MaxView

XPK Start: Sun Apr 19 02:00:30 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-19 02:00:55.128516: 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)
I0419 02:00:55.306768 137073140729664 max_utils.py:273] Attempting to initialize the jax distributed system...
I0419 02:01:04.349076 137073140729664 distributed.py:149] Starting JAX distributed service on [::]:8482
I0419 02:01:04.351522 137073140729664 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-bikz6-slice-job-0-0.mt-11-optimizer-offload-bikz6:8482
I0419 02:01:05.673047 137073140729664 max_utils.py:284] Jax distributed system initialized!
I0419 02:01:11.881883 137073140729664 max_utils.py:800] System Information: Jax Version: 0.9.2
I0419 02:01:11.881983 137073140729664 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0419 02:01:11.882022 137073140729664 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
I0419 02:01:11.882055 137073140729664 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0419 02:01:12.574151 137073140729664 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0419 02:01:12.688645 137073140729664 checkpointing.py:688] Setting up checkpoint logger...
I0419 02:01:12.688759 137073140729664 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0419 02:01:12.688818 137073140729664 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0419 02:01:12.689018 137073140729664 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 0x7caa1c14c0e0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0419 02:01:15.600327 137073140729664 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0419 02:01:15.600560 137073140729664 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c9610482210>}, handler_registry=None
I0419 02:01:15.600796 137073140729664 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c9610482210>` 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`.
I0419 02:01:15.600857 137073140729664 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c961049fda0>` 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`.
I0419 02:01:15.600893 137073140729664 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 0x7c9610482210>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c9610482210>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c961049fda0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c961049fda0>}).
I0419 02:01:15.601209 137073140729664 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0419 02:01:15.601279 137073140729664 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c9610418fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0419 02:01:17.994410 137073140729664 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints
I0419 02:01:18.509649 137073140729664 checkpoint_manager.py:929] [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_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_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c96281031d0>
I0419 02:01:18.509848 137073140729664 checkpointing.py:302] Checkpoint manager created!
I0419 02:01:19.461430 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to host
I0419 02:01:19.461543 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='decoder_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to host
I0419 02:01:19.461580 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_0'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461611 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_1'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461641 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wo'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461666 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='post_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to host
I0419 02:01:19.461692 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='pre_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to host
I0419 02:01:19.461718 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='key'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461741 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='out'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461764 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='query'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461788 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='value'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461823 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='logits_dense'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461848 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='token_embedder'), DictKey(key='embedding'), GetAttrKey(name='value')) to host
I0419 02:01:19.461871 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='decoder_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to host
I0419 02:01:19.461895 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_0'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461921 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_1'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461945 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wo'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.461968 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='post_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to host
I0419 02:01:19.461990 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='pre_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to host
I0419 02:01:19.462014 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='key'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.462036 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='out'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.462059 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='query'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.462084 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='value'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.462107 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='logits_dense'), DictKey(key='kernel'), GetAttrKey(name='value')) to host
I0419 02:01:19.462131 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='token_embedder'), DictKey(key='embedding'), GetAttrKey(name='value')) to host
I0419 02:01:19.462153 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to host
I0419 02:01:19.462174 137073140729664 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to host
I0419 02:01:19.463264 137073140729664 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0419 02:01:19.463316 137073140729664 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0419 02:01:20.903407 137073140729664 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0419 02:01:20.903520 137073140729664 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0419 02:01:20.903570 137073140729664 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0419 02:01:20.903611 137073140729664 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0419 02:01:20.903646 137073140729664 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0419 02:01:20.903681 137073140729664 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0419 02:01:20.903716 137073140729664 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 02:01:20.903751 137073140729664 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0419 02:01:20.903784 137073140729664 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 02:01:20.903829 137073140729664 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0419 02:01:20.903862 137073140729664 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0419 02:01:20.903894 137073140729664 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0419 02:01:21.308303 137073140729664 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 02:01:21.308397 137073140729664 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:01:21.313720 137073140729664 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0419 02:01:21.313774 137073140729664 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:01:21.329985 137073140729664 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 02:01:21.330044 137073140729664 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:01:21.345554 137073140729664 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0419 02:01:21.345611 137073140729664 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0419 02:01:21.370435 137073140729664 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 02:01:21.370500 137073140729664 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:01:21.386183 137073140729664 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 02:01:21.386241 137073140729664 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:01:21.401872 137073140729664 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0419 02:01:21.401930 137073140729664 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:01:21.431772 137073140729664 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0419 02:01:21.431859 137073140729664 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0419 02:01:21.451421 137073140729664 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0419 02:01:21.451484 137073140729664 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0419 02:01:21.609203 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to device
I0419 02:01:21.609315 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='decoder_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:21.609357 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_0'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609392 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_1'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609421 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wo'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609449 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='post_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:21.609476 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='pre_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:21.609503 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='key'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609529 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='out'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609555 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='query'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609579 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='value'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609604 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='logits_dense'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609632 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='token_embedder'), DictKey(key='embedding'), GetAttrKey(name='value')) to device
I0419 02:01:21.609657 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='decoder_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:21.609681 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_0'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609705 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_1'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609729 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wo'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609752 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='post_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:21.609774 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='pre_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:21.609797 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='key'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609851 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='out'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609879 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='query'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609904 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='value'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609930 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='logits_dense'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:21.609955 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='token_embedder'), DictKey(key='embedding'), GetAttrKey(name='value')) to device
I0419 02:01:21.609978 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to device
I0419 02:01:21.610001 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to device
I0419 02:01:32.571377 137073140729664 max_utils.py:791] Total memory size: 1.3 GB, Output size: 0.1 GB, Temp size: 1.1 GB, Argument size: 0.1 GB, Host temp size: 0.0 GB.
I0419 02:01:32.580534 137073140729664 metric_logger.py:301] number parameters: 1.104 billion
I0419 02:01:33.428549 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to device
I0419 02:01:33.428663 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='decoder_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:33.428704 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_0'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428738 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_1'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428766 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wo'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428794 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='post_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:33.428840 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='pre_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:33.428869 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='key'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428897 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='out'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428925 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='query'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428951 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='value'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.428975 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='decoder'), DictKey(key='logits_dense'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429000 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='mu'), DictKey(key='token_embedder'), DictKey(key='embedding'), GetAttrKey(name='value')) to device
I0419 02:01:33.429030 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='decoder_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:33.429055 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_0'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429079 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wi_1'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429102 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='mlp'), DictKey(key='wo'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429131 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='post_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:33.429155 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='pre_self_attention_layer_norm'), DictKey(key='scale'), GetAttrKey(name='value')) to device
I0419 02:01:33.429179 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='key'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429203 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='out'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429225 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='query'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429252 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='layers'), DictKey(key='self_attention'), DictKey(key='value'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429277 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='decoder'), DictKey(key='logits_dense'), DictKey(key='kernel'), GetAttrKey(name='value')) to device
I0419 02:01:33.429301 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='nu'), DictKey(key='token_embedder'), DictKey(key='embedding'), GetAttrKey(name='value')) to device
I0419 02:01:33.429324 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to device
I0419 02:01:33.429347 137073140729664 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to device
I0419 02:01:44.144452 137073140729664 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0419 02:01:44.312173 137073140729664 checkpointing.py:798] Waited 0.16770648956298828 seconds for step 0 to finish before starting checkpointing.
I0419 02:01:44.314724 137073140729664 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 02:01:44.316506 137073140729664 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0419 02:01:44.317852 137073140729664 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0.
I0419 02:01:45.471133 137073140729664 signaling_client.py:364] Using JaxDistributedSignalingClient
I0419 02:01:45.493348 137073140729664 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0419 02:01:45.493430 137073140729664 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
I0419 02:01:45.723474 136938447169280 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0
I0419 02:01:45.765315 137073140729664 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.273408s
I0419 02:01:45.766138 137073140729664 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.404 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28551721572875977 s) (per-host)
I0419 02:01:45.766215 137073140729664 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285614s (batch_requests_ready=0.003073s, total_serialization_initiated=0.281796s, others=0.000746s)
I0419 02:01:45.766303 137073140729664 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289562s (all_items=0.000017s, per_item={'items': '0.00001669'}, temp_paths=0.289546)
I0419 02:01:45.767116 137073140729664 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.45 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0.
I0419 02:01:45.767413 136938455561984 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:21:45.767377
I0419 02:01:45.770465 137073140729664 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 02:01:45.770728 136924429809408 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0419 02:01:45.770871 137073140729664 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776564104.3147056, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776564104.316545, 'checkpointer_blocking_duration_secs': 1.4510071277618408, 'get_old_steps_start_time': 1776564105.767578, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776564104.3125534, 'checkpoint_manager_blocking_duration_secs': 1.4582805633544922}
I0419 02:01:45.771028 137073140729664 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0419 02:01:45.771102 137073140729664 max_utils.py:750] 
Memstats: After params initialized:
I0419 02:01:45.771152 137073140729664 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_0(process=0,(0,0,0,0))
I0419 02:01:45.771185 137073140729664 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_1(process=0,(1,0,0,0))
I0419 02:01:45.771219 137073140729664 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_4(process=0,(0,1,0,0))
I0419 02:01:45.771245 137073140729664 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_5(process=0,(1,1,0,0))
I0419 02:01:46.124237 137073140729664 metric_logger.py:196] completed step: 0, seconds: 11.253, TFLOP/s/device: 1.207, Tokens/s/device: 181.991, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0419 02:01:46.127014 137073140729664 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/tensorboard/'
I0419 02:01:46.549797 137073140729664 metric_logger.py:196] completed step: 1, seconds: 1.976, TFLOP/s/device: 6.876, Tokens/s/device: 1036.473, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0419 02:01:46.707057 137073140729664 metric_logger.py:196] completed step: 2, seconds: 0.427, TFLOP/s/device: 31.847, Tokens/s/device: 4800.379, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133
I0419 02:01:46.870199 137073140729664 metric_logger.py:196] completed step: 3, seconds: 0.012, TFLOP/s/device: 1111.876, Tokens/s/device: 167594.108, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16852.779
I0419 02:01:47.540640 136938430383872 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/items
I0419 02:01:48.116869 136924413024000 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776564107288298977, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776564107288298977, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA
I0419 02:01:49.956667    2788 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0419 02:01:52.163083 136938413598464 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_0
I0419 02:02:09.910069 137073140729664 metric_logger.py:196] completed step: 4, seconds: 0.159, TFLOP/s/device: 85.516, Tokens/s/device: 12889.907, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10659.797
I0419 02:02:09.919734 137073140729664 metric_logger.py:196] completed step: 5, seconds: 0.162, TFLOP/s/device: 83.705, Tokens/s/device: 12616.897, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7244.862
I0419 02:02:10.080020 137073140729664 metric_logger.py:196] completed step: 6, seconds: 23.038, TFLOP/s/device: 0.590, Tokens/s/device: 88.897, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.227
I0419 02:02:10.242924 137073140729664 metric_logger.py:196] completed step: 7, seconds: 0.008, TFLOP/s/device: 1792.970, Tokens/s/device: 270256.004, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.796
I0419 02:02:10.405835 137073140729664 metric_logger.py:196] completed step: 8, seconds: 0.162, TFLOP/s/device: 83.920, Tokens/s/device: 12649.315, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.198
I0419 02:02:10.569743 137073140729664 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0419 02:02:10.570436 137073140729664 checkpointing.py:798] Waited 0.0007059574127197266 seconds for step 9 to finish before starting checkpointing.
I0419 02:02:10.572364 137073140729664 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 02:02:27.890957 136938405205760 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 37.942084s (commit=36.314369s, array_metadata_write=1.627715s)
I0419 02:02:27.892146 136938455561984 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.255 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.41149687767029 s) (per-host)
I0419 02:02:27.892273 136938455561984 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.124750s.
I0419 02:02:28.549373 136938455561984 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776564107288298977, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA
I0419 02:02:29.635258 136938455561984 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000450s.
I0419 02:02:29.686092 136924413024000 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776564107288298977, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/_CHECKPOINT_METADATA
I0419 02:02:29.916645 136938455561984 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0419 02:02:30.856607 136938455561984 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.184848s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/items
I0419 02:02:30.858179 136938455561984 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0/items
I0419 02:02:31.400074 136938455561984 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0
I0419 02:02:33.210457 136938455561984 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0`.
I0419 02:02:33.211269 136938455561984 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.89 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/0
I0419 02:02:33.212797 136938455561984 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.445273s.
I0419 02:02:33.212940 136924429809408 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 02:02:33.213013 136924429809408 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0419 02:02:33.213093 136924429809408 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0419 02:02:33.215503 136924429809408 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0419 02:02:33.215678 137073140729664 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0419 02:02:33.215824 137073140729664 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.643444 seconds. If this number is high, consider checkpointing less frequently.
I0419 02:02:33.217590 137073140729664 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0419 02:02:33.219620 137073140729664 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9.
I0419 02:02:34.364359 137073140729664 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0419 02:02:34.364457 137073140729664 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
I0419 02:02:34.385599 137073140729664 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.022618s
I0419 02:02:34.386404 137073140729664 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 49.413 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03122687339782715 s) (per-host)
I0419 02:02:34.386475 137073140729664 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.031314s (batch_requests_ready=0.002618s, total_serialization_initiated=0.027936s, others=0.000761s)
I0419 02:02:34.386550 137073140729664 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.035426s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.035409)
I0419 02:02:34.387229 137073140729664 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.17 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9.
I0419 02:02:34.387542 136924429809408 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 02:22:34.387509
I0419 02:02:34.397142 137073140729664 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0419 02:02:34.397412 136926501791488 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0419 02:02:34.397571 137073140729664 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776564130.5723333, 'wait_for_prev_duration_secs': 22.643444299697876, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776564153.2176287, 'checkpointer_blocking_duration_secs': 1.1700425148010254, 'get_old_steps_start_time': 1776564154.3876953, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776564130.570658, 'checkpoint_manager_blocking_duration_secs': 23.8268780708313}
I0419 02:02:34.397686 137073140729664 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0419 02:02:34.397731 137073140729664 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0419 02:02:34.587519 136938455561984 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9
I0419 02:02:36.733348 136926510184192 atomicity.py:140] Creating tmp directory gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9/items
I0419 02:02:40.455088 136938413598464 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_0
I0419 02:03:19.789395 136938405205760 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.509803s (commit=40.308951s, array_metadata_write=1.200852s)
I0419 02:03:19.791266 136924429809408 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.775 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.43605375289917 s) (per-host)
I0419 02:03:19.791390 136924429809408 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.403731s.
I0419 02:03:22.416908 136924429809408 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0419 02:03:22.640407 136924429809408 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000417s.
I0419 02:03:23.476332 136924429809408 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.911856s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9/items
I0419 02:03:23.479011 136924429809408 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9/items
I0419 02:03:24.438995 136924429809408 atomicity.py:666] Finalizing gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9
I0419 02:03:27.106727 136924429809408 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9`.
I0419 02:03:27.107591 136924429809408 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.89 seconds @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/nnx_xpk_feat_nnx_post_train_fixes_20260418_235007_11_optimizer_offload_true/checkpoints/9
I0419 02:03:27.109083 136924429809408 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.721431s.
I0419 02:03:27.109302 136926501791488 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0419 02:03:27.109419 136926501791488 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0419 02:03:27.109485 136926501791488 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0419 02:03:27.112335 136926501791488 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0419 02:03:27.112502 137073140729664 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0419 02:03:27.112666 137073140729664 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0419 02:03:27.113605 137073140729664 metric_logger.py:196] completed step: 9, seconds: 0.162, TFLOP/s/device: 83.661, Tokens/s/device: 12610.372, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3550.804
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Sun Apr 19 02:03:37 UTC 2026
EXIT_CODE=0