2026-04-15 21:43:37.302072: 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)
I0415 21:43:37.419550 136973736066176 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0415 21:44:28.397109 136973736066176 max_utils.py:800] System Information: Jax Version: 0.9.2
I0415 21:44:28.397215 136973736066176 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0415 21:44:28.397249 136973736066176 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
I0415 21:44:28.397274 136973736066176 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system
I0415 21:44:28.397296 136973736066176 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0415 21:44:28.397372 136973736066176 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0415 21:44:28.487832 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:44:28.884758 136973736066176 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0415 21:44:28.885135 136973736066176 checkpointing.py:677] Setting up checkpoint logger...
I0415 21:44:28.885192 136973736066176 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0415 21:44:28.885236 136973736066176 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0415 21:44:28.885818 136973736066176 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 0x7c9317ff9580>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0415 21:44:31.252995 136973736066176 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0415 21:44:31.253220 136973736066176 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 0x7c8c9bd39f70>}, handler_registry=None
I0415 21:44:31.253498 136973736066176 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c8c9bd39f70>` 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`.
I0415 21:44:31.253538 136973736066176 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8c9bd3a210>` 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`.
I0415 21:44:31.253569 136973736066176 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 0x7c8c9bd39f70>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c8c9bd39f70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8c9bd3a210>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8c9bd3a210>}).
I0415 21:44:31.254134 136973736066176 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0415 21:44:31.254190 136973736066176 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7c8c9bd2d120> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0415 21:44:31.368779 136973736066176 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints
I0415 21:44:31.368997 136973736066176 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_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c8c9bd38a70>
I0415 21:44:31.369107 136973736066176 checkpointing.py:301] Checkpoint manager created!
W0415 21:44:31.385411 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:44:31.564574 136973736066176 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0415 21:44:31.564678 136973736066176 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:44:31.660368 136973736066176 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0415 21:44:31.660445 136973736066176 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:44:31.674772 136973736066176 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0415 21:44:31.674823 136973736066176 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:44:31.695235 136973736066176 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0415 21:44:31.695291 136973736066176 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:44:31.709520 136973736066176 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0415 21:44:31.709568 136973736066176 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:44:31.723764 136973736066176 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0415 21:44:31.723822 136973736066176 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:44:31.748594 136973736066176 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0415 21:44:31.748659 136973736066176 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:44:31.767763 136973736066176 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0415 21:44:31.767814 136973736066176 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0415 21:44:31.950906 136973736066176 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0415 21:44:31.951012 136973736066176 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
W0415 21:44:32.001664 3261918 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
I0415 21:44:32.383689 136973736066176 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0415 21:44:32.383854 136973736066176 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)
I0415 21:44:32.383894 136973736066176 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)
I0415 21:44:32.383936 136973736066176 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')
I0415 21:44:32.383975 136973736066176 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)
I0415 21:44:32.384002 136973736066176 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)
I0415 21:44:32.384040 136973736066176 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)
I0415 21:44:32.384098 136973736066176 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')
I0415 21:44:32.384126 136973736066176 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)
I0415 21:44:32.384152 136973736066176 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)
I0415 21:44:32.384187 136973736066176 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0415 21:44:32.384220 136973736066176 maxtext_utils.py:1620] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0415 21:44:32.806510 136973736066176 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0415 21:44:32.806600 136973736066176 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0415 21:44:32.820132 136973736066176 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0415 21:44:32.820180 136973736066176 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
W0415 21:44:33.471377 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:44:33.587303 136973736066176 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.
I0415 21:44:33.587789 136973736066176 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0415 21:44:33.587934 136973736066176 metric_logger.py:289] number parameters: 1.104 billion
W0415 21:44:34.420017 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:44:34.516401 136973736066176 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0415 21:44:34.633599 136973736066176 checkpointing.py:776] Waited 0.1171884536743164 seconds for step 0 to finish before starting checkpointing.
I0415 21:44:34.634034 136973736066176 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:44:34.634217 136973736066176 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0415 21:44:34.634573 136973736066176 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0.
I0415 21:44:34.716919 136973736066176 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0415 21:44:34.807949 136973736066176 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0415 21:44:34.808040 136973736066176 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
I0415 21:44:34.819938 136866231748160 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0
W0415 21:44:35.022356 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.033329 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.047312 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.063408 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.074080 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.079785 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.084746 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.089417 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:44:35.369493 136973736066176 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.561975s
I0415 21:44:35.370183 136973736066176 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 18.924 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 652 milliseconds) (per-host)
I0415 21:44:35.370344 136973736066176 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.652451s (batch_requests_ready=0.085911s, total_serialization_initiated=0.565926s, others=0.000614s)
I0415 21:44:35.370468 136973736066176 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.653068s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.653049)
I0415 21:44:35.371325 136866168833600 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0415 21:44:35.371420 136973736066176 async_checkpointer.py:561] Finished blocking save. Time taken: 0.737163s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0.
I0415 21:44:35.371604 136973736066176 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0415 21:44:35.371783 136866242233920 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0415 21:44:35.371891 136973736066176 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776289474.6340213, 'wait_for_prev_duration_secs': 7.05718994140625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776289474.6342378, 'checkpointer_blocking_duration_secs': 0.7372837066650391, 'get_old_steps_start_time': 1776289475.371541, 'get_old_steps_duration_secs': 2.0503997802734375e-05, 'checkpoint_manager_blocking_start_time': 1776289474.6339386, 'checkpoint_manager_blocking_duration_secs': 0.7379209995269775}
I0415 21:44:35.372002 136973736066176 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0415 21:44:35.372062 136973736066176 max_utils.py:750]
Memstats: After params initialized:
I0415 21:44:35.372163 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0))
I0415 21:44:35.372193 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0415 21:44:35.372215 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0415 21:44:35.372235 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0415 21:44:35.372255 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0415 21:44:35.372274 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0415 21:44:35.372293 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0415 21:44:35.372311 136973736066176 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
W0415 21:44:35.378220 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.395109 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:44:35.400388 3261918 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:44:35.519330 136866221262400 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/items
I0415 21:44:35.693393 136866149959232 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776289475439125599, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776289475439125599, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:44:35.700194 136973736066176 metric_logger.py:185] completed step: 0, seconds: 0.928, TFLOP/s/device: 14.642, Tokens/s/device: 2206.935, total_weights: 16384, loss: 10.874
I0415 21:44:35.701203 136973736066176 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/tensorboard/'
I0415 21:44:35.815737 136973736066176 metric_logger.py:185] completed step: 1, seconds: 1.182, TFLOP/s/device: 11.498, Tokens/s/device: 1733.114, total_weights: 16384, loss: 10.874
I0415 21:44:35.934175 136973736066176 metric_logger.py:185] completed step: 2, seconds: 0.007, TFLOP/s/device: 1994.295, Tokens/s/device: 300601.791, total_weights: 16384, loss: 9.775
I0415 21:44:36.053110 136973736066176 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.478, Tokens/s/device: 17556.794, total_weights: 16384, loss: 8.838
I0415 21:44:36.211220 3264170 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0415 21:44:36.211284 3264170 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0415 21:44:37.168236 136866189805120 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/items/array_metadatas/process_0
I0415 21:44:55.591969 136973736066176 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 114.991, Tokens/s/device: 17332.724, total_weights: 16384, loss: 7.965
I0415 21:44:55.599079 136973736066176 metric_logger.py:185] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.600, Tokens/s/device: 17273.808, total_weights: 16384, loss: 7.233
I0415 21:44:55.716223 136973736066176 metric_logger.py:185] completed step: 6, seconds: 19.540, TFLOP/s/device: 0.695, Tokens/s/device: 104.812, total_weights: 16384, loss: 6.676
I0415 21:44:55.834622 136973736066176 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2893.341, Tokens/s/device: 436115.843, total_weights: 16384, loss: 6.293
I0415 21:44:55.953170 136973736066176 metric_logger.py:185] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.300, Tokens/s/device: 17379.203, total_weights: 16384, loss: 6.055
I0415 21:44:56.070961 136973736066176 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0415 21:44:56.071801 136973736066176 checkpointing.py:776] Waited 0.0008702278137207031 seconds for step 9 to finish before starting checkpointing.
I0415 21:44:56.072118 136973736066176 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0415 21:44:57.557910 136866179319360 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 21.344588s (commit=20.929612s, array_metadata_write=0.414976s)
I0415 21:44:57.558941 136866168833600 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 553.250 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 22 seconds) (per-host)
I0415 21:44:57.559001 136866168833600 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 22.187534s.
I0415 21:44:57.904943 136866168833600 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776289475439125599, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:44:58.086674 136866168833600 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0415 21:44:58.297396 136866149959232 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776289475439125599, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:44:58.519589 136866168833600 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0415 21:44:58.520181 136866168833600 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.573738s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/items
I0415 21:44:58.520837 136866168833600 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0/items
I0415 21:44:58.767242 136866168833600 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0
I0415 21:44:59.456653 136866168833600 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0`.
I0415 21:44:59.457296 136866168833600 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 24.823044s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/0
I0415 21:44:59.457387 136866168833600 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 24.085923s.
I0415 21:44:59.457581 136866242233920 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0415 21:44:59.457699 136866242233920 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0415 21:44:59.457761 136866242233920 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0415 21:44:59.457813 136866242233920 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0415 21:44:59.458108 136973736066176 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.
W0415 21:44:59.458234 136973736066176 checkpoint_manager.py:1441] Waiting for previous save to complete took 3.386147 seconds. If this number is high, consider checkpointing less frequently.
I0415 21:44:59.459442 136973736066176 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0415 21:44:59.459739 136973736066176 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9.
I0415 21:44:59.635124 136973736066176 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0415 21:44:59.635218 136973736066176 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
I0415 21:44:59.638626 136866242233920 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9
I0415 21:45:00.366011 136865961215552 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9/items
I0415 21:45:10.420326 136973736066176 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 10.785637s
I0415 21:45:10.424983 136973736066176 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.135 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 10 seconds) (per-host)
I0415 21:45:10.425137 136973736066176 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 10.873147s (batch_requests_ready=0.079844s, total_serialization_initiated=10.788670s, others=0.004633s)
I0415 21:45:10.425222 136973736066176 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 10.873696s (all_items=0.000031s, per_item={'items': '0.00003076'}, temp_paths=10.873666)
I0415 21:45:10.426632 136865961215552 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0415 21:45:10.426716 136973736066176 async_checkpointer.py:561] Finished blocking save. Time taken: 10.967226s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9.
I0415 21:45:10.426991 136973736066176 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0415 21:45:10.427233 136866189805120 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0415 21:45:10.427347 136973736066176 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776289496.0720596, 'wait_for_prev_duration_secs': 3.3861467838287354, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776289499.4594655, 'checkpointer_blocking_duration_secs': 10.967357635498047, 'get_old_steps_start_time': 1776289510.426861, 'get_old_steps_duration_secs': 6.651878356933594e-05, 'checkpoint_manager_blocking_start_time': 1776289496.0719957, 'checkpoint_manager_blocking_duration_secs': 14.355321168899536}
I0415 21:45:10.427467 136973736066176 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0415 21:45:10.427508 136973736066176 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0415 21:45:11.189035 136865636156992 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9/items/array_metadatas/process_0
I0415 21:45:44.752001 136865353041472 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 34.326468s (commit=33.860658s, array_metadata_write=0.465810s)
I0415 21:45:44.753012 136865961215552 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 279.569 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 45 seconds) (per-host)
I0415 21:45:44.753085 136865961215552 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.326323s.
I0415 21:45:45.179074 136865961215552 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0415 21:45:45.547700 136865961215552 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0415 21:45:45.547873 136865961215552 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.522210s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9/items
I0415 21:45:45.548501 136865961215552 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9/items
I0415 21:45:45.776901 136865961215552 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9
I0415 21:45:46.454747 136865961215552 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9`.
I0415 21:45:46.455383 136865961215552 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 46.995899s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_11_optimizer_offload_false/checkpoints/9
I0415 21:45:46.455466 136865961215552 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.028708s.
I0415 21:45:46.455656 136866189805120 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0415 21:45:46.455774 136866189805120 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0415 21:45:46.455858 136866189805120 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0415 21:45:46.455908 136866189805120 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0415 21:45:46.456295 136973736066176 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.
I0415 21:45:46.456429 136973736066176 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:45:46.457185 136973736066176 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.740, Tokens/s/device: 17294.814, total_weights: 16384, loss: 5.915
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops