2026-04-15 21:16:35.366502: 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 21:16:35.484161 130761292586112 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0415 21:17:21.181296 130761292586112 max_utils.py:800] System Information: Jax Version: 0.9.2 I0415 21:17:21.181400 130761292586112 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0415 21:17:21.181434 130761292586112 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 21:17:21.181463 130761292586112 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system I0415 21:17:21.181545 130761292586112 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. W0415 21:17:21.272784 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 21:17:21.678175 130761292586112 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0415 21:17:21.678442 130761292586112 checkpointing.py:677] Setting up checkpoint logger... I0415 21:17:21.678483 130761292586112 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0415 21:17:21.678527 130761292586112 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0415 21:17:21.678977 130761292586112 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 0x76eca6b143e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0415 21:17:24.144196 130761292586112 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0415 21:17:24.144365 130761292586112 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 0x76e629939490>}, handler_registry=None I0415 21:17:24.144633 130761292586112 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76e629939490>` 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 21:17:24.144672 130761292586112 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76e62993a540>` 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 21:17:24.144702 130761292586112 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 0x76e629939490>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76e629939490>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76e62993a540>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76e62993a540>}). I0415 21:17:24.145160 130761292586112 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0415 21:17:24.145214 130761292586112 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x76e6299351c0> timeout: 600 secs and primary_host=0 for async checkpoint writes I0415 21:17:24.273647 130761292586112 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints I0415 21:17:24.273860 130761292586112 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_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x76e62993a030> I0415 21:17:24.273937 130761292586112 checkpointing.py:301] Checkpoint manager created! I0415 21:17:24.353997 130761292586112 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0415 21:17:24.356923 130761292586112 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 21:17:24.410016 130761292586112 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 21:17:24.439553 130761292586112 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0415 21:17:24.439626 130761292586112 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0415 21:17:25.433981 130761292586112 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_norm_length', 'activation_embed'). I0415 21:17:25.434102 130761292586112 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None). I0415 21:17:25.536309 130761292586112 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 21:17:25.536386 130761292586112 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None). I0415 21:17:25.551237 130761292586112 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0415 21:17:25.551286 130761292586112 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None). I0415 21:17:25.572707 130761292586112 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 21:17:25.572762 130761292586112 attentions.py:1154] attentions/query Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 21:17:25.587636 130761292586112 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 21:17:25.587686 130761292586112 attentions.py:1155] attentions/key Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 21:17:25.603382 130761292586112 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 21:17:25.603436 130761292586112 attentions.py:1156] attentions/value Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 21:17:25.625615 130761292586112 attentions.py:1197] attentions/out Logical: bfloat16[8,128,16,128]...................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0415 21:17:25.625675 130761292586112 attentions.py:1197] attentions/out Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None). I0415 21:17:25.649794 130761292586112 linears.py:525] linears/x Logical: bfloat16[8,128,7168]........................................ ('activation_batch', 'activation_length', 'activation_mlp'). I0415 21:17:25.649851 130761292586112 linears.py:525] linears/x Physical: bfloat16[8,128,7168]........................................ ('fsdp', None, None). I0415 21:17:25.844268 130761292586112 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0415 21:17:25.844361 130761292586112 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. W0415 21:17:25.894010 3230093 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 21:17:26.300641 130761292586112 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0415 21:17:26.300751 130761292586112 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) I0415 21:17:26.300794 130761292586112 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) I0415 21:17:26.300841 130761292586112 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') I0415 21:17:26.300894 130761292586112 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) I0415 21:17:26.300924 130761292586112 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) I0415 21:17:26.300964 130761292586112 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) I0415 21:17:26.301004 130761292586112 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') I0415 21:17:26.301033 130761292586112 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) I0415 21:17:26.301173 130761292586112 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) I0415 21:17:26.301218 130761292586112 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0415 21:17:26.301253 130761292586112 maxtext_utils.py:1620] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0415 21:17:26.760637 130761292586112 train.py:155] train/xent Logical: float32[8,128].............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 21:17:26.760717 130761292586112 train.py:155] train/xent Physical: float32[8,128].............................................. ('fsdp', None). I0415 21:17:26.775017 130761292586112 train.py:162] train/z_loss Logical: float32[8,128].............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0415 21:17:26.775076 130761292586112 train.py:162] train/z_loss Physical: float32[8,128].............................................. ('fsdp', None). W0415 21:17:27.013327 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 21:17:27.104184 130761292586112 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 21:17:27.104654 130761292586112 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0415 21:17:27.104795 130761292586112 metric_logger.py:289] number parameters: 1.104 billion W0415 21:17:30.127227 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 21:17:30.227376 130761292586112 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0415 21:17:30.292611 130761292586112 checkpointing.py:776] Waited 0.06522655487060547 seconds for step 0 to finish before starting checkpointing. I0415 21:17:30.293008 130761292586112 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 21:17:30.293172 130761292586112 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0415 21:17:30.293536 130761292586112 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0. I0415 21:17:30.390348 130761292586112 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0415 21:17:30.476290 130761292586112 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 21:17:30.476377 130761292586112 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 21:17:30.483838 130648073307712 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0 W0415 21:17:30.690288 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.700891 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.714345 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.731649 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.746025 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.752393 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.759092 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:30.763776 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 21:17:31.114117 130761292586112 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.638282s I0415 21:17:31.114543 130761292586112 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 17.064 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 723 milliseconds) (per-host) I0415 21:17:31.114597 130761292586112 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.723319s (batch_requests_ready=0.081218s, total_serialization_initiated=0.641772s, others=0.000329s) I0415 21:17:31.114658 130761292586112 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.723857s (all_items=0.000020s, per_item={'items': '0.00002027'}, temp_paths=0.723837) I0415 21:17:31.115429 130648002004544 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 21:17:31.115527 130761292586112 async_checkpointer.py:561] Finished blocking save. Time taken: 0.822314s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0. I0415 21:17:31.115698 130761292586112 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0415 21:17:31.115948 130647991518784 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 21:17:31.116070 130761292586112 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776287850.2929962, 'wait_for_prev_duration_secs': 4.124641418457031e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776287850.2931945, 'checkpointer_blocking_duration_secs': 0.8224353790283203, 'get_old_steps_start_time': 1776287851.1156452, 'get_old_steps_duration_secs': 1.9788742065429688e-05, 'checkpoint_manager_blocking_start_time': 1776287850.2929134, 'checkpoint_manager_blocking_duration_secs': 0.823115348815918} I0415 21:17:31.116183 130761292586112 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0415 21:17:31.116223 130761292586112 max_utils.py:750] Memstats: After params initialized: I0415 21:17:31.116304 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_0(process=0,(0,0,0,0)) I0415 21:17:31.116331 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_1(process=0,(1,0,0,0)) I0415 21:17:31.116351 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_2(process=0,(0,1,0,0)) I0415 21:17:31.116373 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_3(process=0,(1,1,0,0)) I0415 21:17:31.116390 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_4(process=0,(0,2,0,0)) I0415 21:17:31.116406 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_5(process=0,(1,2,0,0)) I0415 21:17:31.116423 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_6(process=0,(0,3,0,0)) I0415 21:17:31.116439 130761292586112 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_7(process=0,(1,3,0,0)) W0415 21:17:31.122249 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:31.127605 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. W0415 21:17:31.133326 3230093 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions. I0415 21:17:31.223911 130648062821952 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/items I0415 21:17:31.437317 130648083793472 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776287851132219981, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776287851132219981, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0415 21:17:31.447264 130761292586112 metric_logger.py:185] completed step: 0, seconds: 3.122, TFLOP/s/device: 0.257, Tokens/s/device: 40.998, total_weights: 963, loss: 10.834 I0415 21:17:31.448276 130761292586112 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/tensorboard/' I0415 21:17:31.508710 130761292586112 metric_logger.py:185] completed step: 1, seconds: 1.218, TFLOP/s/device: 0.658, Tokens/s/device: 105.089, total_weights: 945, loss: 10.747 I0415 21:17:31.572379 130761292586112 metric_logger.py:185] completed step: 2, seconds: 0.010, TFLOP/s/device: 82.821, Tokens/s/device: 13236.815, total_weights: 962, loss: 10.246 I0415 21:17:31.636003 130761292586112 metric_logger.py:185] completed step: 3, seconds: 0.062, TFLOP/s/device: 12.870, Tokens/s/device: 2056.919, total_weights: 975, loss: 9.829 I0415 21:17:31.833922 3232732 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0415 21:17:31.833990 3232732 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0415 21:17:32.598725 130648022976064 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/items/array_metadatas/process_0 I0415 21:17:42.890447 130761292586112 metric_logger.py:185] completed step: 4, seconds: 0.063, TFLOP/s/device: 12.629, Tokens/s/device: 2018.354, total_weights: 962, loss: 9.393 I0415 21:17:42.898864 130761292586112 metric_logger.py:185] completed step: 5, seconds: 0.064, TFLOP/s/device: 12.550, Tokens/s/device: 2005.767, total_weights: 958, loss: 9.248 I0415 21:17:42.961123 130761292586112 metric_logger.py:185] completed step: 6, seconds: 11.253, TFLOP/s/device: 0.071, Tokens/s/device: 11.374, total_weights: 930, loss: 9.424 I0415 21:17:43.024701 130761292586112 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 129.592, Tokens/s/device: 20711.974, total_weights: 935, loss: 9.317 I0415 21:17:43.088307 130761292586112 metric_logger.py:185] completed step: 8, seconds: 0.063, TFLOP/s/device: 12.739, Tokens/s/device: 2035.947, total_weights: 930, loss: 9.478 I0415 21:17:43.151898 130761292586112 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0415 21:17:43.152689 130761292586112 checkpointing.py:776] Waited 0.0008187294006347656 seconds for step 9 to finish before starting checkpointing. I0415 21:17:43.152944 130761292586112 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 21:18:07.331985 130648012490304 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.500966s (commit=35.062727s, array_metadata_write=0.438239s) I0415 21:18:07.332983 130648002004544 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 342.074 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 36 seconds) (per-host) I0415 21:18:07.333100 130648002004544 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.217526s. I0415 21:18:07.573149 130648002004544 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776287851132219981, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0415 21:18:07.746792 130648002004544 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 21:18:07.977402 130648083793472 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776287851132219981, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/_CHECKPOINT_METADATA I0415 21:18:08.207992 130648002004544 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 21:18:08.208182 130648002004544 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.592534s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/items I0415 21:18:08.208853 130648002004544 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0/items I0415 21:18:08.453284 130648002004544 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0 I0415 21:18:09.175122 130648002004544 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0`. I0415 21:18:09.175764 130648002004544 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 38.882557s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/0 I0415 21:18:09.175840 130648002004544 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 38.060268s. I0415 21:18:09.176022 130647991518784 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 21:18:09.176136 130647991518784 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 21:18:09.176196 130647991518784 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0415 21:18:09.176241 130647991518784 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0415 21:18:09.176369 130761292586112 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 21:18:09.176487 130761292586112 checkpoint_manager.py:1441] Waiting for previous save to complete took 26.023545 seconds. If this number is high, consider checkpointing less frequently. I0415 21:18:09.177218 130761292586112 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0415 21:18:09.177501 130761292586112 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9. I0415 21:18:09.364863 130761292586112 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0415 21:18:09.364963 130761292586112 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 21:18:09.366654 130647991518784 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9 I0415 21:18:10.088408 130648062821952 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9/items I0415 21:18:21.604323 130761292586112 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.239855s I0415 21:18:21.609863 130761292586112 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.001 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host) I0415 21:18:21.609957 130761292586112 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.332962s (batch_requests_ready=0.084641s, total_serialization_initiated=12.242958s, others=0.005363s) I0415 21:18:21.610083 130761292586112 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.333544s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=12.333526) I0415 21:18:21.612359 130648022976064 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0415 21:18:21.612488 130761292586112 async_checkpointer.py:561] Finished blocking save. Time taken: 12.435207s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9. I0415 21:18:21.612811 130761292586112 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0415 21:18:21.613136 130648002004544 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0415 21:18:21.613321 130761292586112 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776287863.1529133, 'wait_for_prev_duration_secs': 26.023545026779175, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776287889.1772459, 'checkpointer_blocking_duration_secs': 12.43539571762085, 'get_old_steps_start_time': 1776287901.6126695, 'get_old_steps_duration_secs': 6.318092346191406e-05, 'checkpoint_manager_blocking_start_time': 1776287863.1528735, 'checkpoint_manager_blocking_duration_secs': 38.460405349731445} I0415 21:18:21.613556 130761292586112 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0415 21:18:21.613598 130761292586112 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0415 21:18:22.397733 130646890513984 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9/items/array_metadatas/process_0 I0415 21:18:53.908787 130648073307712 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 32.297783s (commit=31.847266s, array_metadata_write=0.450516s) I0415 21:18:53.910121 130648022976064 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 283.126 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 44 seconds) (per-host) I0415 21:18:53.910196 130648022976064 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.297649s. I0415 21:18:54.352530 130648022976064 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 21:18:54.774084 130648022976064 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0415 21:18:54.775249 130648022976064 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.574330s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9/items I0415 21:18:54.776101 130648022976064 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9/items I0415 21:18:55.019463 130648022976064 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9 I0415 21:18:55.712792 130648022976064 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9`. I0415 21:18:55.713673 130648022976064 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 46.536411s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_03_dropout/checkpoints/9 I0415 21:18:55.713757 130648022976064 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 34.101211s. I0415 21:18:55.714015 130648002004544 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0415 21:18:55.714182 130648002004544 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0415 21:18:55.714251 130648002004544 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0415 21:18:55.714340 130648002004544 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0415 21:18:55.714735 130761292586112 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 21:18:55.714928 130761292586112 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0415 21:18:55.715925 130761292586112 metric_logger.py:185] completed step: 9, seconds: 0.064, TFLOP/s/device: 12.609, Tokens/s/device: 2015.272, 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