MaxView

XPK Start: Wed Apr 22 14:51:09 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-22 14:51:33.591947: 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)
I0422 14:51:34.122375 139880708081472 max_utils.py:273] Attempting to initialize the jax distributed system...
I0422 14:51:43.164482 139880708081472 distributed.py:149] Starting JAX distributed service on [::]:8482
I0422 14:51:43.167000 139880708081472 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-59r9h-slice-job-0-0.mt-11-optimizer-offload-59r9h:8482
I0422 14:51:44.664214 139880708081472 max_utils.py:284] Jax distributed system initialized!
I0422 14:51:49.871687 139880708081472 max_utils.py:800] System Information: Jax Version: 0.9.2
I0422 14:51:49.871790 139880708081472 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0422 14:51:49.871829 139880708081472 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
I0422 14:51:49.871865 139880708081472 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0422 14:51:50.568124 139880708081472 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0422 14:51:50.682220 139880708081472 checkpointing.py:688] Setting up checkpoint logger...
I0422 14:51:50.682343 139880708081472 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0422 14:51:50.682391 139880708081472 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0422 14:51:50.682590 139880708081472 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 0x7f37f2718110>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0422 14:51:53.559974 139880708081472 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0422 14:51:53.560218 139880708081472 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f235c7303b0>}, handler_registry=None
I0422 14:51:53.560458 139880708081472 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f235c7303b0>` 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`.
I0422 14:51:53.560512 139880708081472 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f235c11f380>` 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`.
I0422 14:51:53.560548 139880708081472 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 0x7f235c7303b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f235c7303b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f235c11f380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f235c11f380>}).
I0422 14:51:53.560870 139880708081472 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0422 14:51:53.560940 139880708081472 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f235c5759e0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0422 14:51:54.981805 139880708081472 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints
I0422 14:51:55.367759 139880708081472 checkpoint_manager.py:929] [process=6][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_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f235c5cf9e0>
I0422 14:51:55.367924 139880708081472 checkpointing.py:302] Checkpoint manager created!
I0422 14:51:56.337167 139880708081472 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to host
I0422 14:51:56.337277 139880708081472 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
I0422 14:51:56.337315 139880708081472 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
I0422 14:51:56.337346 139880708081472 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
I0422 14:51:56.337377 139880708081472 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
I0422 14:51:56.337406 139880708081472 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
I0422 14:51:56.337434 139880708081472 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
I0422 14:51:56.337459 139880708081472 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
I0422 14:51:56.337492 139880708081472 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
I0422 14:51:56.337517 139880708081472 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
I0422 14:51:56.337541 139880708081472 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
I0422 14:51:56.337562 139880708081472 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
I0422 14:51:56.337585 139880708081472 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
I0422 14:51:56.337608 139880708081472 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
I0422 14:51:56.337631 139880708081472 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
I0422 14:51:56.337654 139880708081472 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
I0422 14:51:56.337676 139880708081472 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
I0422 14:51:56.337698 139880708081472 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
I0422 14:51:56.337721 139880708081472 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
I0422 14:51:56.337743 139880708081472 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
I0422 14:51:56.337765 139880708081472 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
I0422 14:51:56.337787 139880708081472 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
I0422 14:51:56.337810 139880708081472 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
I0422 14:51:56.337832 139880708081472 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
I0422 14:51:56.337856 139880708081472 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
I0422 14:51:56.337878 139880708081472 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to host
I0422 14:51:56.337902 139880708081472 maxtext_utils_nnx.py:115] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to host
I0422 14:51:56.338992 139880708081472 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0422 14:51:56.339046 139880708081472 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0422 14:51:57.782452 139880708081472 maxtext_utils.py:1836]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0422 14:51:57.782573 139880708081472 maxtext_utils.py:1836]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 14:51:57.782624 139880708081472 maxtext_utils.py:1836]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0422 14:51:57.782666 139880708081472 maxtext_utils.py:1836]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0422 14:51:57.782704 139880708081472 maxtext_utils.py:1836]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 14:51:57.782741 139880708081472 maxtext_utils.py:1836]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0422 14:51:57.782779 139880708081472 maxtext_utils.py:1836]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 14:51:57.782814 139880708081472 maxtext_utils.py:1836]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0422 14:51:57.782848 139880708081472 maxtext_utils.py:1836]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 14:51:57.782881 139880708081472 maxtext_utils.py:1836]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0422 14:51:57.782912 139880708081472 maxtext_utils.py:1836]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0422 14:51:57.782944 139880708081472 maxtext_utils.py:1836]  token_embedder/embedding/value
    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')

I0422 14:51:58.187523 139880708081472 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 14:51:58.187613 139880708081472 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:51:58.192989 139880708081472 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0422 14:51:58.193046 139880708081472 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:51:58.209472 139880708081472 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 14:51:58.209536 139880708081472 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:51:58.225342 139880708081472 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0422 14:51:58.225401 139880708081472 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0422 14:51:58.249416 139880708081472 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 14:51:58.249484 139880708081472 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:51:58.271537 139880708081472 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 14:51:58.271639 139880708081472 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:51:58.287724 139880708081472 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0422 14:51:58.287785 139880708081472 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:51:58.318382 139880708081472 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0422 14:51:58.318453 139880708081472 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0422 14:51:58.338326 139880708081472 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0422 14:51:58.338393 139880708081472 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0422 14:51:58.500967 139880708081472 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to device
I0422 14:51:58.501085 139880708081472 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
I0422 14:51:58.501140 139880708081472 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
I0422 14:51:58.501177 139880708081472 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
I0422 14:51:58.501208 139880708081472 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
I0422 14:51:58.501237 139880708081472 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
I0422 14:51:58.501265 139880708081472 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
I0422 14:51:58.501291 139880708081472 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
I0422 14:51:58.501319 139880708081472 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
I0422 14:51:58.501347 139880708081472 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
I0422 14:51:58.501372 139880708081472 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
I0422 14:51:58.501404 139880708081472 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
I0422 14:51:58.501431 139880708081472 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
I0422 14:51:58.501458 139880708081472 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
I0422 14:51:58.501488 139880708081472 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
I0422 14:51:58.501512 139880708081472 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
I0422 14:51:58.501535 139880708081472 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
I0422 14:51:58.501559 139880708081472 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
I0422 14:51:58.501582 139880708081472 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
I0422 14:51:58.501604 139880708081472 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
I0422 14:51:58.501631 139880708081472 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
I0422 14:51:58.501654 139880708081472 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
I0422 14:51:58.501677 139880708081472 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
I0422 14:51:58.501700 139880708081472 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
I0422 14:51:58.501723 139880708081472 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
I0422 14:51:58.501746 139880708081472 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to device
I0422 14:51:58.501769 139880708081472 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to device
I0422 14:52:09.784076 139880708081472 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.
I0422 14:52:09.786306 139880708081472 metric_logger.py:301] number parameters: 1.104 billion
I0422 14:52:10.330348 139880708081472 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to device
I0422 14:52:10.330467 139880708081472 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
I0422 14:52:10.330507 139880708081472 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
I0422 14:52:10.330538 139880708081472 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
I0422 14:52:10.330568 139880708081472 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
I0422 14:52:10.330601 139880708081472 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
I0422 14:52:10.330626 139880708081472 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
I0422 14:52:10.330653 139880708081472 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
I0422 14:52:10.330677 139880708081472 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
I0422 14:52:10.330702 139880708081472 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
I0422 14:52:10.330729 139880708081472 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
I0422 14:52:10.330755 139880708081472 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
I0422 14:52:10.330781 139880708081472 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
I0422 14:52:10.330812 139880708081472 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
I0422 14:52:10.330837 139880708081472 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
I0422 14:52:10.330861 139880708081472 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
I0422 14:52:10.330885 139880708081472 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
I0422 14:52:10.330908 139880708081472 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
I0422 14:52:10.330931 139880708081472 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
I0422 14:52:10.330954 139880708081472 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
I0422 14:52:10.330976 139880708081472 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
I0422 14:52:10.331000 139880708081472 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
I0422 14:52:10.331025 139880708081472 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
I0422 14:52:10.331048 139880708081472 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
I0422 14:52:10.331072 139880708081472 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
I0422 14:52:10.331105 139880708081472 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to device
I0422 14:52:10.331129 139880708081472 maxtext_utils_nnx.py:132] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to device
I0422 14:52:21.174149 139880708081472 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0422 14:52:21.341956 139880708081472 checkpointing.py:798] Waited 0.16784906387329102 seconds for step 0 to finish before starting checkpointing.
I0422 14:52:21.344588 139880708081472 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 14:52:21.346286 139880708081472 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0422 14:52:21.347709 139880708081472 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints/0.
I0422 14:52:22.082754 139880708081472 signaling_client.py:364] Using JaxDistributedSignalingClient
I0422 14:52:22.083747 139880708081472 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 14:52:22.083803 139880708081472 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
I0422 14:52:22.365963 139880708081472 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.283999s
I0422 14:52:22.366143 139880708081472 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.245 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.294069766998291 s) (per-host)
I0422 14:52:22.366197 139880708081472 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.294134s (batch_requests_ready=0.003231s, total_serialization_initiated=0.290832s, others=0.000072s)
I0422 14:52:22.366291 139880708081472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.298249s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.298230)
I0422 14:52:22.367117 139880708081472 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.02 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints/0.
I0422 14:52:22.367467 139749268367104 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:12:22.367426
I0422 14:52:22.369447 139880708081472 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 14:52:22.369693 139748315285248 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 14:52:22.369819 139880708081472 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776869541.3445709, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776869541.3463247, 'checkpointer_blocking_duration_secs': 1.0212507247924805, 'get_old_steps_start_time': 1776869542.3676019, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776869541.342367, 'checkpoint_manager_blocking_duration_secs': 1.0274169445037842}
I0422 14:52:22.369982 139880708081472 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0422 14:52:22.370051 139880708081472 max_utils.py:750] 
Memstats: After params initialized:
I0422 14:52:22.370110 139880708081472 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_24(process=6,(0,6,0,0))
I0422 14:52:22.370145 139880708081472 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_25(process=6,(1,6,0,0))
I0422 14:52:22.370172 139880708081472 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_28(process=6,(0,7,0,0))
I0422 14:52:22.370195 139880708081472 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_29(process=6,(1,7,0,0))
I0422 14:52:22.691396 139880708081472 metric_logger.py:196] completed step: 0, seconds: 11.387, TFLOP/s/device: 1.193, Tokens/s/device: 179.860, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0422 14:52:22.891059 139880708081472 metric_logger.py:196] completed step: 1, seconds: 1.515, TFLOP/s/device: 8.966, Tokens/s/device: 1351.437, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469
I0422 14:52:23.340628 139880708081472 metric_logger.py:196] completed step: 2, seconds: 0.038, TFLOP/s/device: 357.518, Tokens/s/device: 53889.064, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133
I0422 14:52:23.503329 139880708081472 metric_logger.py:196] completed step: 3, seconds: 0.439, TFLOP/s/device: 30.921, Tokens/s/device: 4660.795, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16852.779
I0422 14:52:23.829778 139880708081472 metric_logger.py:196] completed step: 4, seconds: 0.180, TFLOP/s/device: 75.531, Tokens/s/device: 11384.798, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10659.797
I0422 14:52:23.837334 139880708081472 metric_logger.py:196] completed step: 5, seconds: 0.162, TFLOP/s/device: 83.735, Tokens/s/device: 12621.485, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7244.862
I0422 14:52:25.661171    2730 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0422 14:52:27.718331 139748332070656 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_6
I0422 14:52:46.480363 139880708081472 metric_logger.py:196] completed step: 6, seconds: 0.327, TFLOP/s/device: 41.534, Tokens/s/device: 6260.470, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.227
I0422 14:52:46.643279 139880708081472 metric_logger.py:196] completed step: 7, seconds: 22.481, TFLOP/s/device: 0.604, Tokens/s/device: 91.098, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.796
I0422 14:52:46.806033 139880708081472 metric_logger.py:196] completed step: 8, seconds: 0.169, TFLOP/s/device: 80.521, Tokens/s/device: 12136.944, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.198
I0422 14:52:46.969431 139880708081472 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0422 14:52:46.970111 139880708081472 checkpointing.py:798] Waited 0.0006778240203857422 seconds for step 9 to finish before starting checkpointing.
I0422 14:52:46.972709 139880708081472 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 14:53:00.137386 139749268367104 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 41.495 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.065272092819214 s) (per-host)
I0422 14:53:00.137518 139749268367104 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.769432s.
I0422 14:53:11.174397 139749268367104 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 48.806293s.
I0422 14:53:11.174675 139748315285248 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 14:53:11.174799 139748315285248 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 14:53:11.174850 139748315285248 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0422 14:53:11.176597 139748315285248 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0422 14:53:11.176762 139880708081472 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0422 14:53:11.176911 139880708081472 checkpoint_manager.py:1452] Waiting for previous save to complete took 24.204196 seconds. If this number is high, consider checkpointing less frequently.
I0422 14:53:11.178922 139880708081472 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0422 14:53:11.180994 139880708081472 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints/9.
I0422 14:53:11.894529 139880708081472 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0422 14:53:11.894621 139880708081472 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
I0422 14:53:11.927700 139880708081472 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034830s
I0422 14:53:11.927864 139880708081472 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 36.725 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.042002201080322266 s) (per-host)
I0422 14:53:11.927913 139880708081472 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042060s (batch_requests_ready=0.002886s, total_serialization_initiated=0.039109s, others=0.000065s)
I0422 14:53:11.928007 139880708081472 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.046317s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.046301)
I0422 14:53:11.928709 139880708081472 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints/9.
I0422 14:53:11.929054 139748315285248 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-22 15:13:11.929017
I0422 14:53:11.932396 139880708081472 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0422 14:53:11.932677 139726015162112 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0422 14:53:11.932834 139880708081472 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776869566.9726799, 'wait_for_prev_duration_secs': 24.204195976257324, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776869591.1789603, 'checkpointer_blocking_duration_secs': 0.7502641677856445, 'get_old_steps_start_time': 1776869591.9292483, 'get_old_steps_duration_secs': 3.0040740966796875e-05, 'checkpoint_manager_blocking_start_time': 1776869566.9703405, 'checkpoint_manager_blocking_duration_secs': 24.962460041046143}
I0422 14:53:11.932943 139880708081472 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0422 14:53:11.932985 139880708081472 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0422 14:53:17.908994 139748332070656 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260422_123906/nnx_xpk_feat_nnx_post_train_fixes_20260422_123906_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_6
I0422 14:53:54.550875 139748315285248 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.022 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.664992570877075 s) (per-host)
I0422 14:53:54.550977 139748315285248 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.621782s.
I0422 14:54:02.626590 139748315285248 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.697379s.
I0422 14:54:02.626867 139726015162112 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0422 14:54:02.626989 139726015162112 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0422 14:54:02.627035 139726015162112 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0422 14:54:02.628535 139726015162112 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0422 14:54:02.628730 139880708081472 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0422 14:54:02.628899 139880708081472 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0422 14:54:02.629927 139880708081472 metric_logger.py:196] completed step: 9, seconds: 0.163, TFLOP/s/device: 83.222, Tokens/s/device: 12544.177, 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: Wed Apr 22 14:54:14 UTC 2026
EXIT_CODE=0