2026-04-23 01:58:30.150783: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered WARNING: All log messages before absl::InitializeLog() is called are written to STDERR E0000 00:00:1776909510.164383 2297650 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776909510.168276 2297650 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776909510.179867 2297650 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776909510.179883 2297650 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776909510.179884 2297650 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776909510.179887 2297650 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} `rope_scaling`'s factor field must be a float >= 1, got 40 `rope_scaling`'s beta_fast field must be a float, got 32 `rope_scaling`'s beta_slow field must be a float, got 1 Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} Unrecognized keys in `rope_scaling` for 'rope_type'='yarn': {'rope_theta'} 2026-04-23 01:58:33.858243: 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) I0423 01:58:33.993119 127053407999808 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0423 01:59:06.230439 127053407999808 max_utils.py:800] System Information: Jax Version: 0.8.1 I0423 01:59:06.230698 127053407999808 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0423 01:59:06.230749 127053407999808 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Nov 12 2025 14:16:36 (1762985796) cl/831091709 I0423 01:59:06.230779 127053407999808 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system I0423 01:59:06.230804 127053407999808 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0423 01:59:06.230967 127053407999808 train.py:703] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0423 01:59:06.883154 127053407999808 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0423 01:59:06.883423 127053407999808 checkpointing.py:677] Setting up checkpoint logger... I0423 01:59:06.883465 127053407999808 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0423 01:59:06.883500 127053407999808 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0423 01:59:06.883926 127053407999808 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 0x738da3bb7110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0423 01:59:09.262760 127053407999808 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0423 01:59:09.263163 127053407999808 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 0x738bd1b5dbb0>}, handler_registry=None I0423 01:59:09.263395 127053407999808 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x738bd1b5dbb0>` 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`. I0423 01:59:09.263437 127053407999808 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x738bd1b5e1e0>` 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`. I0423 01:59:09.263466 127053407999808 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 0x738bd1b5dbb0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x738bd1b5dbb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x738bd1b5e1e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x738bd1b5e1e0>}). I0423 01:59:09.263737 127053407999808 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0423 01:59:09.263788 127053407999808 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x738bd1b70720> timeout: 600 secs and primary_host=0 for async checkpoint writes I0423 01:59:09.389714 127053407999808 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints I0423 01:59:09.389945 127053407999808 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_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x738c4839b1a0> I0423 01:59:09.390242 127053407999808 checkpointing.py:301] Checkpoint manager created! I0423 01:59:09.772131 127053407999808 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0423 01:59:09.772230 127053407999808 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 01:59:10.053353 127053407999808 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 01:59:10.053449 127053407999808 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 01:59:10.065603 127053407999808 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0423 01:59:10.065656 127053407999808 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0423 01:59:10.110537 127053407999808 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 01:59:10.110660 127053407999808 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:59:10.123467 127053407999808 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 01:59:10.123519 127053407999808 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:59:10.135711 127053407999808 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0423 01:59:10.135763 127053407999808 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:59:10.156131 127053407999808 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0423 01:59:10.156191 127053407999808 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0423 01:59:10.171731 127053407999808 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0423 01:59:10.171786 127053407999808 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0423 01:59:10.323252 127053407999808 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0423 01:59:10.323358 127053407999808 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 I0423 01:59:12.956789 127053407999808 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0423 01:59:12.956902 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0423 01:59:12.956944 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: PartitionSpec('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0423 01:59:12.956998 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: PartitionSpec('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0423 01:59:12.957039 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0423 01:59:12.957084 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: PartitionSpec('norm', 'layers') Physical: (None, None) I0423 01:59:12.957125 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0423 01:59:12.957165 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: PartitionSpec('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0423 01:59:12.957194 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0423 01:59:12.957221 127053407999808 maxtext_utils.py:1707] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: PartitionSpec('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0423 01:59:12.957257 127053407999808 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0423 01:59:12.957293 127053407999808 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0423 01:59:13.366680 127053407999808 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 01:59:13.366775 127053407999808 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0423 01:59:13.378341 127053407999808 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0423 01:59:13.378391 127053407999808 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0423 01:59:14.184102 127053407999808 max_utils.py:791] Total memory size: 2.4 GB, Output size: 0.5 GB, Temp size: 1.9 GB, Argument size: 0.5 GB, Host temp size: 0.0 GB. I0423 01:59:14.184703 127053407999808 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0423 01:59:14.184855 127053407999808 metric_logger.py:301] number parameters: 1.104 billion I0423 01:59:16.860250 127053407999808 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0423 01:59:17.056130 127053407999808 checkpointing.py:776] Waited 0.19588232040405273 seconds for step 0 to finish before starting checkpointing. I0423 01:59:17.056792 127053407999808 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 01:59:17.056961 127053407999808 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0423 01:59:17.057391 127053407999808 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0. I0423 01:59:17.142643 127053407999808 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0423 01:59:17.215709 127014962988608 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0 I0423 01:59:17.223709 127053407999808 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 01:59:17.223814 127053407999808 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 I0423 01:59:17.851529 127053407999808 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.628219s I0423 01:59:17.852100 127053407999808 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 17.423 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 708 milliseconds) (per-host) I0423 01:59:17.852161 127053407999808 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.708448s (batch_requests_ready=0.075907s, total_serialization_initiated=0.632131s, others=0.000409s) I0423 01:59:17.852231 127053407999808 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.709039s (all_items=0.000027s, per_item={'items': '0.00002718'}, temp_paths=0.709012) I0423 01:59:17.853101 127014551946816 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 01:59:17.853264 127053407999808 async_checkpointer.py:561] Finished blocking save. Time taken: 0.796256s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0. I0423 01:59:17.853450 127053407999808 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0423 01:59:17.853672 127015220938304 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 01:59:17.853834 127053407999808 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776909557.0567796, 'wait_for_prev_duration_secs': 4.2438507080078125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776909557.056983, 'checkpointer_blocking_duration_secs': 0.7963571548461914, 'get_old_steps_start_time': 1776909557.8533645, 'get_old_steps_duration_secs': 4.291534423828125e-05, 'checkpoint_manager_blocking_start_time': 1776909557.0566769, 'checkpoint_manager_blocking_duration_secs': 0.797114372253418} I0423 01:59:17.854126 127053407999808 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0423 01:59:17.854248 127053407999808 max_utils.py:750] Memstats: After params initialized: I0423 01:59:17.854372 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_0(process=0,(0,0,0,0)) I0423 01:59:17.854411 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_1(process=0,(1,0,0,0)) I0423 01:59:17.854437 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_2(process=0,(0,1,0,0)) I0423 01:59:17.854459 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_3(process=0,(1,1,0,0)) I0423 01:59:17.854480 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_4(process=0,(0,2,0,0)) I0423 01:59:17.854499 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_5(process=0,(1,2,0,0)) I0423 01:59:17.854519 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_6(process=0,(0,3,0,0)) I0423 01:59:17.854538 127053407999808 max_utils.py:756] Using (GB) 0.56 / 31.25 (1.792000%) on TPU_7(process=0,(1,3,0,0)) I0423 01:59:17.986514 127014952502848 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/items I0423 01:59:18.193927 127014541461056 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776909557888452766, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776909557888452766, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0423 01:59:18.201192 127053407999808 metric_logger.py:196] completed step: 0, seconds: 2.675, TFLOP/s/device: 5.080, Tokens/s/device: 765.675, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.840 I0423 01:59:18.201921 127053407999808 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/tensorboard/' I0423 01:59:18.396964 127053407999808 metric_logger.py:196] completed step: 1, seconds: 1.338, TFLOP/s/device: 10.151, Tokens/s/device: 1530.119, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.840 I0423 01:59:18.468628 2299761 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0423 01:59:18.468689 2299761 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0423 01:59:18.595025 127053407999808 metric_logger.py:196] completed step: 2, seconds: 0.007, TFLOP/s/device: 1843.322, Tokens/s/device: 277845.611, total_weights: 16384, loss: 9.775, lm_loss: 9.775, perplexity: 17589.299 I0423 01:59:18.794416 127053407999808 metric_logger.py:196] completed step: 3, seconds: 0.196, TFLOP/s/device: 69.146, Tokens/s/device: 10422.445, total_weights: 16384, loss: 8.838, lm_loss: 8.838, perplexity: 6892.454 I0423 01:59:35.632829 127014742787648 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_0 I0423 01:59:35.637869 127053407999808 metric_logger.py:196] completed step: 4, seconds: 0.198, TFLOP/s/device: 68.541, Tokens/s/device: 10331.277, total_weights: 16384, loss: 7.965, lm_loss: 7.965, perplexity: 2879.114 I0423 01:59:35.681704 127053407999808 metric_logger.py:196] completed step: 5, seconds: 0.199, TFLOP/s/device: 68.197, Tokens/s/device: 10279.370, total_weights: 16384, loss: 7.233, lm_loss: 7.233, perplexity: 1383.691 I0423 01:59:35.888911 127053407999808 metric_logger.py:196] completed step: 6, seconds: 16.880, TFLOP/s/device: 0.805, Tokens/s/device: 121.330, total_weights: 16384, loss: 6.677, lm_loss: 6.677, perplexity: 793.605 I0423 01:59:36.088145 127053407999808 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 1869.960, Tokens/s/device: 281860.721, total_weights: 16384, loss: 6.293, lm_loss: 6.293, perplexity: 540.987 I0423 01:59:36.299855 127053407999808 metric_logger.py:196] completed step: 8, seconds: 0.208, TFLOP/s/device: 65.207, Tokens/s/device: 9828.717, total_weights: 16384, loss: 6.055, lm_loss: 6.055, perplexity: 426.449 I0423 01:59:36.506797 127053407999808 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0423 01:59:36.508116 127053407999808 checkpointing.py:776] Waited 0.0015892982482910156 seconds for step 9 to finish before starting checkpointing. I0423 01:59:36.508458 127053407999808 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 02:00:18.880720 127014562432576 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 60.417177s (commit=59.988958s, array_metadata_write=0.428219s) I0423 02:00:18.881994 127014551946816 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 204.683 MiB/s (total gbytes: 12.3 GiB) (time elapsed: a minute) (per-host) I0423 02:00:18.882120 127014551946816 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 61.028893s. I0423 02:00:19.184750 127014551946816 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776909557888452766, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0423 02:00:19.406228 127014551946816 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0423 02:00:19.627475 127014541461056 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776909557888452766, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA I0423 02:00:19.801135 127014551946816 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0423 02:00:19.801322 127014551946816 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.559229s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/items I0423 02:00:19.802096 127014551946816 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0/items I0423 02:00:20.032452 127014551946816 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0 I0423 02:00:20.711256 127014551946816 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0`. I0423 02:00:20.712039 127014551946816 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 63.655043s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/0 I0423 02:00:20.712137 127014551946816 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 62.858911s. I0423 02:00:20.712361 127015220938304 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 02:00:20.712490 127015220938304 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 02:00:20.712556 127015220938304 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0423 02:00:20.712603 127015220938304 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0423 02:00:20.712711 127053407999808 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. W0423 02:00:20.712806 127053407999808 checkpoint_manager.py:1441] Waiting for previous save to complete took 44.204366 seconds. If this number is high, consider checkpointing less frequently. I0423 02:00:20.713890 127053407999808 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0423 02:00:20.714153 127053407999808 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9. I0423 02:00:20.903216 127053407999808 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0423 02:00:20.903356 127014551946816 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9 I0423 02:00:20.903437 127053407999808 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 I0423 02:00:21.312331 127053407999808 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.409832s I0423 02:00:21.312907 127053407999808 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 24.755 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 498 milliseconds) (per-host) I0423 02:00:21.312973 127053407999808 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.498651s (batch_requests_ready=0.085347s, total_serialization_initiated=0.412863s, others=0.000441s) I0423 02:00:21.313066 127053407999808 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.499232s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=0.499217) I0423 02:00:21.314038 127014023464512 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0423 02:00:21.314190 127053407999808 async_checkpointer.py:561] Finished blocking save. Time taken: 0.600252s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9. I0423 02:00:21.314451 127053407999808 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0423 02:00:21.314660 127015220938304 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0423 02:00:21.314821 127053407999808 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776909576.5084105, 'wait_for_prev_duration_secs': 44.2043662071228, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776909620.7139144, 'checkpointer_blocking_duration_secs': 0.6004014015197754, 'get_old_steps_start_time': 1776909621.3143494, 'get_old_steps_duration_secs': 5.030632019042969e-05, 'checkpoint_manager_blocking_start_time': 1776909576.5083656, 'checkpoint_manager_blocking_duration_secs': 44.80641722679138} I0423 02:00:21.314987 127053407999808 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0423 02:00:21.315031 127053407999808 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0423 02:00:21.600976 127014742787648 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9/items I0423 02:00:22.797175 127014426117696 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_0 I0423 02:01:00.172892 127014214305344 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.106118s (commit=37.647624s, array_metadata_write=0.458494s) I0423 02:01:00.174167 127014023464512 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 321.059 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 39 seconds) (per-host) I0423 02:01:00.174319 127014023464512 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.860069s. I0423 02:01:00.631401 127014023464512 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0423 02:01:01.018030 127014023464512 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0423 02:01:01.018260 127014023464512 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.556450s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9/items I0423 02:01:01.019059 127014023464512 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9/items I0423 02:01:01.257167 127014023464512 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9 I0423 02:01:01.982725 127014023464512 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9`. I0423 02:01:01.983486 127014023464512 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 41.269558s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260423_012248/linen_main_20260423_012248_11_optimizer_offload_true/checkpoints/9 I0423 02:01:01.983590 127014023464512 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 40.669344s. I0423 02:01:01.983851 127015220938304 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0423 02:01:01.984006 127015220938304 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0423 02:01:01.984112 127015220938304 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0423 02:01:01.984170 127015220938304 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0423 02:01:01.984382 127053407999808 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. I0423 02:01:01.984559 127053407999808 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0423 02:01:01.985709 127053407999808 metric_logger.py:196] completed step: 9, seconds: 0.200, TFLOP/s/device: 68.094, Tokens/s/device: 10263.863, total_weights: 16384, loss: 5.915, lm_loss: 5.915, perplexity: 370.686 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops