MaxView

XPK Start: Fri Apr 24 14:04:46 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
`rope_parameters`'s factor field must be a float >= 1, got 40
`rope_parameters`'s beta_fast field must be a float, got 32
`rope_parameters`'s beta_slow field must be a float, got 1
DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 
2026-04-24 14:05:11.417114: 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)
I0424 14:05:11.628381 133596268832576 max_utils.py:273] Attempting to initialize the jax distributed system...
I0424 14:05:20.669641 133596268832576 distributed.py:149] Starting JAX distributed service on [::]:8482
I0424 14:05:20.671958 133596268832576 distributed.py:172] Connecting to JAX distributed service on mt-13-scan-layers-false-998a1-slice-job-0-0.mt-13-scan-layers-false-998a1:8482
I0424 14:05:22.161210 133596268832576 max_utils.py:284] Jax distributed system initialized!
I0424 14:05:28.366994 133596268832576 max_utils.py:800] System Information: Jax Version: 0.9.2
I0424 14:05:28.367100 133596268832576 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0424 14:05:28.367141 133596268832576 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365
I0424 14:05:28.367177 133596268832576 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0424 14:05:29.061616 133596268832576 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0424 14:05:29.875756 133596268832576 checkpointing.py:688] Setting up checkpoint logger...
I0424 14:05:29.875878 133596268832576 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0424 14:05:29.875925 133596268832576 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0424 14:05:29.876176 133596268832576 base_pytree_checkpoint_handler.py:441] 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 0x7980b81ba6c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0424 14:05:32.765450 133596268832576 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0424 14:05:32.765711 133596268832576 checkpoint_manager.py:708] [process=5][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796c787db740>}, handler_registry=None
I0424 14:05:32.766013 133596268832576 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796c787db740>` 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`.
I0424 14:05:32.766066 133596268832576 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x796c787dca70>` 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`.
I0424 14:05:32.766105 133596268832576 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 0x796c787db740>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x796c787db740>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x796c787dca70>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x796c787dca70>}).
I0424 14:05:32.766433 133596268832576 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0424 14:05:32.766503 133596268832576 async_checkpointer.py:192] [process=5][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x796c90533a60> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0424 14:05:33.537956 133596268832576 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints
I0424 14:05:33.557615 133596268832576 checkpoint_manager.py:929] [process=5][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_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, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x796cb02eaff0>
I0424 14:05:33.557756 133596268832576 checkpointing.py:302] Checkpoint manager created!
I0424 14:05:35.683563 133596268832576 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0424 14:05:35.683705 133596268832576 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0424 14:05:43.621544 133596268832576 maxtext_utils.py:1889]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.621695 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.621762 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.621828 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.621878 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.621927 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.621989 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622040 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.622091 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622141 133596268832576 maxtext_utils.py:1889]  decoder/layers/0/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622186 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.622236 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.622285 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.622327 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.622369 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.622421 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622464 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.622510 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622559 133596268832576 maxtext_utils.py:1889]  decoder/layers/1/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622610 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.622680 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.622736 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.622782 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.622825 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.622876 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.622919 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.622970 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623019 133596268832576 maxtext_utils.py:1889]  decoder/layers/2/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623066 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.623112 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.623160 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.623202 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.623244 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.623296 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623340 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.623385 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623432 133596268832576 maxtext_utils.py:1889]  decoder/layers/3/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623474 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.623516 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.623564 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.623604 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.623643 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.623702 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623744 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.623789 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623829 133596268832576 maxtext_utils.py:1889]  decoder/layers/4/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.623878 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.623920 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.623970 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.624015 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.624055 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.624094 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624143 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.624184 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624227 133596268832576 maxtext_utils.py:1889]  decoder/layers/5/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624275 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.624316 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.624357 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.624405 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.624445 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.624485 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624534 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.624576 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624618 133596268832576 maxtext_utils.py:1889]  decoder/layers/6/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624675 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.624719 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.624763 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.624805 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.624845 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.624887 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.624942 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.624984 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625026 133596268832576 maxtext_utils.py:1889]  decoder/layers/7/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625071 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.625112 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.625150 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.625196 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.625237 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.625276 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625324 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.625366 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625407 133596268832576 maxtext_utils.py:1889]  decoder/layers/8/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625455 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.625496 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.625537 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.625583 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.625623 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.625678 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625726 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.625768 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625809 133596268832576 maxtext_utils.py:1889]  decoder/layers/9/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.625858 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.625899 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.625946 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.625996 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.626035 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.626077 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626125 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.626168 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626213 133596268832576 maxtext_utils.py:1889]  decoder/layers/10/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626260 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.626301 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.626342 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.626387 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.626426 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.626467 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626516 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.626558 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626602 133596268832576 maxtext_utils.py:1889]  decoder/layers/11/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626661 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.626705 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.626750 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.626790 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.626832 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.626878 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.626919 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.626965 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627014 133596268832576 maxtext_utils.py:1889]  decoder/layers/12/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627059 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.627104 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.627151 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.627190 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.627228 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.627277 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627318 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.627359 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627406 133596268832576 maxtext_utils.py:1889]  decoder/layers/13/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627447 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.627487 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.627534 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.627573 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.627610 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.627673 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627719 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.627761 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627801 133596268832576 maxtext_utils.py:1889]  decoder/layers/14/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.627849 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/mlp/wi_0/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.627891 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/mlp/wi_1/kernel/value
    Shape:     float32[2048,7168]
    Physical:  ('fsdp', None)
I0424 14:05:43.627934 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/mlp/wo/kernel/value
    Shape:     float32[7168,2048]
    Physical:  (None, 'fsdp')
I0424 14:05:43.627984 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.628024 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0424 14:05:43.628065 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/self_attention/key/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.628113 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/self_attention/out/kernel/value
    Shape:     float32[16,128,2048]
    Physical:  (None, None, 'fsdp')
I0424 14:05:43.628155 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/self_attention/query/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.628196 133596268832576 maxtext_utils.py:1889]  decoder/layers/15/self_attention/value/kernel/value
    Shape:     float32[2048,16,128]
    Physical:  ('fsdp', None, None)
I0424 14:05:43.628244 133596268832576 maxtext_utils.py:1889]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0424 14:05:43.628284 133596268832576 maxtext_utils.py:1889]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0424 14:05:44.123989 133596268832576 nnx_decoders.py:1023] nnx_decoders/y_in Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0424 14:05:44.124081 133596268832576 nnx_decoders.py:1023] nnx_decoders/y_in Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 14:05:44.128676 133596268832576 nnx_decoders.py:1023] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0424 14:05:44.128731 133596268832576 nnx_decoders.py:1023] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 14:05:44.145034 133596268832576 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0424 14:05:44.145093 133596268832576 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 14:05:44.160685 133596268832576 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0424 14:05:44.160741 133596268832576 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0424 14:05:44.183260 133596268832576 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 14:05:44.183323 133596268832576 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 14:05:44.198923 133596268832576 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 14:05:44.198981 133596268832576 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 14:05:44.214460 133596268832576 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0424 14:05:44.214521 133596268832576 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 14:05:44.240826 133596268832576 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0424 14:05:44.240895 133596268832576 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0424 14:05:44.261183 133596268832576 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0424 14:05:44.261247 133596268832576 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0424 14:06:45.115980 133596268832576 max_utils.py:791] Total memory size: 1.8 GB, Output size: 0.4 GB, Temp size: 1.4 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0424 14:06:45.126858 133596268832576 metric_logger.py:301] number parameters: 1.104 billion
I0424 14:07:45.873107 133596268832576 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0424 14:07:46.089944 133596268832576 checkpointing.py:798] Waited 0.21685051918029785 seconds for step 0 to finish before starting checkpointing.
I0424 14:07:46.092962 133596268832576 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 14:07:46.094809 133596268832576 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 0
I0424 14:07:46.096243 133596268832576 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints/0.
I0424 14:07:46.494777 133596268832576 signaling_client.py:364] Using JaxDistributedSignalingClient
I0424 14:07:46.495971 133596268832576 jax_array_handlers.py:360] Scheduling D2H of 648 prioritized jax.Array.
I0424 14:07:46.496083 133596268832576 replica_slices.py:424] 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
I0424 14:07:46.797775 133596268832576 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.307660s
I0424 14:07:46.797936 133596268832576 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 4.224 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.3651745319366455 s) (per-host)
I0424 14:07:46.797998 133596268832576 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.365242s (batch_requests_ready=0.024684s, total_serialization_initiated=0.340482s, others=0.000076s)
I0424 14:07:46.798429 133596268832576 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.369698s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.369681)
I0424 14:07:46.799254 133596268832576 event_tracking.py:125] [process=5] [async] Finished blocking save in 0.70 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints/0.
I0424 14:07:46.799602 133469839316736 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 14:27:46.799565
I0424 14:07:46.853239 133596268832576 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 14:07:46.853589 133469809936128 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0424 14:07:46.853783 133596268832576 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777039666.0929446, 'wait_for_prev_duration_secs': 6.699562072753906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777039666.0948477, 'checkpointer_blocking_duration_secs': 0.7049179077148438, 'get_old_steps_start_time': 1777039666.79979, 'get_old_steps_duration_secs': 2.765655517578125e-05, 'checkpoint_manager_blocking_start_time': 1777039666.0911872, 'checkpoint_manager_blocking_duration_secs': 0.7625570297241211}
I0424 14:07:46.854019 133596268832576 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0424 14:07:46.854215 133596268832576 max_utils.py:750] 
Memstats: After params initialized:
I0424 14:07:46.854276 133596268832576 max_utils.py:756] 	Using (GB) 0.82 / 31.25 (2.624000%) on TPU_18(process=5,(2,4,0,0))
I0424 14:07:46.854311 133596268832576 max_utils.py:756] 	Using (GB) 0.82 / 31.25 (2.624000%) on TPU_19(process=5,(3,4,0,0))
I0424 14:07:46.854337 133596268832576 max_utils.py:756] 	Using (GB) 0.82 / 31.25 (2.624000%) on TPU_22(process=5,(2,5,0,0))
I0424 14:07:46.854362 133596268832576 max_utils.py:756] 	Using (GB) 0.82 / 31.25 (2.624000%) on TPU_23(process=5,(3,5,0,0))
I0424 14:07:47.247906 133596268832576 metric_logger.py:196] completed step: 0, seconds: 60.739, TFLOP/s/device: 0.224, Tokens/s/device: 33.718, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52794.566
I0424 14:07:47.365301 133596268832576 metric_logger.py:196] completed step: 1, seconds: 1.364, TFLOP/s/device: 9.961, Tokens/s/device: 1501.404, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52794.566
I0424 14:07:47.897457 133596268832576 metric_logger.py:196] completed step: 2, seconds: 0.052, TFLOP/s/device: 258.941, Tokens/s/device: 39030.340, total_weights: 65536, loss: 10.265, lm_loss: 10.265, perplexity: 28703.650
I0424 14:07:48.032881 133596268832576 metric_logger.py:196] completed step: 3, seconds: 0.407, TFLOP/s/device: 33.407, Tokens/s/device: 5035.442, total_weights: 65536, loss: 9.736, lm_loss: 9.736, perplexity: 16919.520
I0424 14:07:48.310302 133596268832576 metric_logger.py:196] completed step: 4, seconds: 0.219, TFLOP/s/device: 62.145, Tokens/s/device: 9367.210, total_weights: 65536, loss: 9.279, lm_loss: 9.279, perplexity: 10711.382
I0424 14:07:48.332317 133596268832576 metric_logger.py:196] completed step: 5, seconds: 0.136, TFLOP/s/device: 100.005, Tokens/s/device: 15073.786, total_weights: 65536, loss: 8.894, lm_loss: 8.894, perplexity: 7287.216
I0424 14:07:49.820615    2783 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0424 14:07:52.020792 133469818328832 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 648 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints/0/items/array_metadatas/process_5
I0424 14:08:11.069533 133596268832576 metric_logger.py:196] completed step: 6, seconds: 0.277, TFLOP/s/device: 49.076, Tokens/s/device: 7397.294, total_weights: 65536, loss: 8.595, lm_loss: 8.595, perplexity: 5403.792
I0424 14:08:11.204947 133596268832576 metric_logger.py:196] completed step: 7, seconds: 22.613, TFLOP/s/device: 0.601, Tokens/s/device: 90.568, total_weights: 65536, loss: 8.387, lm_loss: 8.387, perplexity: 4389.297
I0424 14:08:11.340245 133596268832576 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.538, Tokens/s/device: 13948.388, total_weights: 65536, loss: 8.258, lm_loss: 8.258, perplexity: 3856.737
I0424 14:08:11.482979 133596268832576 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0424 14:08:11.487243 133596268832576 checkpointing.py:798] Waited 0.004269838333129883 seconds for step 9 to finish before starting checkpointing.
I0424 14:08:11.490724 133596268832576 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 14:08:19.787105 133469839316736 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 47.358 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 33.3543016910553 s) (per-host)
I0424 14:08:19.787234 133469839316736 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 32.987507s.
I0424 14:08:30.757830 133469839316736 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 43.958088s.
I0424 14:08:30.758173 133469809936128 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 14:08:30.758287 133469809936128 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0424 14:08:30.758336 133469809936128 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0424 14:08:30.759923 133469809936128 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0424 14:08:30.760112 133596268832576 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0424 14:08:30.760247 133596268832576 checkpoint_manager.py:1452] Waiting for previous save to complete took 19.269517 seconds. If this number is high, consider checkpointing less frequently.
I0424 14:08:30.762816 133596268832576 checkpoint_manager.py:1512] [process=5] Saving checkpoint at step 9
I0424 14:08:30.764955 133596268832576 event_tracking.py:70] [process=5] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints/9.
I0424 14:08:31.643345 133596268832576 jax_array_handlers.py:360] Scheduling D2H of 648 prioritized jax.Array.
I0424 14:08:31.643899 133596268832576 replica_slices.py:424] 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
I0424 14:08:31.772760 133596268832576 base_pytree_checkpoint_handler.py:154] [process=5][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.135199s
I0424 14:08:31.772915 133596268832576 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/blocking_gbytes_per_sec: 2.143 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7199001312255859 s) (per-host)
I0424 14:08:31.772963 133596268832576 base_pytree_checkpoint_handler.py:768] [process=5][thread=MainThread] Initiated Pytree async_save. Time taken: 0.719956s (batch_requests_ready=0.023715s, total_serialization_initiated=0.696178s, others=0.000063s)
I0424 14:08:31.773358 133596268832576 composite_checkpoint_handler.py:715] [process=5][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.724276s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.724261)
I0424 14:08:31.774137 133596268832576 event_tracking.py:125] [process=5] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints/9.
I0424 14:08:31.774509 133469809936128 async_checkpointer.py:76] [process=5][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 14:28:31.774473
I0424 14:08:31.784292 133596268832576 checkpoint_manager.py:1560] [process=5][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0424 14:08:31.784585 133469296375552 async_checkpointer.py:280] [process=5][thread=save_finalize] Waiting for background save thread=async_save.
I0424 14:08:31.784732 133596268832576 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777039691.4906943, 'wait_for_prev_duration_secs': 19.269516706466675, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777039710.7628565, 'checkpointer_blocking_duration_secs': 1.011826515197754, 'get_old_steps_start_time': 1777039711.774711, 'get_old_steps_duration_secs': 3.4809112548828125e-05, 'checkpoint_manager_blocking_start_time': 1777039691.4882584, 'checkpoint_manager_blocking_duration_secs': 20.296440362930298}
I0424 14:08:31.785015 133596268832576 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0424 14:08:31.785063 133596268832576 checkpoint_manager.py:2020] [process=5][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0424 14:08:37.043306 133469818328832 array_metadata_store.py:203] [process=5][thread=array_type_handler] Wrote 648 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/nnx_xpk_feat_nnx_post_train_fixes_20260424_120657_13_scan_layers_false/checkpoints/9/items/array_metadatas/process_5
I0424 14:09:13.007511 133469809936128 base_pytree_checkpoint_handler.py:130] [process=5] /jax/orbax/write/gbytes_per_sec: 37.650 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.95445489883423 s) (per-host)
I0424 14:09:13.007640 133469809936128 async_checkpointer.py:90] [process=5][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.233002s.
I0424 14:09:23.359461 133469809936128 async_checkpointer.py:160] [process=5][thread=async_save] Background save thread done. Time taken: 51.584808s.
I0424 14:09:23.359745 133469296375552 async_checkpointer.py:288] [process=5][thread=save_finalize] Done with waiting for background save thread=async_save.
I0424 14:09:23.359871 133469296375552 async_checkpointer.py:298] [process=5][thread=save_finalize] No errors found in background save thread=async_save.
I0424 14:09:23.359919 133469296375552 checkpoint_manager.py:2137] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0424 14:09:23.363731 133469296375552 checkpoint_manager.py:2146] [process=5][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0424 14:09:23.363986 133596268832576 checkpoint_manager.py:2032] [process=5][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0424 14:09:23.364292 133596268832576 checkpoint_manager.py:2009] [process=5][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0424 14:09:23.365265 133596268832576 metric_logger.py:196] completed step: 9, seconds: 0.135, TFLOP/s/device: 100.575, Tokens/s/device: 15159.702, total_weights: 65536, loss: 8.182, lm_loss: 8.182, perplexity: 3574.626
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Fri Apr 24 14:09:36 UTC 2026
EXIT_CODE=0