2026-04-16 13:05:48.513693: 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)
I0416 13:05:48.632629 133502278925440 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0416 13:06:44.521720 133502278925440 max_utils.py:800] System Information: Jax Version: 0.9.2
I0416 13:06:44.521827 133502278925440 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0416 13:06:44.521867 133502278925440 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Apr 6 2026 20:48:10 (1775533690) cl/895581894
I0416 13:06:44.521893 133502278925440 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system
I0416 13:06:44.521988 133502278925440 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0416 13:06:44.615676 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:06:45.012237 133502278925440 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 13:06:45.012554 133502278925440 checkpointing.py:677] Setting up checkpoint logger...
I0416 13:06:45.012603 133502278925440 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0416 13:06:45.012645 133502278925440 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0416 13:06:45.013328 133502278925440 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 0x796ad53f4d70>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0416 13:06:47.443987 133502278925440 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0416 13:06:47.444186 133502278925440 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 0x796458938e90>}, handler_registry=None
I0416 13:06:47.444469 133502278925440 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796458938e90>` 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`.
I0416 13:06:47.444509 133502278925440 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79645893a4b0>` 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`.
I0416 13:06:47.444539 133502278925440 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 0x796458938e90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796458938e90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79645893a4b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x79645893a4b0>}).
I0416 13:06:47.444855 133502278925440 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0416 13:06:47.444904 133502278925440 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x796458929300> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0416 13:06:47.573791 133502278925440 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints
I0416 13:06:47.574005 133502278925440 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_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x796458938c50>
I0416 13:06:47.574101 133502278925440 checkpointing.py:301] Checkpoint manager created!
I0416 13:06:47.654273 133502278925440 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0
I0416 13:06:47.657286 133502278925440 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.
I0416 13:06:47.712542 133502278925440 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
I0416 13:06:47.743615 133502278925440 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2
I0416 13:06:47.743683 133502278925440 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2
I0416 13:06:48.720248 133502278925440 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 13:06:48.720351 133502278925440 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 13:06:48.821976 133502278925440 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 13:06:48.822070 133502278925440 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 13:06:48.836904 133502278925440 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 13:06:48.836951 133502278925440 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 13:06:48.858959 133502278925440 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 13:06:48.859015 133502278925440 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:06:48.873906 133502278925440 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 13:06:48.873955 133502278925440 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:06:48.888798 133502278925440 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 13:06:48.888850 133502278925440 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:06:48.910978 133502278925440 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0416 13:06:48.911034 133502278925440 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:06:48.930095 133502278925440 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0416 13:06:48.930149 133502278925440 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0416 13:06:49.118155 133502278925440 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0416 13:06:49.118246 133502278925440 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
W0416 13:06:49.168014 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
[DECOUPLED NO-OP] gcs_storage: using stubs.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] workload_monitor: using stub.
[DECOUPLED NO-OP] vertex_tensorboard: using stub.
fsdp: 8
I0416 13:06:49.563828 133502278925440 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0416 13:06:49.563945 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0416 13:06:49.563988 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0416 13:06:49.564034 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0416 13:06:49.564095 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0416 13:06:49.564126 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0416 13:06:49.564166 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/key/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0416 13:06:49.564205 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/out/kernel
Shape: float32[16,16,128,2048]
Logical: P('heads', 'layers', 'kv', 'embed')
Physical: (None, None, None, 'fsdp')
I0416 13:06:49.564233 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/query/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'q_heads', 'kv')
Physical: ('fsdp', None, None, None)
I0416 13:06:49.564259 133502278925440 maxtext_utils.py:1620] params/params/decoder/layers/self_attention/value/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0416 13:06:49.564295 133502278925440 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0416 13:06:49.564329 133502278925440 maxtext_utils.py:1620] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0416 13:06:49.996889 133502278925440 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 13:06:49.996970 133502278925440 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0416 13:06:50.011382 133502278925440 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 13:06:50.011432 133502278925440 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
W0416 13:06:50.280536 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:06:50.393143 133502278925440 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.
I0416 13:06:50.393622 133502278925440 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0416 13:06:50.393763 133502278925440 metric_logger.py:289] number parameters: 1.104 billion
W0416 13:06:53.431317 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:06:53.561986 133502278925440 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0416 13:06:53.681849 133502278925440 checkpointing.py:776] Waited 0.11985230445861816 seconds for step 0 to finish before starting checkpointing.
I0416 13:06:53.682281 133502278925440 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 13:06:53.682431 133502278925440 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0416 13:06:53.682802 133502278925440 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0.
I0416 13:06:53.766287 133502278925440 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0416 13:06:53.857273 133389271172672 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0
I0416 13:06:53.868915 133502278925440 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 13:06:53.869004 133502278925440 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
W0416 13:06:54.086154 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.097770 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.111491 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.124360 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.136036 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.142486 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.147709 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:54.152675 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:06:54.574155 133389260686912 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/items
I0416 13:06:54.724962 133389220841024 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776344814485728781, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776344814485728781, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA
I0416 13:06:54.998008 133502278925440 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.129497s
I0416 13:06:54.998718 133502278925440 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 10.022 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host)
I0416 13:06:54.998878 133502278925440 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.231650s (batch_requests_ready=0.097981s, total_serialization_initiated=1.133053s, others=0.000616s)
I0416 13:06:54.999001 133502278925440 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.232262s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=1.232243)
I0416 13:06:54.999857 133389189383744 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 13:06:54.999952 133502278925440 async_checkpointer.py:561] Finished blocking save. Time taken: 1.317482s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0.
I0416 13:06:55.000153 133502278925440 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 13:06:55.000344 133389281658432 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 13:06:55.000450 133502278925440 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776344813.6822681, 'wait_for_prev_duration_secs': 4.267692565917969e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776344813.6824527, 'checkpointer_blocking_duration_secs': 1.3176233768463135, 'get_old_steps_start_time': 1776344815.0000963, 'get_old_steps_duration_secs': 2.1219253540039062e-05, 'checkpoint_manager_blocking_start_time': 1776344813.6821861, 'checkpoint_manager_blocking_duration_secs': 1.3182342052459717}
I0416 13:06:55.000562 133502278925440 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0416 13:06:55.000608 133502278925440 max_utils.py:750]
Memstats: After params initialized:
I0416 13:06:55.000708 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0))
I0416 13:06:55.000741 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0416 13:06:55.000762 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0416 13:06:55.000781 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0416 13:06:55.000800 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0416 13:06:55.000818 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0416 13:06:55.000836 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0416 13:06:55.000853 133502278925440 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
W0416 13:06:55.006860 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:55.012365 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:06:55.017404 3850305 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:06:55.062721 3852900 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0416 13:06:55.062778 3852900 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0416 13:06:55.342450 133502278925440 metric_logger.py:185] completed step: 0, seconds: 3.168, TFLOP/s/device: 4.289, Tokens/s/device: 646.519, total_weights: 13328, loss: 10.828
I0416 13:06:55.343444 133502278925440 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/tensorboard/'
I0416 13:06:55.458507 133502278925440 metric_logger.py:185] completed step: 1, seconds: 1.779, TFLOP/s/device: 7.640, Tokens/s/device: 1151.511, total_weights: 12332, loss: 10.828
I0416 13:06:55.575700 133502278925440 metric_logger.py:185] completed step: 2, seconds: 0.010, TFLOP/s/device: 1405.081, Tokens/s/device: 211789.038, total_weights: 15161, loss: 9.823
I0416 13:06:55.693277 133502278925440 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.389, Tokens/s/device: 17543.409, total_weights: 13327, loss: 9.405
I0416 13:06:56.018661 133389210355264 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/items/array_metadatas/process_0
I0416 13:07:15.056173 133502278925440 metric_logger.py:185] completed step: 4, seconds: 0.117, TFLOP/s/device: 115.925, Tokens/s/device: 17473.508, total_weights: 11939, loss: 9.045
I0416 13:07:15.065631 133502278925440 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 114.856, Tokens/s/device: 17312.358, total_weights: 15502, loss: 8.953
I0416 13:07:15.182278 133502278925440 metric_logger.py:185] completed step: 6, seconds: 19.362, TFLOP/s/device: 0.702, Tokens/s/device: 105.773, total_weights: 13864, loss: 8.775
I0416 13:07:15.299591 133502278925440 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 2209.289, Tokens/s/device: 333008.130, total_weights: 12988, loss: 8.688
I0416 13:07:15.416740 133502278925440 metric_logger.py:185] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.601, Tokens/s/device: 17424.597, total_weights: 13820, loss: 8.740
I0416 13:07:15.534167 133502278925440 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0416 13:07:15.534949 133502278925440 checkpointing.py:776] Waited 0.0008101463317871094 seconds for step 9 to finish before starting checkpointing.
I0416 13:07:15.535241 133502278925440 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 13:07:19.989470 133389199869504 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 24.925149s (commit=24.490476s, array_metadata_write=0.434673s)
I0416 13:07:19.990629 133389189383744 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 481.891 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 26 seconds) (per-host)
I0416 13:07:19.990737 133389189383744 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 24.990739s.
I0416 13:07:20.202871 133389189383744 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776344814485728781, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA
I0416 13:07:20.384705 133389189383744 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0416 13:07:20.578417 133389220841024 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776344814485728781, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/_CHECKPOINT_METADATA
I0416 13:07:20.811292 133389189383744 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 13:07:20.811914 133389189383744 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.571213s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/items
I0416 13:07:20.812552 133389189383744 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0/items
I0416 13:07:21.053229 133389189383744 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0
I0416 13:07:21.742538 133389189383744 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0`.
I0416 13:07:21.743252 133389189383744 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 28.060787s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/0
I0416 13:07:21.743326 133389189383744 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 26.743331s.
I0416 13:07:21.743503 133389281658432 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 13:07:21.743605 133389281658432 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 13:07:21.743661 133389281658432 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0416 13:07:21.743706 133389281658432 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0416 13:07:21.743880 133502278925440 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.
W0416 13:07:21.743999 133502278925440 checkpoint_manager.py:1441] Waiting for previous save to complete took 6.208769 seconds. If this number is high, consider checkpointing less frequently.
I0416 13:07:21.744818 133502278925440 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0416 13:07:21.745117 133502278925440 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9.
I0416 13:07:21.912671 133389281658432 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9
I0416 13:07:21.939198 133502278925440 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 13:07:21.939291 133502278925440 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
I0416 13:07:22.659606 133389118080576 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9/items
I0416 13:07:29.478377 133502278925440 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 7.539595s
I0416 13:07:29.482982 133502278925440 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.613 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host)
I0416 13:07:29.483041 133502278925440 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 7.653009s (batch_requests_ready=0.105835s, total_serialization_initiated=7.542675s, others=0.004499s)
I0416 13:07:29.483123 133502278925440 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 7.653506s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=7.653489)
I0416 13:07:29.484481 133389281658432 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 13:07:29.484573 133502278925440 async_checkpointer.py:561] Finished blocking save. Time taken: 7.739707s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9.
I0416 13:07:29.484825 133502278925440 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 13:07:29.485069 133389139052096 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 13:07:29.485183 133502278925440 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776344835.5352004, 'wait_for_prev_duration_secs': 6.208768844604492, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776344841.744845, 'checkpointer_blocking_duration_secs': 7.739897966384888, 'get_old_steps_start_time': 1776344849.484759, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1776344835.5351524, 'checkpoint_manager_blocking_duration_secs': 13.950005292892456}
I0416 13:07:29.485320 133502278925440 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0416 13:07:29.485353 133502278925440 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 13:07:30.270450 133389210355264 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9/items/array_metadatas/process_0
I0416 13:08:15.212220 133389118080576 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 45.728527s (commit=45.255713s, array_metadata_write=0.472815s)
I0416 13:08:15.213447 133389281658432 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 236.718 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 53 seconds) (per-host)
I0416 13:08:15.213553 133389281658432 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.728937s.
I0416 13:08:15.678603 133389281658432 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0416 13:08:16.093484 133389281658432 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 13:08:16.094169 133389281658432 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.576437s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9/items
I0416 13:08:16.094826 133389281658432 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9/items
I0416 13:08:16.322118 133389281658432 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9
I0416 13:08:17.033178 133389281658432 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9`.
I0416 13:08:17.033812 133389281658432 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 55.288954s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_10_shardy_true/checkpoints/9
I0416 13:08:17.033886 133389281658432 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 47.549272s.
I0416 13:08:17.034078 133389139052096 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 13:08:17.034181 133389139052096 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 13:08:17.034241 133389139052096 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0416 13:08:17.034285 133389139052096 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0416 13:08:17.034430 133502278925440 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.
I0416 13:08:17.034570 133502278925440 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 13:08:17.035882 133502278925440 metric_logger.py:185] completed step: 9, seconds: 0.117, TFLOP/s/device: 116.051, Tokens/s/device: 17492.462, total_weights: 12300, loss: 8.710
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops