2026-04-17 13:23:48.629623: 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 13:23:48.746642 127253858557056 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0417 13:24:38.492293 127253858557056 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 13:24:38.492406 127253858557056 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 13:24:38.492447 127253858557056 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 13:24:38.492470 127253858557056 train_utils.py:335] WARNING: 'dataset_path' might be pointing your local file system I0417 13:24:38.492567 127253858557056 train.py:671] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0417 13:24:39.005815 127253858557056 maxtext_utils.py:1548] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 13:24:39.006161 127253858557056 checkpointing.py:677] Setting up checkpoint logger... I0417 13:24:39.006207 127253858557056 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 13:24:39.006270 127253858557056 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0417 13:24:39.006893 127253858557056 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 0x73bc02bfce30>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 13:24:41.441163 127253858557056 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0417 13:24:41.441579 127253858557056 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 0x73b6fd5a08f0>}, handler_registry=None I0417 13:24:41.442091 127253858557056 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x73b6fd5a08f0>` 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 13:24:41.442140 127253858557056 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x73b5863323f0>` 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 13:24:41.442178 127253858557056 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 0x73b6fd5a08f0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x73b6fd5a08f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x73b5863323f0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x73b5863323f0>}). I0417 13:24:41.442966 127253858557056 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0417 13:24:41.443053 127253858557056 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x73b58632a200> timeout: 600 secs and primary_host=0 for async checkpoint writes I0417 13:24:41.593123 127253858557056 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints I0417 13:24:41.593382 127253858557056 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_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x73b586332270> I0417 13:24:41.593514 127253858557056 checkpointing.py:301] Checkpoint manager created! I0417 13:24:42.042010 127253858557056 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0417 13:24:42.044950 127253858557056 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 13:24:42.098694 127253858557056 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 13:24:42.121870 127253858557056 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0417 13:24:42.121932 127253858557056 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0417 13:24:42.764609 127253858557056 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 13:24:42.764733 127253858557056 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 13:24:42.867470 127253858557056 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 13:24:42.867547 127253858557056 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 13:24:42.882458 127253858557056 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 13:24:42.882507 127253858557056 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0417 13:24:42.904689 127253858557056 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 13:24:42.904750 127253858557056 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 13:24:42.919804 127253858557056 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 13:24:42.919852 127253858557056 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 13:24:42.934780 127253858557056 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 13:24:42.934828 127253858557056 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 13:24:42.957055 127253858557056 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 13:24:42.957118 127253858557056 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0417 13:24:42.976160 127253858557056 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 13:24:42.976213 127253858557056 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0417 13:24:43.174028 127253858557056 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0417 13:24:43.174142 127253858557056 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint. [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 13:24:44.548270 127253858557056 maxtext_utils.py:1651] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 13:24:44.548382 127253858557056 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 13:24:44.548424 127253858557056 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 13:24:44.548470 127253858557056 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 13:24:44.548509 127253858557056 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 13:24:44.548538 127253858557056 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 13:24:44.548576 127253858557056 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 13:24:44.548613 127253858557056 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 13:24:44.548640 127253858557056 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 13:24:44.548665 127253858557056 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 13:24:44.548702 127253858557056 maxtext_utils.py:1651] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 13:24:44.548735 127253858557056 maxtext_utils.py:1651] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 13:24:44.982452 127253858557056 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 13:24:44.982532 127253858557056 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0417 13:24:44.996660 127253858557056 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0417 13:24:44.996710 127253858557056 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0417 13:24:52.478980 127253858557056 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 13:24:52.479916 127253858557056 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0417 13:24:52.480100 127253858557056 metric_logger.py:301] number parameters: 1.104 billion I0417 13:25:02.640632 127253858557056 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0417 13:25:02.770534 127253858557056 checkpointing.py:776] Waited 0.12989592552185059 seconds for step 0 to finish before starting checkpointing. I0417 13:25:02.771095 127253858557056 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 13:25:02.771262 127253858557056 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0417 13:25:02.771696 127253858557056 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0. I0417 13:25:02.857485 127253858557056 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0417 13:25:03.353870 127138986460736 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0 I0417 13:25:03.354356 127253858557056 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 13:25:03.354585 127253858557056 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 13:25:04.057697 127253858557056 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.704325s I0417 13:25:04.058222 127253858557056 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 10.287 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0417 13:25:04.058300 127253858557056 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.199807s (batch_requests_ready=0.093115s, total_serialization_initiated=1.106315s, others=0.000377s) I0417 13:25:04.058381 127253858557056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.200398s (all_items=0.000028s, per_item={'items': '0.00002766'}, temp_paths=1.200371) I0417 13:25:04.059331 127138873214528 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 13:25:04.059455 127253858557056 async_checkpointer.py:561] Finished blocking save. Time taken: 1.288142s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0. I0417 13:25:04.059693 127253858557056 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:25:04.059894 127138862728768 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:25:04.060012 127253858557056 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776432302.7710814, 'wait_for_prev_duration_secs': 4.2438507080078125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776432302.7712834, 'checkpointer_blocking_duration_secs': 1.28830885887146, 'get_old_steps_start_time': 1776432304.0596194, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776432302.7709703, 'checkpoint_manager_blocking_duration_secs': 1.2889907360076904} I0417 13:25:04.060141 127253858557056 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0417 13:25:04.060191 127253858557056 max_utils.py:750] Memstats: After params initialized: I0417 13:25:04.060467 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0417 13:25:04.060503 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0417 13:25:04.060526 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0417 13:25:04.060546 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0417 13:25:04.060564 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0417 13:25:04.060582 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0417 13:25:04.060599 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0417 13:25:04.060616 127253858557056 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0417 13:25:04.158601 127138929837632 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0/items I0417 13:25:04.305005 127138996946496 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776432304067545530, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776432304067545530, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0417 13:25:04.513919 127253858557056 metric_logger.py:196] completed step: 0, seconds: 10.159, TFLOP/s/device: 1.337, Tokens/s/device: 201.586, total_weights: 13328, loss: 10.828, lm_loss: 10.828, perplexity: 50404.086 I0417 13:25:04.514941 127253858557056 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/tensorboard/' I0417 13:25:04.629129 127253858557056 metric_logger.py:196] completed step: 1, seconds: 1.871, TFLOP/s/device: 7.263, Tokens/s/device: 1094.827, total_weights: 12332, loss: 10.828, lm_loss: 10.828, perplexity: 50391.758 I0417 13:25:04.746866 127253858557056 metric_logger.py:196] completed step: 2, seconds: 0.011, TFLOP/s/device: 1255.278, Tokens/s/device: 189209.165, total_weights: 15161, loss: 9.823, lm_loss: 9.823, perplexity: 18444.867 I0417 13:25:04.842309 461831 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0417 13:25:04.842384 461831 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0417 13:25:04.864837 127253858557056 metric_logger.py:196] completed step: 3, seconds: 0.116, TFLOP/s/device: 117.142, Tokens/s/device: 17656.847, total_weights: 13327, loss: 9.405, lm_loss: 9.405, perplexity: 12149.331 I0417 13:25:05.798338 127138894186048 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_10_shardy_false/checkpoints/0/items/array_metadatas/process_0 I0417 13:25:24.936180 127253858557056 metric_logger.py:196] completed step: 4, seconds: 0.118, TFLOP/s/device: 115.317, Tokens/s/device: 17381.858, total_weights: 11939, loss: 9.045, lm_loss: 9.045, perplexity: 8471.842 I0417 13:25:24.945482 127253858557056 metric_logger.py:196] completed step: 5, seconds: 0.119, TFLOP/s/device: 113.951, Tokens/s/device: 17175.877, total_weights: 15502, loss: 8.953, lm_loss: 8.953, perplexity: 7728.546 I0417 13:25:25.062416 127253858557056 metric_logger.py:196] completed step: 6, seconds: 20.069, TFLOP/s/device: 0.677, Tokens/s/device: 102.047, total_weights: 13864, loss: 8.775, lm_loss: 8.775, perplexity: 6468.536 I0417 13:25:25.179774 127253858557056 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2119.346, Tokens/s/device: 319450.944, total_weights: 12988, loss: 8.688, lm_loss: 8.688, perplexity: 5928.887 I0417 13:25:25.297020 127253858557056 metric_logger.py:196] completed step: 8, seconds: 0.118, TFLOP/s/device: 114.964, Tokens/s/device: 17328.618, total_weights: 13820, loss: 8.740, lm_loss: 8.740, perplexity: 6250.123 I0417 13:25:25.413699 127253858557056 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0417 13:25:25.414537 127253858557056 checkpointing.py:776] Waited 0.0008623600006103516 seconds for step 9 to finish before starting checkpointing. I0417 13:25:25.414813 127253858557056 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:25:28.475619 127138883700288 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 23.631372s (commit=23.179430s, array_metadata_write=0.451942s) I0417 13:25:28.476660 127138873214528 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 493.276 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 25 seconds) (per-host) I0417 13:25:28.476768 127138873214528 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 24.417251s. I0417 13:25:28.726421 127138873214528 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776432304067545530, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0417 13:25:28.896710 127138873214528 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 13:25:29.118367 127138996946496 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776432304067545530, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0417 13:25:29.329240 127138873214528 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:25:29.329406 127138873214528 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.560766s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0/items I0417 13:25:29.330584 127138873214528 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0/items I0417 13:25:29.574613 127138873214528 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0 I0417 13:25:30.262670 127138873214528 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_10_shardy_false/checkpoints/0`. I0417 13:25:30.263321 127138873214528 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 27.492021s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/0 I0417 13:25:30.263402 127138873214528 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 26.203887s. I0417 13:25:30.263526 127138862728768 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:25:30.263575 127138862728768 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:25:30.263629 127138862728768 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:25:30.263669 127138862728768 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 13:25:30.263791 127253858557056 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 13:25:30.263925 127253858557056 checkpoint_manager.py:1441] Waiting for previous save to complete took 4.849115 seconds. If this number is high, consider checkpointing less frequently. I0417 13:25:30.264635 127253858557056 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0417 13:25:30.264944 127253858557056 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9. I0417 13:25:30.439149 127253858557056 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0417 13:25:30.439240 127253858557056 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 13:25:30.440780 127138862728768 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9 I0417 13:25:31.121879 127138726413888 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9/items I0417 13:25:34.206056 127253858557056 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 3.767331s I0417 13:25:34.206569 127253858557056 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.200 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 3 seconds) (per-host) I0417 13:25:34.206637 127253858557056 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 3.856780s (batch_requests_ready=0.085853s, total_serialization_initiated=3.770512s, others=0.000415s) I0417 13:25:34.206703 127253858557056 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 3.857292s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=3.857273) I0417 13:25:34.212670 127138862728768 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0417 13:25:34.212788 127253858557056 async_checkpointer.py:561] Finished blocking save. Time taken: 3.948101s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9. I0417 13:25:34.213015 127253858557056 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 13:25:34.213197 127138919351872 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0417 13:25:34.213273 127253858557056 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776432325.414777, 'wait_for_prev_duration_secs': 4.849114656448364, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776432330.2646592, 'checkpointer_blocking_duration_secs': 3.94826340675354, 'get_old_steps_start_time': 1776432334.2129462, 'get_old_steps_duration_secs': 2.9802322387695312e-05, 'checkpoint_manager_blocking_start_time': 1776432325.414732, 'checkpoint_manager_blocking_duration_secs': 8.798513889312744} I0417 13:25:34.213382 127253858557056 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0417 13:25:34.213417 127253858557056 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 13:25:34.960529 127138986460736 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_10_shardy_false/checkpoints/9/items/array_metadatas/process_0 I0417 13:26:07.691060 127138726413888 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 33.480182s (commit=33.058791s, array_metadata_write=0.421391s) I0417 13:26:07.692088 127138862728768 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 338.405 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 37 seconds) (per-host) I0417 13:26:07.692213 127138862728768 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 33.479368s. I0417 13:26:08.154291 127138862728768 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 13:26:08.577526 127138862728768 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0417 13:26:08.578300 127138862728768 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.569951s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9/items I0417 13:26:08.578965 127138862728768 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9/items I0417 13:26:08.838308 127138862728768 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9 I0417 13:26:09.524251 127138862728768 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_10_shardy_false/checkpoints/9`. I0417 13:26:09.524889 127138862728768 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 39.260215s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260417_125240/linen_main_20260417_125240_10_shardy_false/checkpoints/9 I0417 13:26:09.524967 127138862728768 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 35.312124s. I0417 13:26:09.525079 127138919351872 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 13:26:09.525126 127138919351872 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0417 13:26:09.525173 127138919351872 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 13:26:09.525213 127138919351872 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 13:26:09.525359 127253858557056 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 13:26:09.525487 127253858557056 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 13:26:09.526350 127253858557056 metric_logger.py:196] completed step: 9, seconds: 0.117, TFLOP/s/device: 116.051, Tokens/s/device: 17492.462, 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