MaxView

2026-04-16 20:21:49.956830: 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 20:21:50.072529 134996612451456 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0416 20:22:42.921523 134996612451456 max_utils.py:800] System Information: Jax Version: 0.9.2
I0416 20:22:42.921629 134996612451456 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0416 20:22:42.921665 134996612451456 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 20:22:42.921689 134996612451456 train_utils.py:364] WARNING: 'dataset_path' might be pointing your local file system
I0416 20:22:42.921712 134996612451456 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0416 20:22:42.921783 134996612451456 train.py:811] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0416 20:22:43.013474 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 20:22:43.401566 134996612451456 maxtext_utils.py:1687] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 20:22:44.449474 134996612451456 checkpointing.py:688] Setting up checkpoint logger...
I0416 20:22:44.449576 134996612451456 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0416 20:22:44.449618 134996612451456 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0416 20:22:44.449894 134996612451456 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 0x7ac6c23f4f80>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0416 20:22:46.849207 134996612451456 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0416 20:22:46.849379 134996612451456 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 0x7ac1bd648890>}, handler_registry=None
I0416 20:22:46.849596 134996612451456 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ac1bd648890>` 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 20:22:46.849641 134996612451456 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ac1bd649130>` 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 20:22:46.849670 134996612451456 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 0x7ac1bd648890>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ac1bd648890>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ac1bd649130>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ac1bd649130>}).
I0416 20:22:46.849984 134996612451456 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0416 20:22:46.850032 134996612451456 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x7ac1c40d0cc0> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0416 20:22:46.978514 134996612451456 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints
I0416 20:22:46.978729 134996612451456 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/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ac1bd64a450>
I0416 20:22:46.978805 134996612451456 checkpointing.py:302] Checkpoint manager created!
W0416 20:22:46.994854 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 20:22:47.812737 134996612451456 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0416 20:22:47.812844 134996612451456 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
W0416 20:22:48.986493 4099044 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 20:22:50.057363 134996612451456 maxtext_utils.py:1805]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.057462 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.057502 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.057533 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.057560 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.057595 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.057622 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.057651 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.057676 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.057700 134996612451456 maxtext_utils.py:1805]  decoder/layers/0/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.057723 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.057745 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.057767 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.057789 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.057809 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.057830 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.057853 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.057874 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.057895 134996612451456 maxtext_utils.py:1805]  decoder/layers/1/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.057915 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.057936 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.057958 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.057978 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.057998 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058020 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058055 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.058079 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058101 134996612451456 maxtext_utils.py:1805]  decoder/layers/2/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058122 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058143 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058165 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.058186 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058206 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058227 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058249 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.058270 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058290 134996612451456 maxtext_utils.py:1805]  decoder/layers/3/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058311 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058331 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058354 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.058374 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058395 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058416 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058437 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.058458 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058479 134996612451456 maxtext_utils.py:1805]  decoder/layers/4/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058500 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058520 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058540 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.058559 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058578 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058598 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058618 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.058642 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058663 134996612451456 maxtext_utils.py:1805]  decoder/layers/5/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058683 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058703 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058723 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.058742 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058761 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058781 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058802 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.058823 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058843 134996612451456 maxtext_utils.py:1805]  decoder/layers/6/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058863 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058883 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.058903 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.058921 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058940 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.058960 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.058981 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.059001 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059021 134996612451456 maxtext_utils.py:1805]  decoder/layers/7/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059050 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059072 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059092 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.059113 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059133 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059154 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059176 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.059197 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059217 134996612451456 maxtext_utils.py:1805]  decoder/layers/8/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059238 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059258 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059278 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.059297 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059317 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059337 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059358 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.059378 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059399 134996612451456 maxtext_utils.py:1805]  decoder/layers/9/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059418 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059438 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059458 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.059478 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059497 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059519 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059540 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.059561 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059581 134996612451456 maxtext_utils.py:1805]  decoder/layers/10/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059601 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059621 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059644 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.059664 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059683 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059703 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059724 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.059743 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059763 134996612451456 maxtext_utils.py:1805]  decoder/layers/11/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059783 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059803 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059823 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.059842 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059861 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.059881 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059901 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.059922 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059942 134996612451456 maxtext_utils.py:1805]  decoder/layers/12/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.059962 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.059985 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.060006 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.060026 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.060057 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.060078 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060099 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.060119 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060140 134996612451456 maxtext_utils.py:1805]  decoder/layers/13/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060160 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.060180 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.060200 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.060219 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.060238 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.060259 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060279 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.060299 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060319 134996612451456 maxtext_utils.py:1805]  decoder/layers/14/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060339 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.060359 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0416 20:22:50.060379 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0416 20:22:50.060399 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.060418 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 20:22:50.060438 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060458 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0416 20:22:50.060479 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060499 134996612451456 maxtext_utils.py:1805]  decoder/layers/15/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0416 20:22:50.060519 134996612451456 maxtext_utils.py:1805]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0416 20:22:50.060538 134996612451456 maxtext_utils.py:1805]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0416 20:22:50.604663 134996612451456 nnx_decoders.py:1023] nnx_decoders/y_in Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 20:22:50.604740 134996612451456 nnx_decoders.py:1023] nnx_decoders/y_in Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 20:22:50.609245 134996612451456 nnx_decoders.py:1023] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 20:22:50.609289 134996612451456 nnx_decoders.py:1023] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 20:22:50.623522 134996612451456 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 20:22:50.623571 134996612451456 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 20:22:50.637188 134996612451456 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 20:22:50.637236 134996612451456 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 20:22:50.657119 134996612451456 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 20:22:50.657174 134996612451456 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 20:22:50.670993 134996612451456 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 20:22:50.671054 134996612451456 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 20:22:50.684867 134996612451456 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 20:22:50.684914 134996612451456 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 20:22:50.708241 134996612451456 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0416 20:22:50.708297 134996612451456 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 20:22:50.726758 134996612451456 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0416 20:22:50.726810 134996612451456 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
W0416 20:22:55.543213 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 20:22:57.826252 134996612451456 max_utils.py:791] Total memory size: 3.3 GB, Output size: 1.5 GB, Temp size: 1.8 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB.
I0416 20:22:57.827087 134996612451456 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0416 20:22:57.835754 134996612451456 metric_logger.py:289] number parameters: 1.104 billion
W0416 20:23:02.871702 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 20:23:04.567346 134996612451456 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0416 20:23:04.667906 134996612451456 checkpointing.py:798] Waited 0.10054492950439453 seconds for step 0 to finish before starting checkpointing.
I0416 20:23:04.669062 134996612451456 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 20:23:04.669235 134996612451456 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0416 20:23:04.669595 134996612451456 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0.
I0416 20:23:04.767958 134996612451456 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0416 20:23:04.855103 134889267529280 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0
I0416 20:23:04.923225 134996612451456 jax_array_handlers.py:347] Scheduling D2H of 648 prioritized jax.Array.
I0416 20:23:04.923665 134996612451456 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 20:23:05.555004 134996612451456 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.635844s
I0416 20:23:05.555558 134996612451456 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 15.695 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 786 milliseconds) (per-host)
I0416 20:23:05.555627 134996612451456 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.786424s (batch_requests_ready=0.100974s, total_serialization_initiated=0.685050s, others=0.000401s)
I0416 20:23:05.555823 134996612451456 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.787392s (all_items=0.000020s, per_item={'items': '0.00001955'}, temp_paths=0.787372)
I0416 20:23:05.556682 134889196226112 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 20:23:05.556846 134996612451456 async_checkpointer.py:561] Finished blocking save. Time taken: 0.887564s. Continuing background save to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0.
I0416 20:23:05.557016 134996612451456 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 20:23:05.557248 134996612451456 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776370984.6690376, 'wait_for_prev_duration_secs': 5.340576171875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776370984.6692562, 'checkpointer_blocking_duration_secs': 0.8876662254333496, 'get_old_steps_start_time': 1776370985.5569487, 'get_old_steps_duration_secs': 2.6941299438476562e-05, 'checkpoint_manager_blocking_start_time': 1776370984.6689506, 'checkpoint_manager_blocking_duration_secs': 0.8882675170898438}
I0416 20:23:05.557510 134996612451456 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0416 20:23:05.557749 134996612451456 max_utils.py:750] 
Memstats: After params initialized:
I0416 20:23:05.558072 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_0(process=0,(0,0,0,0))
I0416 20:23:05.558109 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_1(process=0,(1,0,0,0))
I0416 20:23:05.558130 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_2(process=0,(0,1,0,0))
I0416 20:23:05.558150 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_3(process=0,(1,1,0,0))
I0416 20:23:05.558168 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_4(process=0,(0,2,0,0))
I0416 20:23:05.558186 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_5(process=0,(1,2,0,0))
I0416 20:23:05.558204 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_6(process=0,(0,3,0,0))
I0416 20:23:05.558221 134996612451456 max_utils.py:756] 	Using (GB) 2.16 / 31.25 (6.912000%) on TPU_7(process=0,(1,3,0,0))
W0416 20:23:05.564991 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 20:23:05.570676 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 20:23:05.574800 4099044 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 20:23:05.575766 134889257043520 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/items
I0416 20:23:05.742170 134889217197632 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776370985493791454, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776370985493791454, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/_CHECKPOINT_METADATA
I0416 20:23:05.742381 134889278015040 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 20:23:05.910390 134996612451456 metric_logger.py:185] completed step: 0, seconds: 6.726, TFLOP/s/device: 2.020, Tokens/s/device: 304.507, total_weights: 16384, loss: 10.877
I0416 20:23:05.911393 134996612451456 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/tensorboard/'
I0416 20:23:05.999789 134996612451456 metric_logger.py:185] completed step: 1, seconds: 1.316, TFLOP/s/device: 10.327, Tokens/s/device: 1556.583, total_weights: 16384, loss: 10.877
I0416 20:23:06.102280 134996612451456 metric_logger.py:185] completed step: 2, seconds: 0.104, TFLOP/s/device: 130.510, Tokens/s/device: 19671.879, total_weights: 16384, loss: 9.773
I0416 20:23:06.133022 4101996 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0416 20:23:06.133094 4101996 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0416 20:23:06.268606 134996612451456 metric_logger.py:185] completed step: 3, seconds: 0.032, TFLOP/s/device: 426.102, Tokens/s/device: 64226.801, total_weights: 16384, loss: 8.836
I0416 20:23:06.939387 134889246557760 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 648 array_metadata.ArrayMetadata to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/items/array_metadatas/process_0
I0416 20:23:26.606400 134996612451456 metric_logger.py:185] completed step: 4, seconds: 0.172, TFLOP/s/device: 79.130, Tokens/s/device: 11927.295, total_weights: 16384, loss: 7.959
I0416 20:23:26.635786 134996612451456 metric_logger.py:185] completed step: 5, seconds: 0.100, TFLOP/s/device: 135.385, Tokens/s/device: 20406.740, total_weights: 16384, loss: 7.220
I0416 20:23:26.726415 134996612451456 metric_logger.py:185] completed step: 6, seconds: 20.337, TFLOP/s/device: 0.668, Tokens/s/device: 100.703, total_weights: 16384, loss: 6.661
I0416 20:23:26.837011 134996612451456 metric_logger.py:185] completed step: 7, seconds: 0.024, TFLOP/s/device: 577.046, Tokens/s/device: 86978.680, total_weights: 16384, loss: 6.278
I0416 20:23:26.947936 134996612451456 metric_logger.py:185] completed step: 8, seconds: 0.092, TFLOP/s/device: 147.300, Tokens/s/device: 22202.708, total_weights: 16384, loss: 6.040
I0416 20:23:27.063060 134996612451456 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0416 20:23:27.071001 134996612451456 checkpointing.py:798] Waited 0.007970571517944336 seconds for step 9 to finish before starting checkpointing.
I0416 20:23:27.071752 134996612451456 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 20:23:49.175176 134889206711872 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 43.079488s (commit=42.589188s, array_metadata_write=0.490300s)
I0416 20:23:49.177039 134889196226112 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 284.562 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 44 seconds) (per-host)
I0416 20:23:49.177104 134889196226112 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 43.620310s.
I0416 20:23:49.413923 134889196226112 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776370985493791454, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/_CHECKPOINT_METADATA
I0416 20:23:49.591988 134889196226112 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 20:23:49.821890 134889217197632 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776370985493791454, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/_CHECKPOINT_METADATA
I0416 20:23:50.054528 134889196226112 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 20:23:50.055461 134889196226112 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.602660s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/items
I0416 20:23:50.056143 134889196226112 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0/items
I0416 20:23:50.300854 134889196226112 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0
I0416 20:23:51.018820 134889196226112 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0`.
I0416 20:23:51.019455 134889196226112 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 46.350186s. directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/0
I0416 20:23:51.019542 134889196226112 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 45.462748s.
I0416 20:23:51.019665 134889278015040 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 20:23:51.019716 134889278015040 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 20:23:51.019768 134889278015040 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0416 20:23:51.019812 134889278015040 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0416 20:23:51.019945 134996612451456 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 20:23:51.020085 134996612451456 checkpoint_manager.py:1441] Waiting for previous save to complete took 23.948339 seconds. If this number is high, consider checkpointing less frequently.
I0416 20:23:51.020866 134996612451456 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0416 20:23:51.021159 134996612451456 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9.
I0416 20:23:51.191577 134889278015040 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9
I0416 20:23:51.822039 134996612451456 jax_array_handlers.py:347] Scheduling D2H of 648 prioritized jax.Array.
I0416 20:23:51.822241 134996612451456 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 20:23:52.158644 134996612451456 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.888425s
I0416 20:23:52.159087 134996612451456 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 11.792 GiB/s (total gbytes: 12.3 GiB) (time elapsed: a second) (per-host)
I0416 20:23:52.159146 134996612451456 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 1.046617s (batch_requests_ready=0.107146s, total_serialization_initiated=0.939182s, others=0.000289s)
I0416 20:23:52.159339 134996612451456 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 1.047255s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=1.047237)
I0416 20:23:52.160825 134889147991616 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 20:23:52.160975 134996612451456 async_checkpointer.py:561] Finished blocking save. Time taken: 1.140061s. Continuing background save to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9.
I0416 20:23:52.161228 134996612451456 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 20:23:52.161495 134889137505856 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 20:23:52.161663 134996612451456 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776371007.0717137, 'wait_for_prev_duration_secs': 23.948338508605957, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776371031.0208895, 'checkpointer_blocking_duration_secs': 1.1402366161346436, 'get_old_steps_start_time': 1776371032.1611502, 'get_old_steps_duration_secs': 2.86102294921875e-05, 'checkpoint_manager_blocking_start_time': 1776371007.071667, 'checkpoint_manager_blocking_duration_secs': 25.089959621429443}
I0416 20:23:52.161887 134996612451456 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0416 20:23:52.161928 134996612451456 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 20:23:52.458105 134889257043520 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9/items
I0416 20:23:53.751298 134889206711872 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 648 array_metadata.ArrayMetadata to gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9/items/array_metadatas/process_0
I0416 20:24:26.355873 134889177351744 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 33.377661s (commit=32.891182s, array_metadata_write=0.486480s)
I0416 20:24:26.359061 134889147991616 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 358.527 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 35 seconds) (per-host)
I0416 20:24:26.359260 134889147991616 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 34.198219s.
I0416 20:24:26.780949 134889147991616 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 20:24:27.234924 134889147991616 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 20:24:27.236377 134889147991616 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.604595s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9/items
I0416 20:24:27.237376 134889147991616 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9/items
I0416 20:24:27.506012 134889147991616 atomicity.py:608] Finalizing gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9
I0416 20:24:28.221559 134889147991616 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9`.
I0416 20:24:28.222518 134889147991616 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 37.201614s. directory=gs://wanglance-maxtext/nnx_ckpt_feat_nnx_post_train_fixes_20260416_181521/nnx_feat_nnx_post_train_fixes_20260416_181521_13_scan_layers_false/checkpoints/9
I0416 20:24:28.222621 134889147991616 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 36.061583s.
I0416 20:24:28.223001 134889137505856 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 20:24:28.223194 134889137505856 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 20:24:28.223340 134889137505856 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0416 20:24:28.223408 134889137505856 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0416 20:24:28.223659 134996612451456 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 20:24:28.224132 134996612451456 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 20:24:28.225207 134996612451456 metric_logger.py:185] completed step: 9, seconds: 0.112, TFLOP/s/device: 121.837, Tokens/s/device: 18364.584, total_weights: 16384, loss: 5.900
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops