2026-04-22 18:19:38.339363: 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 00:00:1776881978.352382 2054995 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered E0000 00:00:1776881978.356190 2054995 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered W0000 00:00:1776881978.367666 2054995 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776881978.367681 2054995 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776881978.367683 2054995 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once. W0000 00:00:1776881978.367684 2054995 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 18:19:42.045885: 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 18:19:42.180743 140024073193280 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag. I0422 18:20:33.652535 140024073193280 max_utils.py:800] System Information: Jax Version: 0.8.1 I0422 18:20:33.652783 140024073193280 max_utils.py:801] System Information: Jaxlib Version: 0.8.1 I0422 18:20:33.652833 140024073193280 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 18:20:33.652863 140024073193280 train_utils.py:348] WARNING: 'dataset_path' might be pointing your local file system I0422 18:20:33.653060 140024073193280 train.py:705] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper. I0422 18:20:34.317041 140024073193280 maxtext_utils.py:1604] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0422 18:20:34.317314 140024073193280 checkpointing.py:677] Setting up checkpoint logger... I0422 18:20:34.317359 140024073193280 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True I0422 18:20:34.317394 140024073193280 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None I0422 18:20:34.317837 140024073193280 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 0x7f599b9aaea0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0422 18:20:36.740099 140024073193280 checkpointing.py:265] Enabling policy for fixed interval checkpointing. I0422 18:20:36.740476 140024073193280 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 0x7f57c4f55850>}, handler_registry=None I0422 18:20:36.740709 140024073193280 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f57c4f55850>` 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 18:20:36.740749 140024073193280 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f57c4f55df0>` 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 18:20:36.740779 140024073193280 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 0x7f57c4f55850>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f57c4f55850>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f57c4f55df0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f57c4f55df0>}). I0422 18:20:36.741063 140024073193280 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28 I0422 18:20:36.741114 140024073193280 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7f57c4f64900> timeout: 600 secs and primary_host=0 for async checkpoint writes I0422 18:20:36.868073 140024073193280 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints I0422 18:20:36.868305 140024073193280 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_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f5840b7e2a0> I0422 18:20:36.868580 140024073193280 checkpointing.py:301] Checkpoint manager created! I0422 18:20:36.947115 140024073193280 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0 I0422 18:20:36.949958 140024073193280 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. I0422 18:20:37.010137 140024073193280 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 I0422 18:20:37.032735 140024073193280 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2 I0422 18:20:37.032800 140024073193280 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2 I0422 18:20:37.634748 140024073193280 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0422 18:20:37.634849 140024073193280 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:20:37.937142 140024073193280 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:20:37.937234 140024073193280 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:20:37.950038 140024073193280 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0422 18:20:37.950109 140024073193280 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None). I0422 18:20:37.969383 140024073193280 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 18:20:37.969445 140024073193280 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:20:37.982385 140024073193280 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 18:20:37.982441 140024073193280 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:20:37.995242 140024073193280 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 18:20:37.995292 140024073193280 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:20:38.015448 140024073193280 attentions.py:1198] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0422 18:20:38.015506 140024073193280 attentions.py:1198] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None). I0422 18:20:38.031602 140024073193280 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0422 18:20:38.031662 140024073193280 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None). I0422 18:20:38.187622 140024073193280 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint I0422 18:20:38.187729 140024073193280 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 18:20:38.763352 140024073193280 maxtext_utils.py:1707] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: PartitionSpec('norm',) Physical: (None,) I0422 18:20:38.763468 140024073193280 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 18:20:38.763512 140024073193280 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 18:20:38.763560 140024073193280 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 18:20:38.763602 140024073193280 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 18:20:38.763630 140024073193280 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 18:20:38.763670 140024073193280 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 18:20:38.763711 140024073193280 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 18:20:38.763742 140024073193280 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 18:20:38.763770 140024073193280 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 18:20:38.763808 140024073193280 maxtext_utils.py:1707] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: PartitionSpec('embed_vocab', 'vocab') Physical: ('fsdp', None) I0422 18:20:38.763842 140024073193280 maxtext_utils.py:1707] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: PartitionSpec('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0422 18:20:39.452215 140024073193280 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0422 18:20:39.452303 140024073193280 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None). I0422 18:20:39.464384 140024073193280 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length'). I0422 18:20:39.464436 140024073193280 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None). I0422 18:20:47.272604 140024073193280 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 18:20:47.273373 140024073193280 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter. I0422 18:20:47.273525 140024073193280 metric_logger.py:301] number parameters: 1.104 billion I0422 18:20:57.185718 140024073193280 checkpointing.py:772] Waiting for step 0 to finish before checkpoint... I0422 18:20:57.313695 140024073193280 checkpointing.py:776] Waited 0.12796568870544434 seconds for step 0 to finish before starting checkpointing. I0422 18:20:57.318181 140024073193280 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:20:57.318362 140024073193280 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0 I0422 18:20:57.318805 140024073193280 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0. I0422 18:20:57.416610 140024073193280 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient I0422 18:20:57.504812 139899678230080 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0 I0422 18:20:57.510388 140024073193280 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:20:57.510481 140024073193280 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 18:20:58.325133 139899667744320 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0/items I0422 18:20:58.482771 139899462223424 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776882058233106702, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776882058233106702, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0422 18:20:58.833798 2058161 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json I0422 18:20:58.833857 2058161 google_auth_provider.cc:156] Using OAuth2 AuthProvider I0422 18:20:59.940909 140024073193280 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.430996s I0422 18:20:59.946805 140024073193280 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.880 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0422 18:20:59.946872 140024073193280 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.528709s (batch_requests_ready=0.088062s, total_serialization_initiated=2.434880s, others=0.005767s) I0422 18:20:59.946953 140024073193280 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.529774s (all_items=0.000026s, per_item={'items': '0.00002551'}, temp_paths=2.529749) I0422 18:20:59.948389 139899688715840 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:20:59.948493 140024073193280 async_checkpointer.py:561] Finished blocking save. Time taken: 2.630089s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0. I0422 18:20:59.948686 140024073193280 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:20:59.948853 139898847757888 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:20:59.948951 140024073193280 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882057.3181684, 'wait_for_prev_duration_secs': 4.649162292480469e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776882057.318384, 'checkpointer_blocking_duration_secs': 2.6302218437194824, 'get_old_steps_start_time': 1776882059.9486213, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776882057.3180737, 'checkpoint_manager_blocking_duration_secs': 2.6308505535125732} I0422 18:20:59.949082 140024073193280 checkpointing.py:408] Started an asynchronous checkpoint save for step 0 I0422 18:20:59.949142 140024073193280 max_utils.py:750] Memstats: After params initialized: I0422 18:20:59.949193 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0)) I0422 18:20:59.949218 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0)) I0422 18:20:59.949238 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0)) I0422 18:20:59.949257 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0)) I0422 18:20:59.949274 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0)) I0422 18:20:59.949291 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0)) I0422 18:20:59.949308 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0)) I0422 18:20:59.949325 140024073193280 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0)) I0422 18:21:00.308248 140024073193280 metric_logger.py:196] completed step: 0, seconds: 9.912, TFLOP/s/device: 1.371, Tokens/s/device: 206.626, total_weights: 13328, loss: 10.828, lm_loss: 10.828, perplexity: 50404.086 I0422 18:21:00.309174 140024073193280 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/tensorboard/' I0422 18:21:00.423835 140024073193280 metric_logger.py:196] completed step: 1, seconds: 3.120, TFLOP/s/device: 4.354, Tokens/s/device: 656.354, total_weights: 12332, loss: 10.828, lm_loss: 10.828, perplexity: 50391.758 I0422 18:21:00.541881 140024073193280 metric_logger.py:196] completed step: 2, seconds: 0.010, TFLOP/s/device: 1328.295, Tokens/s/device: 200215.075, total_weights: 15161, loss: 9.823, lm_loss: 9.823, perplexity: 18444.867 I0422 18:21:00.659622 140024073193280 metric_logger.py:196] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.575, Tokens/s/device: 17571.405, total_weights: 13327, loss: 9.405, lm_loss: 9.405, perplexity: 12149.331 I0422 18:21:00.752651 139899646772800 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_10_shardy_false/checkpoints/0/items/array_metadatas/process_0 I0422 18:21:17.978091 140024073193280 metric_logger.py:196] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.743, Tokens/s/device: 17295.252, total_weights: 11939, loss: 9.045, lm_loss: 9.045, perplexity: 8471.842 I0422 18:21:18.007327 140024073193280 metric_logger.py:196] completed step: 5, seconds: 0.118, TFLOP/s/device: 115.588, Tokens/s/device: 17422.670, total_weights: 15502, loss: 8.953, lm_loss: 8.953, perplexity: 7728.546 I0422 18:21:18.125196 140024073193280 metric_logger.py:196] completed step: 6, seconds: 17.337, TFLOP/s/device: 0.784, Tokens/s/device: 118.128, total_weights: 13864, loss: 8.775, lm_loss: 8.775, perplexity: 6468.536 I0422 18:21:18.242682 140024073193280 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1694.156, Tokens/s/device: 255361.596, total_weights: 12988, loss: 8.688, lm_loss: 8.688, perplexity: 5928.887 I0422 18:21:18.361008 140024073193280 metric_logger.py:196] completed step: 8, seconds: 0.120, TFLOP/s/device: 113.336, Tokens/s/device: 17083.180, total_weights: 13820, loss: 8.740, lm_loss: 8.740, perplexity: 6250.123 I0422 18:21:18.478179 140024073193280 checkpointing.py:772] Waiting for step 9 to finish before checkpoint... I0422 18:21:18.479759 140024073193280 checkpointing.py:776] Waited 0.0020399093627929688 seconds for step 9 to finish before starting checkpointing. I0422 18:21:18.480271 140024073193280 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:21:47.082256 139899040695872 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 47.134608s (commit=46.646569s, array_metadata_write=0.488039s) I0422 18:21:47.083492 139899688715840 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 254.439 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 49 seconds) (per-host) I0422 18:21:47.083612 139899688715840 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 47.135072s. I0422 18:21:47.410152 139899688715840 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776882058233106702, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0422 18:21:47.636283 139899688715840 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 18:21:47.825668 139899462223424 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776882058233106702, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0422 18:21:48.053109 139899688715840 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0422 18:21:48.053282 139899688715840 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.600570s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0/items I0422 18:21:48.054061 139899688715840 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0/items I0422 18:21:48.302856 139899688715840 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0 I0422 18:21:49.011508 139899688715840 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_10_shardy_false/checkpoints/0`. I0422 18:21:49.012240 139899688715840 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 51.693843s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/0 I0422 18:21:49.012324 139899688715840 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 49.063787s. I0422 18:21:49.012596 139898847757888 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:21:49.012720 139898847757888 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:21:49.012785 139898847757888 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:21:49.012831 139898847757888 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0422 18:21:49.012985 140024073193280 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 18:21:49.013157 140024073193280 checkpoint_manager.py:1441] Waiting for previous save to complete took 30.532911 seconds. If this number is high, consider checkpointing less frequently. I0422 18:21:49.014264 140024073193280 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9 I0422 18:21:49.014601 140024073193280 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9. I0422 18:21:49.184263 139899688715840 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9 I0422 18:21:49.189401 140024073193280 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array. I0422 18:21:49.189507 140024073193280 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 18:21:49.896658 139899646772800 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9/items I0422 18:21:51.832922 140024073193280 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.643954s I0422 18:21:51.837150 140024073193280 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 4.503 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host) I0422 18:21:51.837286 140024073193280 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.740539s (batch_requests_ready=0.089117s, total_serialization_initiated=2.647256s, others=0.004165s) I0422 18:21:51.837372 140024073193280 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.741160s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=2.741141) I0422 18:21:51.839112 139898847757888 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started. I0422 18:21:51.839217 140024073193280 async_checkpointer.py:561] Finished blocking save. Time taken: 2.824908s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9. I0422 18:21:51.839419 140024073193280 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0422 18:21:51.839587 139899688715840 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0422 18:21:51.839693 140024073193280 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776882078.480206, 'wait_for_prev_duration_secs': 30.5329110622406, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776882109.0142899, 'checkpointer_blocking_duration_secs': 2.825038433074951, 'get_old_steps_start_time': 1776882111.839348, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776882078.4801402, 'checkpoint_manager_blocking_duration_secs': 33.35952854156494} I0422 18:21:51.839814 140024073193280 checkpointing.py:408] Started an asynchronous checkpoint save for step 9 I0422 18:21:51.839850 140024073193280 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0422 18:21:52.716648 139899497875008 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_10_shardy_false/checkpoints/9/items/array_metadatas/process_0 I0422 18:22:32.569525 139899646772800 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 40.731807s (commit=40.270407s, array_metadata_write=0.461401s) I0422 18:22:32.570616 139898847757888 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 290.676 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 43 seconds) (per-host) I0422 18:22:32.570733 139898847757888 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.731469s. I0422 18:22:33.034705 139898847757888 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 18:22:33.480971 139898847757888 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389). I0422 18:22:33.481238 139898847757888 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.616070s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9/items I0422 18:22:33.481972 139898847757888 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9/items I0422 18:22:33.731929 139898847757888 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9 I0422 18:22:34.466147 139898847757888 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_10_shardy_false/checkpoints/9`. I0422 18:22:34.466894 139898847757888 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 45.452591s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260422_174509/linen_main_20260422_174509_10_shardy_false/checkpoints/9 I0422 18:22:34.466973 139898847757888 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 42.627712s. I0422 18:22:34.467162 139899688715840 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0422 18:22:34.467281 139899688715840 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0422 18:22:34.467330 139899688715840 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0422 18:22:34.467367 139899688715840 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0422 18:22:34.467529 140024073193280 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 18:22:34.467673 140024073193280 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0422 18:22:34.468486 140024073193280 metric_logger.py:196] completed step: 9, seconds: 0.122, TFLOP/s/device: 111.022, Tokens/s/device: 16734.487, 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