2026-04-16 13:37:11.121676: 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 13:37:11.237746 129603052530816 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0416 13:38:10.711280 129603052530816 max_utils.py:800] System Information: Jax Version: 0.9.2
I0416 13:38:10.711386 129603052530816 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0416 13:38:10.711424 129603052530816 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 13:38:10.711452 129603052530816 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system
I0416 13:38:10.711475 129603052530816 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0416 13:38:10.711550 129603052530816 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0416 13:38:10.804037 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:38:11.207930 129603052530816 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 13:38:11.208263 129603052530816 checkpointing.py:677] Setting up checkpoint logger...
I0416 13:38:11.208309 129603052530816 checkpointing.py:233] Creating checkpoint manager with ocdbt=False and zarr3=True
I0416 13:38:11.208375 129603052530816 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0416 13:38:11.209007 129603052530816 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x75def99fc890>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0416 13:38:13.587078 129603052530816 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0416 13:38:13.587280 129603052530816 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 0x75d87dfcaf90>}, handler_registry=None
I0416 13:38:13.587590 129603052530816 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x75d87dfcaf90>` 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 13:38:13.587638 129603052530816 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x75d87e59edb0>` 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 13:38:13.587667 129603052530816 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 0x75d87dfcaf90>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x75d87dfcaf90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x75d87e59edb0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x75d87e59edb0>}).
I0416 13:38:13.587972 129603052530816 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0416 13:38:13.588030 129603052530816 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x75d87d131300> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0416 13:38:13.713150 129603052530816 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints
I0416 13:38:13.713376 129603052530816 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=5), 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_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x75da3e378b90>
I0416 13:38:13.713462 129603052530816 checkpointing.py:301] Checkpoint manager created!
W0416 13:38:13.729897 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:38:13.907584 129603052530816 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 13:38:13.907685 129603052530816 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 13:38:13.995687 129603052530816 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 13:38:13.995762 129603052530816 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 13:38:14.009945 129603052530816 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 13:38:14.009994 129603052530816 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 13:38:14.030377 129603052530816 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 13:38:14.030431 129603052530816 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:38:14.044631 129603052530816 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 13:38:14.044678 129603052530816 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:38:14.058769 129603052530816 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 13:38:14.058818 129603052530816 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:38:14.080252 129603052530816 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0416 13:38:14.080307 129603052530816 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 13:38:14.098478 129603052530816 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0416 13:38:14.098531 129603052530816 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0416 13:38:14.281028 129603052530816 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0416 13:38:14.281130 129603052530816 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
W0416 13:38:14.331301 3879656 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 13:38:14.711639 129603052530816 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0416 13:38:14.711746 129603052530816 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 13:38:14.711787 129603052530816 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 13:38:14.711833 129603052530816 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 13:38:14.711872 129603052530816 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 13:38:14.711906 129603052530816 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 13:38:14.711946 129603052530816 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 13:38:14.711985 129603052530816 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 13:38:14.712013 129603052530816 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 13:38:14.712039 129603052530816 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 13:38:14.712091 129603052530816 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0416 13:38:14.712126 129603052530816 maxtext_utils.py:1620] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0416 13:38:15.131833 129603052530816 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 13:38:15.131910 129603052530816 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0416 13:38:15.145618 129603052530816 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 13:38:15.145668 129603052530816 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
W0416 13:38:15.766503 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:38:15.883387 129603052530816 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 13:38:15.883859 129603052530816 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0416 13:38:15.883996 129603052530816 metric_logger.py:289] number parameters: 1.104 billion
W0416 13:38:16.717157 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:38:16.814265 129603052530816 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0416 13:38:16.931370 129603052530816 checkpointing.py:776] Waited 0.11709451675415039 seconds for step 0 to finish before starting checkpointing.
I0416 13:38:16.931811 129603052530816 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 13:38:16.931967 129603052530816 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0416 13:38:16.932333 129603052530816 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0.
I0416 13:38:17.023627 129603052530816 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0416 13:38:17.110095 129495585523264 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0
I0416 13:38:17.124111 129603052530816 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 13:38:17.124194 129603052530816 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 13:38:17.334684 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.346639 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.370607 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.383987 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.389614 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.394135 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.398354 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:17.402606 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:38:17.814264 129495575037504 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/items
I0416 13:38:17.963855 129495543580224 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776346697738822180, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776346697738822180, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0416 13:38:18.309688 3881778 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0416 13:38:18.309741 3881778 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0416 13:38:19.165092 129603052530816 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 2.041331s
I0416 13:38:19.169207 129603052530816 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 5.754 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 2 seconds) (per-host)
I0416 13:38:19.169267 129603052530816 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 2.144665s (batch_requests_ready=0.095465s, total_serialization_initiated=2.045189s, others=0.004011s)
I0416 13:38:19.169338 129603052530816 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 2.145195s (all_items=0.000021s, per_item={'items': '0.00002146'}, temp_paths=2.145173)
I0416 13:38:19.170658 129495596009024 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 13:38:19.170744 129603052530816 async_checkpointer.py:561] Finished blocking save. Time taken: 2.238726s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0.
I0416 13:38:19.170944 129603052530816 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 13:38:19.171130 129495476471360 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 13:38:19.171225 129603052530816 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776346696.9317982, 'wait_for_prev_duration_secs': 4.076957702636719e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776346696.931988, 'checkpointer_blocking_duration_secs': 2.2388527393341064, 'get_old_steps_start_time': 1776346699.170872, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776346696.9317157, 'checkpoint_manager_blocking_duration_secs': 2.239481210708618}
I0416 13:38:19.171343 129603052530816 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0416 13:38:19.171385 129603052530816 max_utils.py:750]
Memstats: After params initialized:
I0416 13:38:19.171437 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0))
I0416 13:38:19.171464 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0416 13:38:19.171483 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0416 13:38:19.171501 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0416 13:38:19.171519 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0416 13:38:19.171536 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0416 13:38:19.171554 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0416 13:38:19.171571 129603052530816 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
W0416 13:38:19.177246 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:19.185210 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 13:38:19.190038 3879656 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 13:38:19.381738 129603052530816 metric_logger.py:185] completed step: 0, seconds: 0.930, TFLOP/s/device: 14.613, Tokens/s/device: 2202.650, total_weights: 16384, loss: 10.874
I0416 13:38:19.382625 129603052530816 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/tensorboard/'
I0416 13:38:19.498257 129603052530816 metric_logger.py:185] completed step: 1, seconds: 2.566, TFLOP/s/device: 5.296, Tokens/s/device: 798.205, total_weights: 16384, loss: 10.874
I0416 13:38:19.616759 129603052530816 metric_logger.py:185] completed step: 2, seconds: 0.005, TFLOP/s/device: 2481.214, Tokens/s/device: 373995.617, total_weights: 16384, loss: 10.276
I0416 13:38:19.736000 129603052530816 metric_logger.py:185] completed step: 3, seconds: 0.118, TFLOP/s/device: 115.547, Tokens/s/device: 17416.447, total_weights: 16384, loss: 9.282
I0416 13:38:19.743150 129603052530816 checkpointing.py:772] Waiting for step 5 to finish before checkpoint...
I0416 13:38:19.951580 129495554065984 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_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_0
I0416 13:38:19.973865 129603052530816 checkpointing.py:776] Waited 0.23067593574523926 seconds for step 5 to finish before starting checkpointing.
I0416 13:38:19.974514 129603052530816 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 13:38:24.368495 129495497442880 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 5.198552s (commit=4.835838s, array_metadata_write=0.362714s)
I0416 13:38:24.369472 129495596009024 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 1.680 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host)
I0416 13:38:24.369570 129495596009024 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 5.198867s.
I0416 13:38:24.580594 129495596009024 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776346697738822180, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
W0416 13:38:24.670861 129495596009024 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/items,
I0416 13:38:24.745587 129495596009024 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 13:38:24.745691 129495596009024 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.128393s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/items
I0416 13:38:24.746341 129495596009024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/items
I0416 13:38:24.977503 129495543580224 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776346697738822180, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0/_CHECKPOINT_METADATA
I0416 13:38:24.980413 129495596009024 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0
I0416 13:38:25.633439 129495596009024 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_15_ocdbt_false/checkpoints/0`.
I0416 13:38:25.634081 129495596009024 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 8.702081s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/0
I0416 13:38:25.634157 129495596009024 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 6.463456s.
I0416 13:38:25.634332 129495476471360 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 13:38:25.634441 129495476471360 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 13:38:25.634504 129495476471360 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0416 13:38:25.634549 129495476471360 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0416 13:38:25.634738 129603052530816 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 13:38:25.634794 129603052530816 checkpoint_manager.py:1441] Waiting for previous save to complete took 5.660304 seconds. If this number is high, consider checkpointing less frequently.
I0416 13:38:25.635527 129603052530816 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 5
I0416 13:38:25.635742 129603052530816 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5.
I0416 13:38:25.809806 129603052530816 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 13:38:25.809902 129603052530816 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 13:38:25.819388 129495596009024 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5
I0416 13:38:26.484319 129495497442880 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5/items
I0416 13:38:36.696353 129603052530816 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 10.886904s
I0416 13:38:36.700538 129603052530816 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.125 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 10 seconds) (per-host)
I0416 13:38:36.700603 129603052530816 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 10.970875s (batch_requests_ready=0.076507s, total_serialization_initiated=10.890291s, others=0.004076s)
I0416 13:38:36.700685 129603052530816 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 10.971387s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=10.971370)
I0416 13:38:36.702153 129495486957120 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 13:38:36.702235 129603052530816 async_checkpointer.py:561] Finished blocking save. Time taken: 11.066666s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5.
I0416 13:38:36.702430 129603052530816 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 13:38:36.702618 129495476471360 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 13:38:36.702729 129603052530816 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776346699.9744778, 'wait_for_prev_duration_secs': 5.660303592681885, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776346705.635548, 'checkpointer_blocking_duration_secs': 11.066792488098145, 'get_old_steps_start_time': 1776346716.7023635, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1776346699.9743774, 'checkpoint_manager_blocking_duration_secs': 16.728323936462402}
I0416 13:38:36.702866 129603052530816 checkpointing.py:408] Started an asynchronous checkpoint save for step 5
I0416 13:38:36.703440 129603052530816 metric_logger.py:185] completed step: 4, seconds: 0.119, TFLOP/s/device: 114.201, Tokens/s/device: 17213.556, total_weights: 16384, loss: 8.327
I0416 13:38:36.710710 129603052530816 metric_logger.py:185] completed step: 5, seconds: 0.121, TFLOP/s/device: 112.450, Tokens/s/device: 16949.714, total_weights: 16384, loss: 7.367
I0416 13:38:36.826563 129603052530816 metric_logger.py:185] completed step: 6, seconds: 16.966, TFLOP/s/device: 0.801, Tokens/s/device: 120.712, total_weights: 16384, loss: 6.459
I0416 13:38:36.945417 129603052530816 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2522.207, Tokens/s/device: 380174.494, total_weights: 16384, loss: 5.622
I0416 13:38:37.064199 129603052530816 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 116.336, Tokens/s/device: 17535.448, total_weights: 16384, loss: 4.864
I0416 13:38:37.182154 129603052530816 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0416 13:38:37.183174 129603052530816 checkpointing.py:776] Waited 0.0010483264923095703 seconds for step 9 to finish before starting checkpointing.
I0416 13:38:37.183506 129603052530816 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 13:38:37.486652 129495554065984 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_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_0
I0416 13:38:42.319821 129495497442880 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 5.618472s (commit=5.246171s, array_metadata_write=0.372301s)
I0416 13:38:42.320868 129495486957120 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 761.660 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 16 seconds) (per-host)
I0416 13:38:42.320926 129495486957120 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 5.618647s.
W0416 13:38:42.640068 129495486957120 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5/items,
I0416 13:38:42.754498 129495486957120 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 13:38:42.754601 129495486957120 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.164805s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5/items
I0416 13:38:42.755247 129495486957120 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5/items
I0416 13:38:43.007944 129495486957120 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5
I0416 13:38:43.663333 129495486957120 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_15_ocdbt_false/checkpoints/5`.
I0416 13:38:43.663937 129495486957120 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 18.028375s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/5
I0416 13:38:43.664007 129495486957120 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 6.961729s.
I0416 13:38:43.664193 129495476471360 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 13:38:43.664292 129495476471360 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 13:38:43.664341 129495476471360 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts...
I0416 13:38:43.664385 129495476471360 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts.
I0416 13:38:43.664480 129603052530816 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5.
W0416 13:38:43.664533 129603052530816 checkpoint_manager.py:1441] Waiting for previous save to complete took 6.481046 seconds. If this number is high, consider checkpointing less frequently.
I0416 13:38:43.664721 129603052530816 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0416 13:38:43.664987 129603052530816 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9.
I0416 13:38:43.817356 129495486957120 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9
I0416 13:38:43.827903 129603052530816 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 13:38:43.827992 129603052530816 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 13:38:44.474526 129495497442880 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9/items
I0416 13:38:51.524235 129603052530816 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 7.696653s
I0416 13:38:51.528219 129603052530816 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.586 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 7 seconds) (per-host)
I0416 13:38:51.528275 129603052530816 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 7.781050s (batch_requests_ready=0.076894s, total_serialization_initiated=7.700282s, others=0.003873s)
I0416 13:38:51.528344 129603052530816 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 7.781541s (all_items=0.000010s, per_item={'items': '0.00001025'}, temp_paths=7.781531)
I0416 13:38:51.529736 129495465985600 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 13:38:51.529819 129603052530816 async_checkpointer.py:561] Finished blocking save. Time taken: 7.865058s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9.
I0416 13:38:51.529994 129603052530816 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 13:38:51.530177 129495476471360 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 13:38:51.530285 129603052530816 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776346717.1834762, 'wait_for_prev_duration_secs': 6.481045961380005, 'time_between_consecutive_saves_sec': 11.548911571502686, 'checkpointer_blocking_start_time': 1776346723.6647434, 'checkpointer_blocking_duration_secs': 7.865182876586914, 'get_old_steps_start_time': 1776346731.5299382, 'get_old_steps_duration_secs': 1.6927719116210938e-05, 'checkpoint_manager_blocking_start_time': 1776346717.1834373, 'checkpoint_manager_blocking_duration_secs': 14.346824645996094}
I0416 13:38:51.530406 129603052530816 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0416 13:38:51.530436 129603052530816 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 13:38:52.223819 129495554065984 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_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_0
I0416 13:38:57.517952 129495497442880 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 5.988919s (commit=5.600226s, array_metadata_write=0.388693s)
I0416 13:38:57.518882 129495465985600 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 917.596 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 13 seconds) (per-host)
I0416 13:38:57.518941 129495465985600 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 5.989074s.
W0416 13:38:57.826435 129495465985600 ocdbt_utils.py:175] [process=0][thread=async_save] Skipping merge of OCDBT checkpoints: No per-process OCDBT checkpoint subdirs found in gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9/items,
I0416 13:38:57.927165 129495465985600 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 13:38:57.927274 129495465985600 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.155833s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9/items
I0416 13:38:57.927913 129495465985600 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9/items
I0416 13:38:58.177200 129495465985600 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9
I0416 13:38:58.838662 129495465985600 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_15_ocdbt_false/checkpoints/9`.
I0416 13:38:58.839283 129495465985600 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 15.174529s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260416_123638/linen_main_20260416_123638_15_ocdbt_false/checkpoints/9
I0416 13:38:58.839356 129495465985600 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 7.309490s.
I0416 13:38:58.839565 129495476471360 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 13:38:58.839671 129495476471360 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 13:38:58.839716 129495476471360 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0416 13:38:58.839755 129495476471360 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0416 13:38:58.839858 129603052530816 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 13:38:58.839938 129603052530816 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 13:38:58.840609 129603052530816 metric_logger.py:185] completed step: 9, seconds: 0.123, TFLOP/s/device: 110.637, Tokens/s/device: 16676.438, total_weights: 16384, loss: 4.189
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops