2026-04-16 12:42:48.462623: 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 12:42:48.581197 125095393639552 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0416 12:43:42.986913 125095393639552 max_utils.py:800] System Information: Jax Version: 0.9.2
I0416 12:43:42.987016 125095393639552 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0416 12:43:42.987062 125095393639552 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 12:43:42.987087 125095393639552 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system
I0416 12:43:42.987190 125095393639552 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0416 12:43:43.079494 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 12:43:43.483944 125095393639552 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 12:43:43.484287 125095393639552 checkpointing.py:677] Setting up checkpoint logger...
I0416 12:43:43.484334 125095393639552 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0416 12:43:43.484379 125095393639552 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0416 12:43:43.484912 125095393639552 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 0x71c577fdae70>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0416 12:43:45.922400 125095393639552 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0416 12:43:45.922561 125095393639552 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 0x71bef7f45580>}, handler_registry=None
I0416 12:43:45.922823 125095393639552 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x71bef7f45580>` 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 12:43:45.922871 125095393639552 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x71bef8900500>` 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 12:43:45.922900 125095393639552 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 0x71bef7f45580>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x71bef7f45580>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x71bef8900500>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x71bef8900500>}).
I0416 12:43:45.923376 125095393639552 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0416 12:43:45.923426 125095393639552 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x71bef7b3d120> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0416 12:43:46.070688 125095393639552 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints
I0416 12:43:46.070915 125095393639552 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_03_dropout/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x71bef7b39580>
I0416 12:43:46.070996 125095393639552 checkpointing.py:301] Checkpoint manager created!
I0416 12:43:46.152319 125095393639552 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0
I0416 12:43:46.155347 125095393639552 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 12:43:46.209161 125095393639552 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 12:43:46.241035 125095393639552 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2
I0416 12:43:46.241117 125095393639552 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2
I0416 12:43:47.224156 125095393639552 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 12:43:47.224256 125095393639552 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None).
I0416 12:43:47.326002 125095393639552 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 12:43:47.326100 125095393639552 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None).
I0416 12:43:47.341272 125095393639552 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,128,2048]........................................ ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 12:43:47.341321 125095393639552 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,128,2048]........................................ ('fsdp', None, None).
I0416 12:43:47.362989 125095393639552 attentions.py:1154] attentions/query Logical: bfloat16[8,128,16,128]...................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 12:43:47.363056 125095393639552 attentions.py:1154] attentions/query Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None).
I0416 12:43:47.378203 125095393639552 attentions.py:1155] attentions/key Logical: bfloat16[8,128,16,128]...................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 12:43:47.378254 125095393639552 attentions.py:1155] attentions/key Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None).
I0416 12:43:47.393477 125095393639552 attentions.py:1156] attentions/value Logical: bfloat16[8,128,16,128]...................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 12:43:47.393534 125095393639552 attentions.py:1156] attentions/value Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None).
I0416 12:43:47.415975 125095393639552 attentions.py:1197] attentions/out Logical: bfloat16[8,128,16,128]...................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0416 12:43:47.416030 125095393639552 attentions.py:1197] attentions/out Physical: bfloat16[8,128,16,128]...................................... ('fsdp', None, None, None).
I0416 12:43:47.439894 125095393639552 linears.py:525] linears/x Logical: bfloat16[8,128,7168]........................................ ('activation_batch', 'activation_length', 'activation_mlp').
I0416 12:43:47.439950 125095393639552 linears.py:525] linears/x Physical: bfloat16[8,128,7168]........................................ ('fsdp', None, None).
I0416 12:43:47.638872 125095393639552 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0416 12:43:47.638964 125095393639552 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
W0416 12:43:47.688322 3825161 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 12:43:48.096627 125095393639552 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0416 12:43:48.096735 125095393639552 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 12:43:48.096777 125095393639552 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 12:43:48.096824 125095393639552 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 12:43:48.096869 125095393639552 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 12:43:48.096898 125095393639552 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 12:43:48.096937 125095393639552 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 12:43:48.096976 125095393639552 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 12:43:48.097003 125095393639552 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 12:43:48.097028 125095393639552 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 12:43:48.097081 125095393639552 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0416 12:43:48.097119 125095393639552 maxtext_utils.py:1620] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0416 12:43:48.560857 125095393639552 train.py:155] train/xent Logical: float32[8,128].............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 12:43:48.560936 125095393639552 train.py:155] train/xent Physical: float32[8,128].............................................. ('fsdp', None).
I0416 12:43:48.575601 125095393639552 train.py:162] train/z_loss Logical: float32[8,128].............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 12:43:48.575652 125095393639552 train.py:162] train/z_loss Physical: float32[8,128].............................................. ('fsdp', None).
W0416 12:43:48.813593 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 12:43:48.906359 125095393639552 max_utils.py:791] Total memory size: 2.5 GB, Output size: 1.5 GB, Temp size: 1.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB.
I0416 12:43:48.906822 125095393639552 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0416 12:43:48.906973 125095393639552 metric_logger.py:289] number parameters: 1.104 billion
W0416 12:43:51.937166 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 12:43:52.032569 125095393639552 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0416 12:43:52.097895 125095393639552 checkpointing.py:776] Waited 0.06534457206726074 seconds for step 0 to finish before starting checkpointing.
I0416 12:43:52.098359 125095393639552 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 12:43:52.098511 125095393639552 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0416 12:43:52.098930 125095393639552 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0.
I0416 12:43:52.184059 125095393639552 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0416 12:43:52.269224 125095393639552 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 12:43:52.269310 125095393639552 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 12:43:52.279496 124982134834752 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0
W0416 12:43:52.485753 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.499231 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.509867 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.525461 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.531031 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.536182 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.540930 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:52.545098 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 12:43:53.030867 124982124348992 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0/items
I0416 12:43:53.172105 124982084503104 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776343432929224244, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776343432929224244, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0/_CHECKPOINT_METADATA
I0416 12:43:53.574064 3827739 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0416 12:43:53.574126 3827739 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0416 12:43:53.876904 125095393639552 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.608125s
I0416 12:43:53.881909 125095393639552 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 7.273 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host)
I0416 12:43:53.882130 125095393639552 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.697122s (batch_requests_ready=0.080401s, total_serialization_initiated=1.611756s, others=0.004964s)
I0416 12:43:53.882249 125095393639552 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.697728s (all_items=0.000022s, per_item={'items': '0.00002217'}, temp_paths=1.697706)
I0416 12:43:53.883522 124982000617024 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 12:43:53.883629 125095393639552 async_checkpointer.py:561] Finished blocking save. Time taken: 1.785066s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0.
I0416 12:43:53.883876 125095393639552 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 12:43:53.884078 124982145320512 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 12:43:53.884178 125095393639552 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776343432.0983472, 'wait_for_prev_duration_secs': 4.172325134277344e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776343432.0985324, 'checkpointer_blocking_duration_secs': 1.7852530479431152, 'get_old_steps_start_time': 1776343433.8838089, 'get_old_steps_duration_secs': 2.3365020751953125e-05, 'checkpoint_manager_blocking_start_time': 1776343432.0982637, 'checkpoint_manager_blocking_duration_secs': 1.785886287689209}
I0416 12:43:53.884302 125095393639552 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0416 12:43:53.884354 125095393639552 max_utils.py:750]
Memstats: After params initialized:
I0416 12:43:53.884405 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_0(process=0,(0,0,0,0))
I0416 12:43:53.884433 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_1(process=0,(1,0,0,0))
I0416 12:43:53.884455 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_2(process=0,(0,1,0,0))
I0416 12:43:53.884475 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_3(process=0,(1,1,0,0))
I0416 12:43:53.884494 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_4(process=0,(0,2,0,0))
I0416 12:43:53.884513 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_5(process=0,(1,2,0,0))
I0416 12:43:53.884531 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_6(process=0,(0,3,0,0))
I0416 12:43:53.884548 125095393639552 max_utils.py:756] Using (GB) 1.56 / 31.25 (4.992000%) on TPU_7(process=0,(1,3,0,0))
W0416 12:43:53.890629 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:53.896949 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 12:43:53.901493 3825161 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 12:43:54.207491 125095393639552 metric_logger.py:185] completed step: 0, seconds: 3.125, TFLOP/s/device: 0.256, Tokens/s/device: 40.959, total_weights: 963, loss: 10.834
I0416 12:43:54.208513 125095393639552 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/tensorboard/'
I0416 12:43:54.269143 125095393639552 metric_logger.py:185] completed step: 1, seconds: 2.173, TFLOP/s/device: 0.369, Tokens/s/device: 58.902, total_weights: 945, loss: 10.747
I0416 12:43:54.332777 125095393639552 metric_logger.py:185] completed step: 2, seconds: 0.010, TFLOP/s/device: 84.117, Tokens/s/device: 13443.966, total_weights: 962, loss: 10.246
I0416 12:43:54.396447 125095393639552 metric_logger.py:185] completed step: 3, seconds: 0.062, TFLOP/s/device: 12.863, Tokens/s/device: 2055.763, total_weights: 975, loss: 9.829
I0416 12:43:54.656019 124982032074304 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_03_dropout/checkpoints/0/items/array_metadatas/process_0
I0416 12:44:05.761804 125095393639552 metric_logger.py:185] completed step: 4, seconds: 0.064, TFLOP/s/device: 12.447, Tokens/s/device: 1989.400, total_weights: 962, loss: 9.393
I0416 12:44:05.770419 125095393639552 metric_logger.py:185] completed step: 5, seconds: 0.063, TFLOP/s/device: 12.679, Tokens/s/device: 2026.407, total_weights: 958, loss: 9.248
I0416 12:44:05.832836 125095393639552 metric_logger.py:185] completed step: 6, seconds: 11.365, TFLOP/s/device: 0.070, Tokens/s/device: 11.263, total_weights: 930, loss: 9.424
I0416 12:44:05.896393 125095393639552 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 129.949, Tokens/s/device: 20769.106, total_weights: 935, loss: 9.317
I0416 12:44:05.960084 125095393639552 metric_logger.py:185] completed step: 8, seconds: 0.063, TFLOP/s/device: 12.733, Tokens/s/device: 2034.976, total_weights: 930, loss: 9.478
I0416 12:44:06.023369 125095393639552 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0416 12:44:06.024146 125095393639552 checkpointing.py:776] Waited 0.0007994174957275391 seconds for step 9 to finish before starting checkpointing.
I0416 12:44:06.024401 125095393639552 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 12:44:22.070522 124982021588544 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 28.187692s (commit=27.704563s, array_metadata_write=0.483130s)
I0416 12:44:22.071537 124982000617024 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 422.826 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 29 seconds) (per-host)
I0416 12:44:22.071638 124982000617024 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.187949s.
I0416 12:44:22.304759 124982000617024 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776343432929224244, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0/_CHECKPOINT_METADATA
I0416 12:44:22.480138 124982000617024 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 12:44:22.701001 124982084503104 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776343432929224244, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0/_CHECKPOINT_METADATA
I0416 12:44:22.986749 124982000617024 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 12:44:22.987364 124982000617024 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.642575s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0/items
I0416 12:44:22.988056 124982000617024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0/items
I0416 12:44:23.234717 124982000617024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0
I0416 12:44:23.900987 124982000617024 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_03_dropout/checkpoints/0`.
I0416 12:44:23.901627 124982000617024 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 31.803082s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/0
I0416 12:44:23.901704 124982000617024 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.018016s.
I0416 12:44:23.901856 124982145320512 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 12:44:23.901957 124982145320512 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 12:44:23.902016 124982145320512 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0416 12:44:23.902072 124982145320512 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0416 12:44:23.902303 125095393639552 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 12:44:23.902434 125095393639552 checkpoint_manager.py:1441] Waiting for previous save to complete took 17.878032 seconds. If this number is high, consider checkpointing less frequently.
I0416 12:44:23.903136 125095393639552 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0416 12:44:23.903426 125095393639552 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9.
I0416 12:44:24.078088 125095393639552 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 12:44:24.078180 125095393639552 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 12:44:24.084832 124982145320512 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9
I0416 12:44:24.826518 124981990131264 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9/items
I0416 12:44:36.400427 125095393639552 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 12.322748s
I0416 12:44:36.405228 125095393639552 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1018.378 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 12 seconds) (per-host)
I0416 12:44:36.405296 125095393639552 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 12.408839s (batch_requests_ready=0.078426s, total_serialization_initiated=12.325722s, others=0.004691s)
I0416 12:44:36.405381 125095393639552 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 12.409372s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=12.409354)
I0416 12:44:36.406746 124982000617024 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 12:44:36.406948 125095393639552 async_checkpointer.py:561] Finished blocking save. Time taken: 12.503761s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9.
I0416 12:44:36.407157 125095393639552 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 12:44:36.407317 124982145320512 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 12:44:36.407391 125095393639552 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776343446.0243683, 'wait_for_prev_duration_secs': 17.878032207489014, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776343463.9031649, 'checkpointer_blocking_duration_secs': 12.503855466842651, 'get_old_steps_start_time': 1776343476.4070594, 'get_old_steps_duration_secs': 4.792213439941406e-05, 'checkpoint_manager_blocking_start_time': 1776343446.024328, 'checkpoint_manager_blocking_duration_secs': 30.383037328720093}
I0416 12:44:36.407507 125095393639552 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0416 12:44:36.407539 125095393639552 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 12:44:37.154189 124993669170752 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_03_dropout/checkpoints/9/items/array_metadatas/process_0
I0416 12:45:12.072816 124981990131264 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 35.666857s (commit=35.158389s, array_metadata_write=0.508468s)
I0416 12:45:12.073836 124982000617024 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 262.843 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 48 seconds) (per-host)
I0416 12:45:12.073946 124982000617024 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 35.667035s.
I0416 12:45:12.510541 124982000617024 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 12:45:12.970765 124982000617024 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 12:45:12.970938 124982000617024 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.606250s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9/items
I0416 12:45:12.971598 124982000617024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9/items
I0416 12:45:13.217735 124982000617024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9
I0416 12:45:13.985649 124982000617024 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_03_dropout/checkpoints/9`.
I0416 12:45:13.986269 124982000617024 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 50.083087s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_03_dropout/checkpoints/9
I0416 12:45:13.986342 124982000617024 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 37.579437s.
I0416 12:45:13.986445 124982145320512 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 12:45:13.986498 124982145320512 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 12:45:13.986562 124982145320512 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0416 12:45:13.986616 124982145320512 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0416 12:45:13.986875 125095393639552 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 12:45:13.987009 125095393639552 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 12:45:13.987718 125095393639552 metric_logger.py:185] completed step: 9, seconds: 0.064, TFLOP/s/device: 12.575, Tokens/s/device: 2009.861, total_weights: 948, loss: 9.141
Per train step:
Total TFLOPs: 0.80
split as 99.60% learnable weight flops and 0.40% attention flops