2026-04-22 18:35:30.601391: 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 00:00:1776882930.614416 2069808 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776882930.618260 2069808 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776882930.629686 2069808 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882930.629699 2069808 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882930.629701 2069808 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776882930.629703 2069808 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'} 2026-04-22 18:35:34.302001: 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) I0422 18:35:34.437156 130708424578880 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. ~/maxtext/maxtext_venv/lib/python3.12/site-packages/jax/_src/xla_bridge.py:204: UserWarning: TPU backend initialization is taking more than 60.0 seconds. Did you run your code on all TPU hosts? See https://docs.jax.dev/en/latest/multi_process.html for more information. warnings.warn( I0422 18:36:36.674673 130708424578880 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:36:36.674934 130708424578880 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:36:36.674983 130708424578880 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 I0422 18:36:36.675011 130708424578880 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system I0422 18:36:36.675037 130708424578880 train_utils.py:361] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0422 18:36:36.675215 130708424578880 train.py:705] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0422 18:36:37.336745 130708424578880 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:36:37.337021 130708424578880 checkpointing.py:677] Setting up checkpoint logger... I0422 18:36:37.337081 130708424578880 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 18:36:37.337117 130708424578880 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0422 18:36:37.337545 130708424578880 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 0x76e0a83ad9d0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:36:39.761167 130708424578880 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0422 18:36:39.761627 130708424578880 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 0x76df483d3170>}, handler_registry=None I0422 18:36:39.761934 130708424578880 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76df483d3170>` 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`. I0422 18:36:39.761978 130708424578880 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76decd35a7b0>` 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`. I0422 18:36:39.762011 130708424578880 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 0x76df483d3170>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x76df483d3170>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76decd35a7b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x76decd35a7b0>}). I0422 18:36:39.762485 130708424578880 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0422 18:36:39.762549 130708424578880 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x76decd380900> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:36:39.899330 130708424578880 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints I0422 18:36:39.899580 130708424578880 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_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x76df46f3c980> I0422 18:36:39.899866 130708424578880 checkpointing.py:301] Checkpoint manager created! I0422 18:36:40.289545 130708424578880 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:36:40.289649 130708424578880 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:36:40.576878 130708424578880 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:36:40.576968 130708424578880 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:36:40.589225 130708424578880 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:36:40.589277 130708424578880 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:36:40.608501 130708424578880 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:36:40.608556 130708424578880 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:36:40.620897 130708424578880 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:36:40.620951 130708424578880 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:36:40.633255 130708424578880 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0422 18:36:40.633308 130708424578880 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:36:40.652847 130708424578880 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:36:40.652915 130708424578880 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:36:40.668407 130708424578880 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:36:40.668461 130708424578880 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0422 18:36:40.819917 130708424578880 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:36:40.820019 130708424578880 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 I0422 18:36:41.458570 130708424578880 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0422 18:36:41.458690 130708424578880 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) I0422 18:36:41.458742 130708424578880 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) I0422 18:36:41.458790 130708424578880 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') I0422 18:36:41.458831 130708424578880 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) I0422 18:36:41.458860 130708424578880 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) I0422 18:36:41.458900 130708424578880 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) I0422 18:36:41.458940 130708424578880 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') I0422 18:36:41.458970 130708424578880 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) I0422 18:36:41.458997 130708424578880 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) I0422 18:36:41.459035 130708424578880 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0422 18:36:41.459097 130708424578880 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0422 18:36:41.870681 130708424578880 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0422 18:36:41.870781 130708424578880 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0422 18:36:41.882412 130708424578880 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0422 18:36:41.882464 130708424578880 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0422 18:36:42.643263 130708424578880 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. I0422 18:36:42.643781 130708424578880 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0422 18:36:42.643935 130708424578880 metric_logger.py:301] number parameters: 1.104 billion I0422 18:36:45.007182 130708424578880 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0422 18:36:45.125170 130708424578880 checkpointing.py:776] Waited 0.11797070503234863 seconds for step 0 to finish before starting checkpointing. I0422 18:36:45.125824 130708424578880 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:36:45.125978 130708424578880 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0422 18:36:45.126400 130708424578880 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0. I0422 18:36:45.214714 130708424578880 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0422 18:36:45.332290 130677557167680 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0 I0422 18:36:45.339726 130708424578880 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:36:45.339820 130708424578880 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 I0422 18:36:46.075129 130677546681920 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/items I0422 18:36:46.229038 130677150320192 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776883005984821655, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776883005984821655, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0422 18:36:46.589084 2071966 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0422 18:36:46.589143 2071966 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0422 18:36:46.655306 130708424578880 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.316019s I0422 18:36:46.660324 130708424578880 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 8.543 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host) I0422 18:36:46.660403 130708424578880 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.444629s (batch_requests_ready=0.119576s, total_serialization_initiated=1.320177s, others=0.004876s) I0422 18:36:46.660511 130708424578880 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.445234s (all_items=0.000020s, per_item={'items': '0.00002027'}, temp_paths=1.445213) I0422 18:36:46.662313 130676154172992 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:36:46.662431 130708424578880 async_checkpointer.py:561] Finished blocking save. Time taken: 1.536407s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0. I0422 18:36:46.662757 130708424578880 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:36:46.663100 130677567653440 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:36:46.663222 130708424578880 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883005.1258118, 'wait_for_prev_duration_secs': 4.172325134277344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883005.1259997, 'checkpointer_blocking_duration_secs': 1.5365819931030273, 'get_old_steps_start_time': 1776883006.6626058, 'get_old_steps_duration_secs': 8.988380432128906e-05, 'checkpoint_manager_blocking_start_time': 1776883005.125727, 'checkpoint_manager_blocking_duration_secs': 1.5374634265899658} I0422 18:36:46.663353 130708424578880 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0422 18:36:46.663400 130708424578880 max_utils.py:750] Memstats: After params initialized: I0422 18:36:46.663456 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0422 18:36:46.663484 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0422 18:36:46.663507 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0422 18:36:46.663527 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0422 18:36:46.663545 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0422 18:36:46.663563 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0422 18:36:46.663581 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0422 18:36:46.663599 130708424578880 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0422 18:36:47.018226 130708424578880 metric_logger.py:196] completed step: 0, seconds: 2.363, TFLOP/s/device: 5.751, Tokens/s/device: 866.814, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0422 18:36:47.019308 130708424578880 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/tensorboard/' I0422 18:36:47.134584 130708424578880 metric_logger.py:196] completed step: 1, seconds: 2.009, TFLOP/s/device: 6.764, Tokens/s/device: 1019.523, total_weights: 16384, loss: 10.874, lm_loss: 10.874, perplexity: 52770.941 I0422 18:36:47.254363 130708424578880 metric_logger.py:196] completed step: 2, seconds: 0.008, TFLOP/s/device: 1780.284, Tokens/s/device: 268343.816, total_weights: 16384, loss: 9.775, lm_loss: 9.775, perplexity: 17589.287 I0422 18:36:47.374332 130708424578880 metric_logger.py:196] completed step: 3, seconds: 0.118, TFLOP/s/device: 114.745, Tokens/s/device: 17295.690, total_weights: 16384, loss: 8.838, lm_loss: 8.838, perplexity: 6892.933 I0422 18:36:47.428400 130676546340416 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0 I0422 18:37:05.075689 130708424578880 metric_logger.py:196] completed step: 4, seconds: 0.119, TFLOP/s/device: 114.428, Tokens/s/device: 17247.768, total_weights: 16384, loss: 7.965, lm_loss: 7.965, perplexity: 2878.644 I0422 18:37:05.086396 130708424578880 metric_logger.py:196] completed step: 5, seconds: 0.120, TFLOP/s/device: 113.578, Tokens/s/device: 17119.738, total_weights: 16384, loss: 7.233, lm_loss: 7.233, perplexity: 1383.750 I0422 18:37:05.203265 130708424578880 metric_logger.py:196] completed step: 6, seconds: 17.705, TFLOP/s/device: 0.767, Tokens/s/device: 115.677, total_weights: 16384, loss: 6.676, lm_loss: 6.676, perplexity: 793.531 I0422 18:37:05.322901 130708424578880 metric_logger.py:196] completed step: 7, seconds: 0.006, TFLOP/s/device: 2442.850, Tokens/s/device: 368212.873, total_weights: 16384, loss: 6.293, lm_loss: 6.293, perplexity: 540.860 I0422 18:37:05.442294 130708424578880 metric_logger.py:196] completed step: 8, seconds: 0.118, TFLOP/s/device: 114.878, Tokens/s/device: 17315.725, total_weights: 16384, loss: 6.055, lm_loss: 6.055, perplexity: 426.371 I0422 18:37:05.560574 130708424578880 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0422 18:37:05.561636 130708424578880 checkpointing.py:776] Waited 0.0010895729064941406 seconds for step 9 to finish before starting checkpointing. I0422 18:37:05.561992 130708424578880 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:37:29.177134 130676355499584 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.515901s (commit=42.004148s, array_metadata_write=0.511753s) I0422 18:37:29.178337 130676154172992 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 287.444 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 43 seconds) (per-host) I0422 18:37:29.178406 130676154172992 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.515912s. I0422 18:37:29.479643 130676154172992 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776883005984821655, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0422 18:37:29.708565 130676154172992 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0422 18:37:29.896848 130677150320192 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776883005984821655, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA I0422 18:37:30.116065 130676154172992 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0422 18:37:30.116296 130676154172992 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.596643s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/items I0422 18:37:30.117102 130676154172992 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0/items I0422 18:37:30.376341 130676154172992 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0 I0422 18:37:31.070009 130676154172992 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0`. I0422 18:37:31.070795 130676154172992 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 45.944781s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/0 I0422 18:37:31.070882 130676154172992 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 44.408392s. I0422 18:37:31.071092 130677567653440 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:37:31.071210 130677567653440 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:37:31.071277 130677567653440 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:37:31.071332 130677567653440 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 18:37:31.071437 130708424578880 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. W0422 18:37:31.071519 130708424578880 checkpoint_manager.py:1441] Waiting for previous save to complete took 25.509577 seconds. If this number is high, consider checkpointing less frequently. I0422 18:37:31.072804 130708424578880 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0422 18:37:31.073103 130708424578880 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9. I0422 18:37:31.238859 130676154172992 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9 I0422 18:37:31.253205 130708424578880 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:37:31.253316 130708424578880 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 I0422 18:37:31.981817 130676546340416 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9/items I0422 18:37:34.717076 130708424578880 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 3.464317s I0422 18:37:34.717438 130708424578880 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 3.466 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 3 seconds) (per-host) I0422 18:37:34.717498 130708424578880 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 3.560615s (batch_requests_ready=0.092849s, total_serialization_initiated=3.467528s, others=0.000238s) I0422 18:37:34.717563 130708424578880 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 3.561191s (all_items=0.000024s, per_item={'items': '0.00002432'}, temp_paths=3.561167) I0422 18:37:34.722875 130676355499584 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:37:34.722985 130708424578880 async_checkpointer.py:561] Finished blocking save. Time taken: 3.650132s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9. I0422 18:37:34.723204 130708424578880 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:37:34.723368 130677567653440 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:37:34.723474 130708424578880 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776883025.5619123, 'wait_for_prev_duration_secs': 25.509576559066772, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776883051.0728278, 'checkpointer_blocking_duration_secs': 3.6502835750579834, 'get_old_steps_start_time': 1776883054.7231293, 'get_old_steps_duration_secs': 3.7670135498046875e-05, 'checkpoint_manager_blocking_start_time': 1776883025.561861, 'checkpoint_manager_blocking_duration_secs': 29.161587953567505} I0422 18:37:34.723585 130708424578880 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0422 18:37:34.723618 130708424578880 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:37:35.492108 130677196457536 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0 I0422 18:38:10.657126 130676154172992 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.939124s (commit=35.456477s, array_metadata_write=0.482646s) I0422 18:38:10.658283 130676355499584 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 319.908 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 39 seconds) (per-host) I0422 18:38:10.658413 130676355499584 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.935378s. I0422 18:38:11.156801 130676355499584 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0. I0422 18:38:11.545892 130676355499584 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0422 18:38:11.546110 130676355499584 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.590936s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9/items I0422 18:38:11.546871 130676355499584 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9/items I0422 18:38:11.800515 130676355499584 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9 I0422 18:38:12.520396 130676355499584 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9`. I0422 18:38:12.521175 130676355499584 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 41.448333s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_13_scan_layers_true/checkpoints/9 I0422 18:38:12.521254 130676355499584 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 37.798222s. I0422 18:38:12.521412 130677567653440 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:38:12.521527 130677567653440 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:38:12.521589 130677567653440 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:38:12.521628 130677567653440 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 18:38:12.525472 130708424578880 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. I0422 18:38:12.525621 130708424578880 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:38:12.526977 130708424578880 metric_logger.py:196] completed step: 9, seconds: 0.119, TFLOP/s/device: 113.723, Tokens/s/device: 17141.518, total_weights: 16384, loss: 5.915, lm_loss: 5.915, perplexity: 370.653 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops