2026-04-15 21:51:50.029223: 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 21:51:50.146746 126659001232512 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0415 21:52:20.529215 126659001232512 max_utils.py:800] System Information: Jax Version: 0.9.2
I0415 21:52:20.529339 126659001232512 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0415 21:52:20.529376 126659001232512 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 21:52:20.529400 126659001232512 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system
I0415 21:52:20.529424 126659001232512 train_utils.py:347] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0415 21:52:20.529505 126659001232512 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0415 21:52:20.621786 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:52:21.038495 126659001232512 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0415 21:52:21.038778 126659001232512 checkpointing.py:677] Setting up checkpoint logger...
I0415 21:52:21.038825 126659001232512 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0415 21:52:21.038887 126659001232512 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0415 21:52:21.040222 126659001232512 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 0x733181df4ce0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0415 21:52:23.485614 126659001232512 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0415 21:52:23.485834 126659001232512 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 0x732b05b38e30>}, handler_registry=None
I0415 21:52:23.486147 126659001232512 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x732b05b38e30>` 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 21:52:23.486191 126659001232512 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x732b05b3a390>` 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 21:52:23.486221 126659001232512 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 0x732b05b38e30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x732b05b38e30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x732b05b3a390>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x732b05b3a390>}).
I0415 21:52:23.486533 126659001232512 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0415 21:52:23.486585 126659001232512 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x732b05b31260> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0415 21:52:23.615751 126659001232512 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints
I0415 21:52:23.616026 126659001232512 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_13_scan_layers_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x732b05b38c50>
I0415 21:52:23.616128 126659001232512 checkpointing.py:301] Checkpoint manager created!
W0415 21:52:23.633223 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:52:23.819408 126659001232512 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0415 21:52:23.819518 126659001232512 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:52:23.907737 126659001232512 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0415 21:52:23.907828 126659001232512 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:52:23.922132 126659001232512 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0415 21:52:23.922187 126659001232512 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:52:23.942612 126659001232512 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 21:52:23.942670 126659001232512 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:52:23.956879 126659001232512 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 21:52:23.956933 126659001232512 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:52:23.971169 126659001232512 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 21:52:23.971221 126659001232512 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:52:23.993612 126659001232512 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0415 21:52:23.993674 126659001232512 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:52:24.012064 126659001232512 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0415 21:52:24.012119 126659001232512 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0415 21:52:24.196675 126659001232512 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0415 21:52:24.196779 126659001232512 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
W0415 21:52:24.247493 3270127 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 21:52:24.629947 126659001232512 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0415 21:52:24.630086 126659001232512 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 21:52:24.630129 126659001232512 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 21:52:24.630173 126659001232512 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 21:52:24.630211 126659001232512 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 21:52:24.630238 126659001232512 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 21:52:24.630280 126659001232512 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 21:52:24.630318 126659001232512 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 21:52:24.630346 126659001232512 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 21:52:24.630372 126659001232512 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 21:52:24.630405 126659001232512 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0415 21:52:24.630437 126659001232512 maxtext_utils.py:1620] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0415 21:52:25.054664 126659001232512 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0415 21:52:25.054756 126659001232512 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0415 21:52:25.068404 126659001232512 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0415 21:52:25.068458 126659001232512 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
W0415 21:52:25.730420 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:52:25.847512 126659001232512 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 21:52:25.848156 126659001232512 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0415 21:52:25.848319 126659001232512 metric_logger.py:289] number parameters: 1.104 billion
W0415 21:52:26.698849 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:52:26.796372 126659001232512 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0415 21:52:26.913489 126659001232512 checkpointing.py:776] Waited 0.11710023880004883 seconds for step 0 to finish before starting checkpointing.
I0415 21:52:26.914024 126659001232512 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:52:26.914211 126659001232512 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0415 21:52:26.914676 126659001232512 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0.
I0415 21:52:27.008753 126659001232512 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0415 21:52:27.099996 126659001232512 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0415 21:52:27.100119 126659001232512 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 21:52:27.107009 126551565796928 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0
W0415 21:52:27.326773 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.340335 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.350009 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.362379 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.375123 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.384269 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.389715 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:27.394403 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:52:27.814918 126551555311168 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0/items
I0415 21:52:27.963563 126551523853888 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776289947717298657, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776289947717298657, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:52:28.342122 3272280 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0415 21:52:28.342189 3272280 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0415 21:52:28.574897 126659001232512 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 1.475453s
I0415 21:52:28.575466 126659001232512 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 7.885 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host)
I0415 21:52:28.575662 126659001232512 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.565346s (batch_requests_ready=0.085682s, total_serialization_initiated=1.479204s, others=0.000459s)
I0415 21:52:28.580124 126659001232512 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.570785s (all_items=0.000026s, per_item={'items': '0.00002599'}, temp_paths=1.570759)
I0415 21:52:28.581642 126551576282688 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0415 21:52:28.581755 126659001232512 async_checkpointer.py:561] Finished blocking save. Time taken: 1.667498s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0.
I0415 21:52:28.581967 126659001232512 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0415 21:52:28.582167 126551429482048 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0415 21:52:28.582241 126659001232512 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776289946.9140105, 'wait_for_prev_duration_secs': 5.888938903808594e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776289946.9142342, 'checkpointer_blocking_duration_secs': 1.667647123336792, 'get_old_steps_start_time': 1776289948.581907, 'get_old_steps_duration_secs': 2.4080276489257812e-05, 'checkpoint_manager_blocking_start_time': 1776289946.9139113, 'checkpoint_manager_blocking_duration_secs': 1.6682989597320557}
I0415 21:52:28.582365 126659001232512 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0415 21:52:28.582415 126659001232512 max_utils.py:750]
Memstats: After params initialized:
I0415 21:52:28.582475 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_0(process=0,(0,0,0,0))
I0415 21:52:28.582503 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0415 21:52:28.582524 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0415 21:52:28.582543 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0415 21:52:28.582562 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0415 21:52:28.582581 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0415 21:52:28.582598 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0415 21:52:28.582615 126659001232512 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
W0415 21:52:28.589143 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:28.595840 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:52:28.601822 3270127 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:52:28.962563 126659001232512 metric_logger.py:185] completed step: 0, seconds: 0.948, TFLOP/s/device: 14.340, Tokens/s/device: 2161.448, total_weights: 16384, loss: 10.874
I0415 21:52:28.963598 126659001232512 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/tensorboard/'
I0415 21:52:29.078765 126659001232512 metric_logger.py:185] completed step: 1, seconds: 2.164, TFLOP/s/device: 6.278, Tokens/s/device: 946.282, total_weights: 16384, loss: 10.874
I0415 21:52:29.198097 126659001232512 metric_logger.py:185] completed step: 2, seconds: 0.007, TFLOP/s/device: 1988.748, Tokens/s/device: 299765.808, total_weights: 16384, loss: 9.775
I0415 21:52:29.316506 126659001232512 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 115.920, Tokens/s/device: 17472.763, total_weights: 16384, loss: 8.838
I0415 21:52:29.550131 126551471425088 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_13_scan_layers_true/checkpoints/0/items/array_metadatas/process_0
I0415 21:52:48.687892 126659001232512 metric_logger.py:185] completed step: 4, seconds: 0.119, TFLOP/s/device: 114.013, Tokens/s/device: 17185.245, total_weights: 16384, loss: 7.965
I0415 21:52:48.696181 126659001232512 metric_logger.py:185] completed step: 5, seconds: 0.119, TFLOP/s/device: 114.644, Tokens/s/device: 17280.367, total_weights: 16384, loss: 7.233
I0415 21:52:48.813273 126659001232512 metric_logger.py:185] completed step: 6, seconds: 19.373, TFLOP/s/device: 0.701, Tokens/s/device: 105.716, total_weights: 16384, loss: 6.676
I0415 21:52:48.931365 126659001232512 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2772.318, Tokens/s/device: 417873.903, total_weights: 16384, loss: 6.293
I0415 21:52:49.049890 126659001232512 metric_logger.py:185] completed step: 8, seconds: 0.118, TFLOP/s/device: 115.218, Tokens/s/device: 17366.971, total_weights: 16384, loss: 6.055
I0415 21:52:49.167920 126659001232512 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0415 21:52:49.168792 126659001232512 checkpointing.py:776] Waited 0.0009045600891113281 seconds for step 9 to finish before starting checkpointing.
I0415 21:52:49.169103 126659001232512 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0415 21:52:56.950825 126551460939328 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 28.369903s (commit=27.945826s, array_metadata_write=0.424077s)
I0415 21:52:56.952208 126551576282688 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 422.044 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 29 seconds) (per-host)
I0415 21:52:56.952336 126551576282688 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.370532s.
I0415 21:52:57.178963 126551576282688 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776289947717298657, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:52:57.372875 126551576282688 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 21:52:57.570936 126551523853888 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776289947717298657, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:52:57.840606 126551576282688 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0415 21:52:57.841288 126551576282688 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.617160s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0/items
I0415 21:52:57.842073 126551576282688 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0/items
I0415 21:52:58.093874 126551576282688 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0
I0415 21:52:58.789854 126551576282688 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_13_scan_layers_true/checkpoints/0`.
I0415 21:52:58.790639 126551576282688 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 31.876393s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/0
I0415 21:52:58.790717 126551576282688 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.208914s.
I0415 21:52:58.790852 126551429482048 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0415 21:52:58.790904 126551429482048 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0415 21:52:58.790956 126551429482048 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0415 21:52:58.791001 126551429482048 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0415 21:52:58.791252 126659001232512 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 21:52:58.791402 126659001232512 checkpoint_manager.py:1441] Waiting for previous save to complete took 9.622299 seconds. If this number is high, consider checkpointing less frequently.
I0415 21:52:58.792269 126659001232512 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0415 21:52:58.792577 126659001232512 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9.
I0415 21:52:58.966167 126659001232512 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0415 21:52:58.966276 126659001232512 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 21:52:58.973547 126551429482048 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9
I0415 21:52:59.665065 126551439967808 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9/items
I0415 21:53:09.064813 126659001232512 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 10.099152s
I0415 21:53:09.069454 126659001232512 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.211 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 10 seconds) (per-host)
I0415 21:53:09.069523 126659001232512 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 10.189085s (batch_requests_ready=0.081868s, total_serialization_initiated=10.102729s, others=0.004489s)
I0415 21:53:09.069646 126659001232512 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 10.190012s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=10.189992)
I0415 21:53:09.071405 126551460939328 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0415 21:53:09.071495 126659001232512 async_checkpointer.py:561] Finished blocking save. Time taken: 10.279178s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9.
I0415 21:53:09.071699 126659001232512 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0415 21:53:09.071871 126551576282688 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0415 21:53:09.071983 126659001232512 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776289969.1690702, 'wait_for_prev_duration_secs': 9.622299194335938, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776289978.792296, 'checkpointer_blocking_duration_secs': 10.279306173324585, 'get_old_steps_start_time': 1776289989.0716195, 'get_old_steps_duration_secs': 3.0279159545898438e-05, 'checkpoint_manager_blocking_start_time': 1776289969.169009, 'checkpoint_manager_blocking_duration_secs': 19.902949810028076}
I0415 21:53:09.072132 126659001232512 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0415 21:53:09.072166 126659001232512 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0415 21:53:09.880570 126565843207744 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_13_scan_layers_true/checkpoints/9/items/array_metadatas/process_0
I0415 21:53:37.377095 126551439967808 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 28.306624s (commit=27.889828s, array_metadata_write=0.416796s)
I0415 21:53:37.378255 126551460939328 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 328.247 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 38 seconds) (per-host)
I0415 21:53:37.378323 126551460939328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 28.306777s.
I0415 21:53:37.809592 126551460939328 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 21:53:38.236141 126551460939328 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0415 21:53:38.236984 126551460939328 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.569815s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9/items
I0415 21:53:38.237821 126551460939328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9/items
I0415 21:53:38.476354 126551460939328 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9
I0415 21:53:39.169496 126551460939328 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_13_scan_layers_true/checkpoints/9`.
I0415 21:53:39.170313 126551460939328 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 40.377998s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_13_scan_layers_true/checkpoints/9
I0415 21:53:39.170395 126551460939328 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 30.098853s.
I0415 21:53:39.170562 126551576282688 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0415 21:53:39.170693 126551576282688 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0415 21:53:39.170790 126551576282688 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0415 21:53:39.170847 126551576282688 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0415 21:53:39.170995 126659001232512 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 21:53:39.171152 126659001232512 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:53:39.171842 126659001232512 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.826, Tokens/s/device: 17307.822, 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