2026-04-15 21:19:23.460028: 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 21:19:23.576275 140567111658624 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0415 21:20:13.341711 140567111658624 max_utils.py:800] System Information: Jax Version: 0.9.2
I0415 21:20:13.341815 140567111658624 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0415 21:20:13.341850 140567111658624 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 21:20:13.341874 140567111658624 train_utils.py:334] WARNING: 'dataset_path' might be pointing your local file system
I0415 21:20:13.341954 140567111658624 train.py:683] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0415 21:20:13.432300 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:20:13.818286 140567111658624 maxtext_utils.py:1517] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0415 21:20:13.818794 140567111658624 checkpointing.py:677] Setting up checkpoint logger...
I0415 21:20:13.818841 140567111658624 checkpointing.py:233] Creating checkpoint manager with ocdbt=True and zarr3=True
I0415 21:20:13.818882 140567111658624 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0415 21:20:13.819347 140567111658624 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 0x7fd7bdd03b00>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0415 21:20:16.224536 140567111658624 checkpointing.py:265] Enabling policy for fixed interval checkpointing.
I0415 21:20:16.224712 140567111658624 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 0x7fd141741ca0>}, handler_registry=None
I0415 21:20:16.224980 140567111658624 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fd141741ca0>` 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 21:20:16.225021 140567111658624 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd1417415e0>` 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 21:20:16.225065 140567111658624 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 0x7fd141741ca0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7fd141741ca0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd1417415e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7fd1417415e0>}).
I0415 21:20:16.225502 140567111658624 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0415 21:20:16.225554 140567111658624 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7fd141738fe0> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0415 21:20:16.355392 140567111658624 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints
I0415 21:20:16.355606 140567111658624 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_04_int8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7fd141f67aa0>
I0415 21:20:16.355685 140567111658624 checkpointing.py:301] Checkpoint manager created!
I0415 21:20:16.436588 140567111658624 dataset_info.py:707] Load dataset info from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0
I0415 21:20:16.439590 140567111658624 reader.py:262] Creating a tf.data.Dataset reading 8 files located in folders: tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0.
I0415 21:20:16.496881 140567111658624 logging_logger.py:49] Constructing tf.data.Dataset __local_c4_builder for split train, from tests/assets/local_datasets/c4_en_dataset_minimal/c4/en/3.1.0
I0415 21:20:16.529308 140567111658624 tokenizer.py:245] Tokenizer path: src/maxtext/assets/tokenizers/tokenizer.llama2
I0415 21:20:16.529376 140567111658624 tokenizer.py:187] Loading sentencepiece tokenizer: src/maxtext/assets/tokenizers/tokenizer.llama2
I0415 21:20:17.693848 140567111658624 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0415 21:20:17.693964 140567111658624 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:20:17.795881 140567111658624 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0415 21:20:17.795957 140567111658624 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:20:17.810866 140567111658624 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0415 21:20:17.810912 140567111658624 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0415 21:20:17.868117 140567111658624 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 21:20:17.868187 140567111658624 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:20:17.883157 140567111658624 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 21:20:17.883207 140567111658624 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:20:17.898005 140567111658624 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 21:20:17.898064 140567111658624 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:20:17.920208 140567111658624 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0415 21:20:17.920272 140567111658624 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0415 21:20:17.977596 140567111658624 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0415 21:20:17.977664 140567111658624 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0415 21:20:18.414474 140567111658624 checkpointing.py:577] checkpoint manager exists so trying to load this run's existing checkpoint
I0415 21:20:18.414571 140567111658624 checkpointing.py:665] No existing checkpoints found, not restoring checkpoint.
[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 21:20:20.325982 140567111658624 maxtext_utils.py:1620] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0415 21:20:20.326112 140567111658624 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 21:20:20.326156 140567111658624 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 21:20:20.326200 140567111658624 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 21:20:20.326240 140567111658624 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 21:20:20.326272 140567111658624 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 21:20:20.326313 140567111658624 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 21:20:20.326352 140567111658624 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 21:20:20.326380 140567111658624 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 21:20:20.326405 140567111658624 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 21:20:20.326440 140567111658624 maxtext_utils.py:1620] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0415 21:20:20.326475 140567111658624 maxtext_utils.py:1620] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0415 21:20:21.950152 140567111658624 train.py:155] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0415 21:20:21.950235 140567111658624 train.py:155] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0415 21:20:21.964718 140567111658624 train.py:162] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0415 21:20:21.964767 140567111658624 train.py:162] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
W0415 21:20:22.515529 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:20:22.702652 140567111658624 max_utils.py:791] Total memory size: 3.7 GB, Output size: 1.5 GB, Temp size: 2.2 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB.
I0415 21:20:22.703147 140567111658624 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0415 21:20:22.703291 140567111658624 metric_logger.py:289] number parameters: 1.104 billion
W0415 21:20:27.118180 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:20:27.345499 140567111658624 checkpointing.py:772] Waiting for step 0 to finish before checkpoint...
I0415 21:20:27.470565 140567111658624 checkpointing.py:776] Waited 0.12508869171142578 seconds for step 0 to finish before starting checkpointing.
I0415 21:20:27.471031 140567111658624 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:20:27.471193 140567111658624 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0415 21:20:27.471605 140567111658624 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0.
I0415 21:20:27.564116 140567111658624 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0415 21:20:27.650429 140451770271296 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0
I0415 21:20:27.660104 140567111658624 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0415 21:20:27.660195 140567111658624 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
W0415 21:20:27.891020 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.896190 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.906628 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.914018 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.931669 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.939654 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.943764 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:27.947771 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:20:28.161863 140567111658624 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.502207s
I0415 21:20:28.162227 140567111658624 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 20.676 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 596 milliseconds) (per-host)
I0415 21:20:28.162277 140567111658624 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.596939s (batch_requests_ready=0.090932s, total_serialization_initiated=0.505751s, others=0.000256s)
I0415 21:20:28.162343 140567111658624 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.597770s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.597751)
I0415 21:20:28.163103 140451698968128 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0415 21:20:28.163202 140567111658624 async_checkpointer.py:561] Finished blocking save. Time taken: 0.691970s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0.
I0415 21:20:28.163370 140567111658624 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0415 21:20:28.163524 140451780757056 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0415 21:20:28.163633 140567111658624 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776288027.4710162, 'wait_for_prev_duration_secs': 5.745887756347656e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776288027.4712143, 'checkpointer_blocking_duration_secs': 0.6920862197875977, 'get_old_steps_start_time': 1776288028.163317, 'get_old_steps_duration_secs': 2.0265579223632812e-05, 'checkpoint_manager_blocking_start_time': 1776288027.4709291, 'checkpoint_manager_blocking_duration_secs': 0.692669153213501}
I0415 21:20:28.163743 140567111658624 checkpointing.py:408] Started an asynchronous checkpoint save for step 0
I0415 21:20:28.163780 140567111658624 max_utils.py:750]
Memstats: After params initialized:
I0415 21:20:28.163851 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0))
I0415 21:20:28.163881 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_1(process=0,(1,0,0,0))
I0415 21:20:28.163900 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_2(process=0,(0,1,0,0))
I0415 21:20:28.163918 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_3(process=0,(1,1,0,0))
I0415 21:20:28.163936 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_4(process=0,(0,2,0,0))
I0415 21:20:28.163952 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_5(process=0,(1,2,0,0))
I0415 21:20:28.163968 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_6(process=0,(0,3,0,0))
I0415 21:20:28.163983 140567111658624 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_7(process=0,(1,3,0,0))
W0415 21:20:28.169907 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:28.175361 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0415 21:20:28.180003 3233185 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0415 21:20:28.505894 140567111658624 metric_logger.py:185] completed step: 0, seconds: 4.642, TFLOP/s/device: 2.927, Tokens/s/device: 441.218, total_weights: 13328, loss: 10.840
I0415 21:20:28.506880 140567111658624 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/tensorboard/'
I0415 21:20:28.557367 140451680093760 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776288028298421601, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776288028298421601, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:20:28.601930 140451759785536 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0/items
I0415 21:20:28.625802 140567111658624 metric_logger.py:185] completed step: 1, seconds: 1.159, TFLOP/s/device: 11.728, Tokens/s/device: 1767.745, total_weights: 12332, loss: 10.869
I0415 21:20:28.747005 140567111658624 metric_logger.py:185] completed step: 2, seconds: 0.010, TFLOP/s/device: 1418.281, Tokens/s/device: 213778.706, total_weights: 15161, loss: 9.874
I0415 21:20:28.868788 140567111658624 metric_logger.py:185] completed step: 3, seconds: 0.121, TFLOP/s/device: 112.432, Tokens/s/device: 16947.049, total_weights: 13327, loss: 9.406
I0415 21:20:29.095238 3236260 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0415 21:20:29.095307 3236260 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0415 21:20:29.912597 140451719939648 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_04_int8/checkpoints/0/items/array_metadatas/process_0
I0415 21:20:53.642946 140451709453888 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 24.552466s (commit=24.089194s, array_metadata_write=0.463272s)
I0415 21:20:53.644113 140451698968128 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 484.564 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 26 seconds) (per-host)
I0415 21:20:53.644220 140451698968128 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 25.480973s.
I0415 21:20:53.876201 140451698968128 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776288028298421601, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:20:54.070098 140451698968128 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 21:20:54.273706 140451680093760 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776288028298421601, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0/_CHECKPOINT_METADATA
I0415 21:20:54.500434 140451698968128 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0415 21:20:54.501111 140451698968128 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.587910s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0/items
I0415 21:20:54.501800 140451698968128 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0/items
I0415 21:20:54.751806 140451698968128 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0
I0415 21:20:55.441278 140451698968128 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_04_int8/checkpoints/0`.
I0415 21:20:55.441912 140451698968128 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 27.970686s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/0
I0415 21:20:55.441989 140451698968128 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 27.278744s.
I0415 21:20:55.442167 140451780757056 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0415 21:20:55.442270 140451780757056 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0415 21:20:55.442349 140451780757056 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0415 21:20:55.442399 140451780757056 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0415 21:20:59.638146 140567111658624 metric_logger.py:185] completed step: 4, seconds: 0.122, TFLOP/s/device: 111.546, Tokens/s/device: 16813.484, total_weights: 11939, loss: 9.149
I0415 21:20:59.648005 140567111658624 metric_logger.py:185] completed step: 5, seconds: 0.122, TFLOP/s/device: 111.478, Tokens/s/device: 16803.137, total_weights: 15502, loss: 9.058
I0415 21:20:59.768813 140567111658624 metric_logger.py:185] completed step: 6, seconds: 30.769, TFLOP/s/device: 0.442, Tokens/s/device: 66.561, total_weights: 13864, loss: 8.892
I0415 21:20:59.890767 140567111658624 metric_logger.py:185] completed step: 7, seconds: 0.006, TFLOP/s/device: 2093.548, Tokens/s/device: 315562.404, total_weights: 12988, loss: 8.751
I0415 21:21:00.012733 140567111658624 metric_logger.py:185] completed step: 8, seconds: 0.121, TFLOP/s/device: 111.859, Tokens/s/device: 16860.546, total_weights: 13820, loss: 8.791
I0415 21:21:00.134685 140567111658624 checkpointing.py:772] Waiting for step 9 to finish before checkpoint...
I0415 21:21:00.135524 140567111658624 checkpointing.py:776] Waited 0.0008721351623535156 seconds for step 9 to finish before starting checkpointing.
I0415 21:21:00.135794 140567111658624 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:21:00.135956 140567111658624 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0415 21:21:00.136258 140567111658624 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9.
I0415 21:21:00.303715 140567111658624 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0415 21:21:00.303809 140567111658624 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 21:21:00.325228 140451780757056 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9
I0415 21:21:01.027913 140451709453888 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9/items
I0415 21:21:08.856260 140567111658624 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 8.552938s
I0415 21:21:08.860502 140567111658624 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.428 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 8 seconds) (per-host)
I0415 21:21:08.860563 140567111658624 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 8.643495s (batch_requests_ready=0.083264s, total_serialization_initiated=8.556096s, others=0.004136s)
I0415 21:21:08.860634 140567111658624 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 8.644011s (all_items=0.000019s, per_item={'items': '0.00001931'}, temp_paths=8.643992)
I0415 21:21:08.862678 140451780757056 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0415 21:21:08.862775 140567111658624 async_checkpointer.py:561] Finished blocking save. Time taken: 8.726779s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9.
I0415 21:21:08.862976 140567111658624 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0415 21:21:08.863146 140451698968128 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0415 21:21:08.863248 140567111658624 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776288060.1357532, 'wait_for_prev_duration_secs': 6.651878356933594e-05, 'time_between_consecutive_saves_sec': 4.693339586257935, 'checkpointer_blocking_start_time': 1776288060.135976, 'checkpointer_blocking_duration_secs': 8.726906776428223, 'get_old_steps_start_time': 1776288068.862904, 'get_old_steps_duration_secs': 3.457069396972656e-05, 'checkpoint_manager_blocking_start_time': 1776288060.1357198, 'checkpoint_manager_blocking_duration_secs': 8.727500915527344}
I0415 21:21:08.863375 140567111658624 checkpointing.py:408] Started an asynchronous checkpoint save for step 9
I0415 21:21:08.863409 140567111658624 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0415 21:21:09.600545 140451749299776 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_04_int8/checkpoints/9/items/array_metadatas/process_0
I0415 21:21:43.030289 140451709453888 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 34.169179s (commit=33.685112s, array_metadata_write=0.484067s)
I0415 21:21:43.031239 140451780757056 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 295.154 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 42 seconds) (per-host)
I0415 21:21:43.031295 140451780757056 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.168477s.
I0415 21:21:43.496249 140451780757056 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 21:21:43.949459 140451780757056 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0415 21:21:43.950174 140451780757056 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.603911s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9/items
I0415 21:21:43.950829 140451780757056 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9/items
I0415 21:21:44.179616 140451780757056 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9
I0415 21:21:44.894329 140451780757056 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_04_int8/checkpoints/9`.
I0415 21:21:44.894968 140451780757056 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 44.758977s. directory=gs://wanglance-maxtext/linen_ckpt_main_20260415_211126/linen_main_20260415_211126_04_int8/checkpoints/9
I0415 21:21:44.895066 140451780757056 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.032249s.
I0415 21:21:44.895256 140451698968128 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0415 21:21:44.895373 140451698968128 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0415 21:21:44.895431 140451698968128 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0415 21:21:44.895482 140451698968128 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0415 21:21:44.895960 140567111658624 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 21:21:44.896106 140567111658624 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0415 21:21:44.896680 140567111658624 metric_logger.py:185] completed step: 9, seconds: 0.122, TFLOP/s/device: 111.655, Tokens/s/device: 16829.787, total_weights: 12300, loss: 8.774
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops