:53:53.492554 135148024736896 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 12:53:53.492753 135148024736896 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0417 12:53:53.493139 135148024736896 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0. I0417 12:53:53.598906 135148024736896 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0417 12:53:53.691686 135033924879936 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0 I0417 12:53:53.699010 135148024736896 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 12:53:53.699110 135148024736896 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 W0417 12:53:53.991999 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.002288 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.021560 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.033118 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.039812 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.046029 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.050912 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.055652 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 12:53:54.457637 135033914394176 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/items I0417 12:53:54.548683 135148024736896 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.850047s I0417 12:53:54.549211 135148024736896 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 13.005 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 948 milliseconds) (per-host) I0417 12:53:54.549289 135148024736896 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.949130s (batch_requests_ready=0.094840s, total_serialization_initiated=0.853906s, others=0.000384s) I0417 12:53:54.549368 135148024736896 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.949989s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.949969) I0417 12:53:54.550426 135033861965376 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 12:53:54.550547 135148024736896 async_checkpointer.py:561] Finished blocking save. Time taken: 1.057748s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0. I0417 12:53:54.550796 135148024736896 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 12:53:54.551065 135148024736896 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776430433.4925416, 'wait_for_prev_duration_secs': 4.124641418457031e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776430433.4927752, 'checkpointer_blocking_duration_secs': 1.0579135417938232, 'get_old_steps_start_time': 1776430434.550717, 'get_old_steps_duration_secs': 2.7418136596679688e-05, 'checkpoint_manager_blocking_start_time': 1776430433.492454, 'checkpoint_manager_blocking_duration_secs': 1.0585684776306152} I0417 12:53:54.551161 135148024736896 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0417 12:53:54.551206 135148024736896 max_utils.py:750] Memstats: After params initialized: I0417 12:53:54.551331 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0417 12:53:54.551361 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0417 12:53:54.551382 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0417 12:53:54.551400 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0417 12:53:54.551419 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0417 12:53:54.551437 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0417 12:53:54.551454 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0417 12:53:54.551471 135148024736896 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) W0417 12:53:54.560273 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.567248 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0417 12:53:54.572688 404923 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0417 12:53:54.593163 135033882936896 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776430434360364968, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776430434360364968, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/_CHECKPOINT_METADATA I0417 12:53:54.593379 135033935365696 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 12:53:54.917940 135148024736896 metric_logger.py:196] completed step: 0, seconds: 10.987, TFLOP/s/device: 1.237, Tokens/s/device: 186.403, total_weights: 13328, loss: 10.828, lm_loss: 10.828, perplexity: 50404.086 I0417 12:53:54.919059 135148024736896 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/tensorboard/' I0417 12:53:55.035220 408136 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0417 12:53:55.035282 408136 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0417 12:53:55.038705 135148024736896 metric_logger.py:196] completed step: 1, seconds: 1.553, TFLOP/s/device: 8.750, Tokens/s/device: 1318.892, total_weights: 12332, loss: 10.828, lm_loss: 10.828, perplexity: 50391.758 I0417 12:53:55.150868 135148024736896 metric_logger.py:196] completed step: 2, seconds: 0.011, TFLOP/s/device: 1237.556, Tokens/s/device: 186537.936, total_weights: 15161, loss: 9.823, lm_loss: 9.823, perplexity: 18444.867 I0417 12:53:55.268531 135148024736896 metric_logger.py:196] completed step: 3, seconds: 0.144, TFLOP/s/device: 94.627, Tokens/s/device: 14263.229, total_weights: 13327, loss: 9.405, lm_loss: 9.405, perplexity: 12149.331 I0417 12:53:56.017032 135033903908416 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/items/array_metadatas/process_0 I0417 12:54:15.344687 135148024736896 metric_logger.py:196] completed step: 4, seconds: 0.090, TFLOP/s/device: 150.212, Tokens/s/device: 22641.593, total_weights: 11939, loss: 9.045, lm_loss: 9.045, perplexity: 8471.842 I0417 12:54:15.353892 135148024736896 metric_logger.py:196] completed step: 5, seconds: 0.117, TFLOP/s/device: 115.805, Tokens/s/device: 17455.339, total_weights: 15502, loss: 8.953, lm_loss: 8.953, perplexity: 7728.546 I0417 12:54:15.470127 135148024736896 metric_logger.py:196] completed step: 6, seconds: 20.075, TFLOP/s/device: 0.677, Tokens/s/device: 102.017, total_weights: 13864, loss: 8.775, lm_loss: 8.775, perplexity: 6468.536 I0417 12:54:15.587408 135148024736896 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2074.371, Tokens/s/device: 312671.756, total_weights: 12988, loss: 8.688, lm_loss: 8.688, perplexity: 5928.887 I0417 12:54:15.704940 135148024736896 metric_logger.py:196] completed step: 8, seconds: 0.117, TFLOP/s/device: 115.979, Tokens/s/device: 17481.562, total_weights: 13820, loss: 8.740, lm_loss: 8.740, perplexity: 6250.123 I0417 12:54:15.822037 135148024736896 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0417 12:54:15.822871 135148024736896 checkpointing.py:776] Waited 0.0008540153503417969 seconds for step 9 to finish before starting checkpointing. I0417 12:54:15.823171 135148024736896 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 12:54:24.552275 135033924879936 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 29.514368s (commit=29.085639s, array_metadata_write=0.428729s) I0417 12:54:24.553423 135033861965376 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 408.255 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 30 seconds) (per-host) I0417 12:54:24.553529 135033861965376 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 30.002923s. I0417 12:54:24.782825 135033861965376 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776430434360364968, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/_CHECKPOINT_METADATA I0417 12:54:24.975895 135033861965376 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0417 12:54:25.171748 135033882936896 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776430434360364968, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/_CHECKPOINT_METADATA I0417 12:54:25.451370 135033861965376 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0417 12:54:25.451987 135033861965376 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.627427s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/items I0417 12:54:25.452652 135033861965376 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0/items I0417 12:54:25.694494 135033861965376 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0 I0417 12:54:26.388325 135033861965376 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0`. I0417 12:54:26.388952 135033861965376 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 32.896165s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/0 I0417 12:54:26.389028 135033861965376 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 31.838423s. I0417 12:54:26.389219 135033935365696 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 12:54:26.389325 135033935365696 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 12:54:26.389384 135033935365696 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 12:54:26.389430 135033935365696 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 12:54:26.389519 135148024736896 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. W0417 12:54:26.389591 135148024736896 checkpoint_manager.py:1441] Waiting for previous save to complete took 10.566448 seconds. If this number is high, consider checkpointing less frequently. I0417 12:54:26.390292 135148024736896 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0417 12:54:26.390573 135148024736896 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9. I0417 12:54:26.567563 135148024736896 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 12:54:26.567665 135148024736896 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 I0417 12:54:26.569211 135033935365696 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9 I0417 12:54:27.267403 135033719359040 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9/items I0417 12:54:32.535483 135148024736896 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 5.968334s I0417 12:54:32.540138 135148024736896 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 2.037 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 6 seconds) (per-host) I0417 12:54:32.540206 135148024736896 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 6.059712s (batch_requests_ready=0.083730s, total_serialization_initiated=5.971423s, others=0.004560s) I0417 12:54:32.540278 135148024736896 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 6.060606s (all_items=0.000023s, per_item={'items': '0.00002265'}, temp_paths=6.060583) I0417 12:54:32.541827 135033851479616 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 12:54:32.541916 135148024736896 async_checkpointer.py:561] Finished blocking save. Time taken: 6.151581s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9. I0417 12:54:32.542118 135148024736896 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 12:54:32.542279 135033861965376 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 12:54:32.542368 135148024736896 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776430455.82312, 'wait_for_prev_duration_secs': 10.566447973251343, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776430466.390316, 'checkpointer_blocking_duration_secs': 6.15170431137085, 'get_old_steps_start_time': 1776430472.5420382, 'get_old_steps_duration_secs': 4.1961669921875e-05, 'checkpoint_manager_blocking_start_time': 1776430455.823076, 'checkpoint_manager_blocking_duration_secs': 16.719266653060913} I0417 12:54:32.542490 135148024736896 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0417 12:54:32.542522 135148024736896 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 12:54:33.272892 135033903908416 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9/items/array_metadatas/process_0 I0417 12:55:14.213238 135033719359040 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.672256s (commit=41.195749s, array_metadata_write=0.476506s) I0417 12:55:14.214257 135033851479616 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 264.735 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 47 seconds) (per-host) I0417 12:55:14.214333 135033851479616 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.672369s. I0417 12:55:14.711958 135033851479616 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0417 12:55:15.159133 135033851479616 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0417 12:55:15.159894 135033851479616 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.625245s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9/items I0417 12:55:15.160545 135033851479616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9/items I0417 12:55:15.419573 135033851479616 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9 I0417 12:55:16.112610 135033851479616 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9`. I0417 12:55:16.113259 135033851479616 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 49.722930s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_01_base/checkpoints/9 I0417 12:55:16.113336 135033851479616 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 43.571376s. I0417 12:55:16.113439 135033861965376 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 12:55:16.113488 135033861965376 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 12:55:16.113545 135033861965376 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 12:55:16.113591 135033861965376 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 12:55:16.113714 135148024736896 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. I0417 12:55:16.113837 135148024736896 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 12:55:16.114632 135148024736896 metric_logger.py:196] completed step: 9, seconds: 0.117, TFLOP/s/device: 115.917, Tokens/s/device: 17472.316, total_weights: 12300, loss: 8.710, lm_loss: 8.710, perplexity: 6065.008 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops