MaxView

XPK Start: Thu Apr 16 22:58:15 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-16 22:58:39.333535: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303)
I0416 22:58:39.866880 132640596268864 max_utils.py:273] Attempting to initialize the jax distributed system...
I0416 22:58:48.908583 132640596268864 distributed.py:149] Starting JAX distributed service on [::]:8482
I0416 22:58:48.910953 132640596268864 distributed.py:172] Connecting to JAX distributed service on mt-11-optimizer-offload-jdpxw-slice-job-0-0.mt-11-optimizer-offload-jdpxw:8482
I0416 22:58:49.877427 132640596268864 max_utils.py:284] Jax distributed system initialized!
I0416 22:58:54.944815 132640596268864 max_utils.py:800] System Information: Jax Version: 0.9.2
I0416 22:58:54.944939 132640596268864 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0416 22:58:54.944978 132640596268864 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
I0416 22:58:54.945011 132640596268864 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0416 22:58:55.627221 132640596268864 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 22:58:55.740782 132640596268864 checkpointing.py:688] Setting up checkpoint logger...
I0416 22:58:55.740910 132640596268864 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0416 22:58:55.740955 132640596268864 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0416 22:58:55.741151 132640596268864 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 0x78a232b1ca40>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0416 22:58:58.618680 132640596268864 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0416 22:58:58.619148 132640596268864 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 0x788e7070d8e0>}, handler_registry=None
I0416 22:58:58.619480 132640596268864 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x788e7070d8e0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`.
I0416 22:58:58.619532 132640596268864 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x788e905fdfa0>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`.
I0416 22:58:58.619566 132640596268864 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 0x788e7070d8e0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x788e7070d8e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x788e905fdfa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x788e905fdfa0>}).
I0416 22:58:58.619922 132640596268864 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0416 22:58:58.619994 132640596268864 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x788e90681ee0> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0416 22:58:59.497245 132640596268864 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints
I0416 22:58:59.951296 132640596268864 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x788e907ef4a0>
I0416 22:58:59.951485 132640596268864 checkpointing.py:302] Checkpoint manager created!
I0416 22:59:00.921614 132640596268864 maxtext_utils_nnx.py:112] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to host
I0416 22:59:00.921728 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921766 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921798 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921841 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921871 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921899 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921924 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921948 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921973 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.921998 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922021 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922045 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922069 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922092 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922114 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922137 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922160 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922182 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922205 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922227 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922249 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922272 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922295 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922317 132640596268864 maxtext_utils_nnx.py:112] 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
I0416 22:59:00.922339 132640596268864 maxtext_utils_nnx.py:112] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to host
I0416 22:59:00.922361 132640596268864 maxtext_utils_nnx.py:112] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to host
I0416 22:59:00.923459 132640596268864 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0416 22:59:00.923513 132640596268864 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0416 22:59:02.368878 132640596268864 maxtext_utils.py:1805]  decoder/decoder_norm/scale/value
    Shape:     float32[2048]
    Physical:  (None,)
I0416 22:59:02.368993 132640596268864 maxtext_utils.py:1805]  decoder/layers/mlp/wi_0/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0416 22:59:02.369045 132640596268864 maxtext_utils.py:1805]  decoder/layers/mlp/wi_1/kernel/value
    Shape:     float32[2048,16,7168]
    Physical:  ('fsdp', None, None)
I0416 22:59:02.369087 132640596268864 maxtext_utils.py:1805]  decoder/layers/mlp/wo/kernel/value
    Shape:     float32[7168,16,2048]
    Physical:  (None, None, 'fsdp')
I0416 22:59:02.369125 132640596268864 maxtext_utils.py:1805]  decoder/layers/post_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0416 22:59:02.369161 132640596268864 maxtext_utils.py:1805]  decoder/layers/pre_self_attention_layer_norm/scale/value
    Shape:     float32[2048,16]
    Physical:  (None, None)
I0416 22:59:02.369197 132640596268864 maxtext_utils.py:1805]  decoder/layers/self_attention/key/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0416 22:59:02.369232 132640596268864 maxtext_utils.py:1805]  decoder/layers/self_attention/out/kernel/value
    Shape:     float32[16,16,128,2048]
    Physical:  (None, None, None, 'fsdp')
I0416 22:59:02.369270 132640596268864 maxtext_utils.py:1805]  decoder/layers/self_attention/query/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0416 22:59:02.369303 132640596268864 maxtext_utils.py:1805]  decoder/layers/self_attention/value/kernel/value
    Shape:     float32[2048,16,16,128]
    Physical:  ('fsdp', None, None, None)
I0416 22:59:02.369335 132640596268864 maxtext_utils.py:1805]  decoder/logits_dense/kernel/value
    Shape:     float32[2048,32000]
    Physical:  ('fsdp', None)
I0416 22:59:02.369367 132640596268864 maxtext_utils.py:1805]  token_embedder/embedding/value

    Shape:     float32[32000,2048]
    Physical:  (None, 'fsdp')
I0416 22:59:02.774719 132640596268864 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 22:59:02.774829 132640596268864 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0416 22:59:02.780248 132640596268864 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 22:59:02.780303 132640596268864 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0416 22:59:02.797134 132640596268864 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 22:59:02.797202 132640596268864 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0416 22:59:02.813097 132640596268864 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 22:59:02.813160 132640596268864 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0416 22:59:02.837334 132640596268864 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 22:59:02.837401 132640596268864 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0416 22:59:02.853302 132640596268864 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 22:59:02.853363 132640596268864 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0416 22:59:02.874615 132640596268864 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 22:59:02.874714 132640596268864 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0416 22:59:02.904803 132640596268864 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0416 22:59:02.904892 132640596268864 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0416 22:59:02.924727 132640596268864 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0416 22:59:02.924791 132640596268864 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0416 22:59:03.076126 132640596268864 maxtext_utils_nnx.py:129] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to device
I0416 22:59:03.076235 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076277 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076317 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076348 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076375 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076402 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076429 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076456 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076483 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076511 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076536 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076561 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076586 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076611 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076635 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076659 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076683 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076707 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076732 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076757 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076836 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076887 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076916 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076941 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:03.076966 132640596268864 maxtext_utils_nnx.py:129] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to device
I0416 22:59:03.076988 132640596268864 maxtext_utils_nnx.py:129] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to device
I0416 22:59:13.908830 132640596268864 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.
I0416 22:59:13.911083 132640596268864 metric_logger.py:289] number parameters: 1.104 billion
I0416 22:59:14.451724 132640596268864 maxtext_utils_nnx.py:129] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=0), DictKey(key='count'), GetAttrKey(name='value')) to device
I0416 22:59:14.451849 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.451891 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.451929 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.451960 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.451989 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452018 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452043 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452069 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452096 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452122 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452147 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452172 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452200 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452225 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452268 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452296 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452320 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452344 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452367 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452390 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452414 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452444 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452481 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452518 132640596268864 maxtext_utils_nnx.py:129] 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
I0416 22:59:14.452545 132640596268864 maxtext_utils_nnx.py:129] max_utils.py: Moving (DictKey(key='opt_state'), DictKey(key=2), DictKey(key='count'), GetAttrKey(name='value')) to device
I0416 22:59:14.452568 132640596268864 maxtext_utils_nnx.py:129] max_utils.py: Moving (DictKey(key='step'), GetAttrKey(name='value')) to device
I0416 22:59:25.078552 132640596268864 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0416 22:59:25.247018 132640596268864 checkpointing.py:798] Waited 0.16845107078552246 seconds for step 0 to finish before starting checkpointing.
I0416 22:59:25.249203 132640596268864 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 22:59:25.250926 132640596268864 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0
I0416 22:59:25.252295 132640596268864 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints/0.
I0416 22:59:25.537626 132640596268864 signaling_client.py:364] Using JaxDistributedSignalingClient
I0416 22:59:25.538651 132640596268864 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0416 22:59:25.538710 132640596268864 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
I0416 22:59:25.812185 132640596268864 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275336s
I0416 22:59:25.812360 132640596268864 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 5.409 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2851555347442627 s) (per-host)
I0416 22:59:25.812412 132640596268864 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.285218s (batch_requests_ready=0.002964s, total_serialization_initiated=0.282183s, others=0.000070s)
I0416 22:59:25.812513 132640596268864 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.289377s (all_items=0.000018s, per_item={'items': '0.00001836'}, temp_paths=0.289359)
I0416 22:59:25.813265 132640596268864 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.56 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints/0.
I0416 22:59:25.813588 132509236967168 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-16 23:19:25.813550
I0416 22:59:25.821147 132640596268864 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 22:59:25.821462 132508292278016 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0416 22:59:25.821599 132640596268864 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776380365.249184, 'wait_for_prev_duration_secs': 6.079673767089844e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776380365.2509627, 'checkpointer_blocking_duration_secs': 0.5627748966217041, 'get_old_steps_start_time': 1776380365.8137605, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776380365.2474093, 'checkpoint_manager_blocking_duration_secs': 0.5741522312164307}
I0416 22:59:25.821761 132640596268864 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0416 22:59:25.821858 132640596268864 max_utils.py:750] 
Memstats: After params initialized:
I0416 22:59:25.821911 132640596268864 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_24(process=6,(0,6,0,0))
I0416 22:59:25.821943 132640596268864 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_25(process=6,(1,6,0,0))
I0416 22:59:25.821969 132640596268864 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_28(process=6,(0,7,0,0))
I0416 22:59:25.821993 132640596268864 max_utils.py:756] 	Using (GB) 0.17 / 31.25 (0.544000%) on TPU_29(process=6,(1,7,0,0))
I0416 22:59:26.142892 132640596268864 metric_logger.py:185] completed step: 0, seconds: 11.166, TFLOP/s/device: 1.217, Tokens/s/device: 183.408, total_weights: 65536, loss: 10.874
I0416 22:59:26.320701 132640596268864 metric_logger.py:185] completed step: 1, seconds: 1.063, TFLOP/s/device: 12.785, Tokens/s/device: 1927.168, total_weights: 65536, loss: 10.874
I0416 22:59:26.719898 132640596268864 metric_logger.py:185] completed step: 2, seconds: 0.016, TFLOP/s/device: 835.617, Tokens/s/device: 125953.260, total_weights: 65536, loss: 10.262
I0416 22:59:26.882775 132640596268864 metric_logger.py:185] completed step: 3, seconds: 0.389, TFLOP/s/device: 34.921, Tokens/s/device: 5263.672, total_weights: 65536, loss: 9.732
I0416 22:59:27.209446 132640596268864 metric_logger.py:185] completed step: 4, seconds: 0.180, TFLOP/s/device: 75.657, Tokens/s/device: 11403.880, total_weights: 65536, loss: 9.274
I0416 22:59:27.216805 132640596268864 metric_logger.py:185] completed step: 5, seconds: 0.162, TFLOP/s/device: 83.647, Tokens/s/device: 12608.120, total_weights: 65536, loss: 8.888
I0416 22:59:30.310438    2711 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0416 22:59:32.244246 132509078746880 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints/0/items/array_metadatas/process_6
I0416 22:59:49.182357 132640596268864 metric_logger.py:185] completed step: 6, seconds: 0.327, TFLOP/s/device: 41.514, Tokens/s/device: 6257.486, total_weights: 65536, loss: 8.589
I0416 22:59:49.345444 132640596268864 metric_logger.py:185] completed step: 7, seconds: 21.804, TFLOP/s/device: 0.623, Tokens/s/device: 93.926, total_weights: 65536, loss: 8.380
I0416 22:59:49.508225 132640596268864 metric_logger.py:185] completed step: 8, seconds: 0.168, TFLOP/s/device: 80.862, Tokens/s/device: 12188.445, total_weights: 65536, loss: 8.251
I0416 22:59:49.671981 132640596268864 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0416 22:59:49.672653 132640596268864 checkpointing.py:798] Waited 0.0006849765777587891 seconds for step 9 to finish before starting checkpointing.
I0416 22:59:49.674714 132640596268864 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 23:00:02.012776 132509236967168 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 43.292 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.48553013801575 s) (per-host)
I0416 23:00:02.012927 132509236967168 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.199226s.
I0416 23:00:12.565110 132509236967168 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 46.751393s.
I0416 23:00:12.565383 132508292278016 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 23:00:12.565511 132508292278016 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0416 23:00:12.565561 132508292278016 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0416 23:00:12.567224 132508292278016 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0416 23:00:12.567398 132640596268864 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.
W0416 23:00:12.567531 132640596268864 checkpoint_manager.py:1452] Waiting for previous save to complete took 22.892817 seconds. If this number is high, consider checkpointing less frequently.
I0416 23:00:12.569438 132640596268864 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9
I0416 23:00:12.571566 132640596268864 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints/9.
I0416 23:00:13.283261 132640596268864 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array.
I0416 23:00:13.283354 132640596268864 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
I0416 23:00:13.319136 132640596268864 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037508s
I0416 23:00:13.319303 132640596268864 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 34.548 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04464888572692871 s) (per-host)
I0416 23:00:13.319352 132640596268864 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.044707s (batch_requests_ready=0.002891s, total_serialization_initiated=0.041753s, others=0.000064s)
I0416 23:00:13.319445 132640596268864 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.048968s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.048951)
I0416 23:00:13.320173 132640596268864 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints/9.
I0416 23:00:13.320487 132508292278016 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-16 23:20:13.320454
I0416 23:00:13.322447 132640596268864 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 23:00:13.322714 132485782435584 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save.
I0416 23:00:13.322880 132640596268864 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776380389.674684, 'wait_for_prev_duration_secs': 22.89281725883484, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776380412.5694766, 'checkpointer_blocking_duration_secs': 0.751154899597168, 'get_old_steps_start_time': 1776380413.3206546, 'get_old_steps_duration_secs': 2.8371810913085938e-05, 'checkpoint_manager_blocking_start_time': 1776380389.6729116, 'checkpoint_manager_blocking_duration_secs': 23.64993667602539}
I0416 23:00:13.322988 132640596268864 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0416 23:00:13.323031 132640596268864 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 23:00:18.390691 132509078746880 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_20260416_221011/nnx_xpk_feat_nnx_post_train_fixes_20260416_221011_11_optimizer_offload_true/checkpoints/9/items/array_metadatas/process_6
I0416 23:00:54.415853 132508292278016 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.393 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.14113402366638 s) (per-host)
I0416 23:00:54.415979 132508292278016 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.095374s.
I0416 23:01:02.845915 132508292278016 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.525295s.
I0416 23:01:02.846195 132485782435584 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 23:01:02.846315 132485782435584 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save.
I0416 23:01:02.846363 132485782435584 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0416 23:01:02.847653 132485782435584 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0416 23:01:02.847835 132640596268864 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.
I0416 23:01:02.848002 132640596268864 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 23:01:02.848719 132640596268864 metric_logger.py:185] completed step: 9, seconds: 0.164, TFLOP/s/device: 83.069, Tokens/s/device: 12521.093, total_weights: 65536, loss: 8.175
Per train step:
 Total TFLOPs: 13.59 
 split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Thu Apr 16 23:01:11 UTC 2026
EXIT_CODE=0