2026-04-16 18:18:13.646122: 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 18:18:13.765102 124806203493504 max_utils.py:238] Skipping jax distributed system due to skip_jax_distributed_system=True flag.
I0416 18:19:05.999522 124806203493504 max_utils.py:800] System Information: Jax Version: 0.9.2
I0416 18:19:05.999639 124806203493504 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0416 18:19:05.999679 124806203493504 max_utils.py:802] System Information: Jax Backend: PJRT C API
TFRT TPU v6 lite
Built on Apr 6 2026 20:48:10 (1775533690) cl/895581894
I0416 18:19:05.999706 124806203493504 train_utils.py:364] WARNING: 'dataset_path' might be pointing your local file system
I0416 18:19:05.999732 124806203493504 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0416 18:19:05.999817 124806203493504 train.py:811] [DECOUPLED NO-OP] skipping cloud diagnostics wrapper.
W0416 18:19:06.094751 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 18:19:06.482007 124806203493504 maxtext_utils.py:1687] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 18:19:06.482191 124806203493504 maxtext_utils.py:1687] Num_devices: 8, shape (1, 1, 1, 8, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0416 18:19:06.482404 124806203493504 checkpointing.py:688] Setting up checkpoint logger...
I0416 18:19:06.482443 124806203493504 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0416 18:19:06.482475 124806203493504 pytree_checkpoint_handler.py:577] save_device_host_concurrent_bytes=None
I0416 18:19:06.483144 124806203493504 base_pytree_checkpoint_handler.py:411] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x71821ebf4da0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0416 18:19:08.863899 124806203493504 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0416 18:19:08.864097 124806203493504 checkpoint_manager.py:702] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x717c72dc24b0>}, handler_registry=None
I0416 18:19:08.864330 124806203493504 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x717c72dc24b0>` 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 18:19:08.864369 124806203493504 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x717ba293e4b0>` 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 18:19:08.864400 124806203493504 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 0x717c72dc24b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x717c72dc24b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x717ba293e4b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x717ba293e4b0>}).
I0416 18:19:08.864693 124806203493504 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.28
I0416 18:19:08.864743 124806203493504 async_checkpointer.py:177] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>.<lambda> at 0x717ba2929ee0> timeout: 600 secs and primary_host=0 for async checkpoint writes
I0416 18:19:08.999029 124806203493504 checkpoint_manager.py:1788] Found 0 checkpoint steps in gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints
I0416 18:19:08.999279 124806203493504 checkpoint_manager.py:921] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_hns=False, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False), root_directory=gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x717ba293d700>
I0416 18:19:08.999355 124806203493504 checkpointing.py:302] Checkpoint manager created!
W0416 18:19:09.015197 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 18:19:09.201560 124806203493504 nnx_wrappers.py:437] Unknown Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0416 18:19:09.201659 124806203493504 nnx_wrappers.py:437] Unknown Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 18:19:09.291278 124806203493504 attentions.py:1088] attentions/inputs_q Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 18:19:09.291427 124806203493504 attentions.py:1088] attentions/inputs_q Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 18:19:09.305785 124806203493504 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[8,2048,2048]....................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0416 18:19:09.305832 124806203493504 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[8,2048,2048]....................................... ('fsdp', None, None).
I0416 18:19:09.327785 124806203493504 attentions.py:1154] attentions/query Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 18:19:09.327840 124806203493504 attentions.py:1154] attentions/query Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 18:19:09.342313 124806203493504 attentions.py:1155] attentions/key Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 18:19:09.342361 124806203493504 attentions.py:1155] attentions/key Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 18:19:09.356631 124806203493504 attentions.py:1156] attentions/value Logical: bfloat16[8,2048,16,128]..................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0416 18:19:09.356681 124806203493504 attentions.py:1156] attentions/value Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 18:19:09.378631 124806203493504 attentions.py:1197] attentions/out Logical: bfloat16[8,2048,16,128]..................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0416 18:19:09.378686 124806203493504 attentions.py:1197] attentions/out Physical: bfloat16[8,2048,16,128]..................................... ('fsdp', None, None, None).
I0416 18:19:09.396794 124806203493504 linears.py:525] linears/x Logical: bfloat16[8,2048,7168]....................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0416 18:19:09.396845 124806203493504 linears.py:525] linears/x Physical: bfloat16[8,2048,7168]....................................... ('fsdp', None, None).
I0416 18:19:09.588468 124806203493504 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0416 18:19:09.588568 124806203493504 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
W0416 18:19:09.638469 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
[DECOUPLED NO-OP] gcs_storage: using stubs.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] mldiagnostics: using stub.
[DECOUPLED NO-OP] workload_monitor: using stub.
[DECOUPLED NO-OP] vertex_tensorboard: using stub.
fsdp: 8
I0416 18:19:10.019168 124806203493504 maxtext_utils.py:1796] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0416 18:19:10.019276 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0416 18:19:10.019319 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0416 18:19:10.019363 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0416 18:19:10.019404 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0416 18:19:10.019432 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0416 18:19:10.019471 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/key/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0416 18:19:10.019510 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/out/kernel
Shape: float32[16,16,128,2048]
Logical: P('heads', 'layers', 'kv', 'embed')
Physical: (None, None, None, 'fsdp')
I0416 18:19:10.019538 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/query/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'q_heads', 'kv')
Physical: ('fsdp', None, None, None)
I0416 18:19:10.019563 124806203493504 maxtext_utils.py:1796] params/params/decoder/layers/self_attention/value/kernel
Shape: float32[2048,16,16,128]
Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim')
Physical: ('fsdp', None, None, None)
I0416 18:19:10.019600 124806203493504 maxtext_utils.py:1796] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0416 18:19:10.019635 124806203493504 maxtext_utils.py:1796] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0416 18:19:10.444698 124806203493504 train.py:157] train/xent Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 18:19:10.444778 124806203493504 train.py:157] train/xent Physical: float32[8,2048]............................................. ('fsdp', None).
I0416 18:19:10.458844 124806203493504 train.py:164] train/z_loss Logical: float32[8,2048]............................................. ('activation_embed_and_logits_batch', 'activation_length').
I0416 18:19:10.458895 124806203493504 train.py:164] train/z_loss Physical: float32[8,2048]............................................. ('fsdp', None).
W0416 18:19:11.083918 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 18:19:11.208871 124806203493504 max_utils.py:791] Total memory size: 3.6 GB, Output size: 1.5 GB, Temp size: 2.0 GB, Argument size: 1.5 GB, Host temp size: 0.0 GB.
I0416 18:19:11.209354 124806203493504 max_utils.py:194] tensorboardX not available; using no-op SummaryWriter.
I0416 18:19:11.209496 124806203493504 metric_logger.py:289] number parameters: 1.104 billion
W0416 18:19:12.081250 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 18:19:12.179867 124806203493504 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0416 18:19:12.297173 124806203493504 checkpointing.py:798] Waited 0.11730098724365234 seconds for step 0 to finish before starting checkpointing.
I0416 18:19:12.297605 124806203493504 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 18:19:12.297760 124806203493504 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 0
I0416 18:19:12.298156 124806203493504 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0.
I0416 18:19:12.376388 124806203493504 signaling_client.py:373] Using ThreadSafeKeyValueSignalingClient
I0416 18:19:12.459187 124698666993216 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0
I0416 18:19:12.464237 124806203493504 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 18:19:12.464329 124806203493504 replica_slices.py:410] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
W0416 18:19:12.677969 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.688689 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.701589 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.712412 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.728166 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.734714 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.740033 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:12.744720 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 18:19:13.098417 124806203493504 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.634625s
I0416 18:19:13.098831 124806203493504 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 17.116 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 721 milliseconds) (per-host)
I0416 18:19:13.098880 124806203493504 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.721113s (batch_requests_ready=0.082368s, total_serialization_initiated=0.638428s, others=0.000317s)
I0416 18:19:13.098938 124806203493504 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.722103s (all_items=0.000020s, per_item={'items': '0.00001979'}, temp_paths=0.722084)
I0416 18:19:13.099612 124698604078656 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 18:19:13.099702 124806203493504 async_checkpointer.py:561] Finished blocking save. Time taken: 0.801903s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0.
I0416 18:19:13.099882 124806203493504 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 18:19:13.100106 124806203493504 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776363552.2975926, 'wait_for_prev_duration_secs': 4.0531158447265625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776363552.2977808, 'checkpointer_blocking_duration_secs': 0.8020279407501221, 'get_old_steps_start_time': 1776363553.0998263, 'get_old_steps_duration_secs': 2.1696090698242188e-05, 'checkpoint_manager_blocking_start_time': 1776363552.2974932, 'checkpoint_manager_blocking_duration_secs': 0.8025901317596436}
I0416 18:19:13.100175 124806203493504 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0416 18:19:13.100213 124806203493504 max_utils.py:750]
Memstats: After params initialized:
I0416 18:19:13.100296 124806203493504 max_utils.py:756] Using (GB) 1.6 / 31.25 (5.120000%) on TPU_0(process=0,(0,0,0,0))
I0416 18:19:13.100322 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_1(process=0,(1,0,0,0))
I0416 18:19:13.100344 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_2(process=0,(0,1,0,0))
I0416 18:19:13.100363 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_3(process=0,(1,1,0,0))
I0416 18:19:13.100381 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_4(process=0,(0,2,0,0))
I0416 18:19:13.100400 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_5(process=0,(1,2,0,0))
I0416 18:19:13.100417 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_6(process=0,(0,3,0,0))
I0416 18:19:13.100433 124806203493504 max_utils.py:756] Using (GB) 1.59 / 31.25 (5.088000%) on TPU_7(process=0,(1,3,0,0))
W0416 18:19:13.106200 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:13.111193 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
W0416 18:19:13.116021 3971316 pjrt_executable.cc:642] Assume version compatibility. PjRt-IFRT does not track XLA executable versions.
I0416 18:19:13.138404 124698656507456 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/items
I0416 18:19:13.276099 124698625050176 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776363553039949308, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776363553039949308, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA
I0416 18:19:13.276354 124698677478976 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 18:19:13.433166 124806203493504 metric_logger.py:185] completed step: 0, seconds: 0.970, TFLOP/s/device: 14.009, Tokens/s/device: 2111.545, total_weights: 16384, loss: 10.874
I0416 18:19:13.434182 124806203493504 metric_logger.py:269] To see full metrics 'tensorboard --logdir=gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/tensorboard/'
I0416 18:19:13.549257 124806203493504 metric_logger.py:185] completed step: 1, seconds: 1.251, TFLOP/s/device: 10.858, Tokens/s/device: 1636.615, total_weights: 16384, loss: 10.874
I0416 18:19:13.667486 124806203493504 metric_logger.py:185] completed step: 2, seconds: 0.007, TFLOP/s/device: 2025.511, Tokens/s/device: 305307.096, total_weights: 16384, loss: 9.775
I0416 18:19:13.785879 124806203493504 metric_logger.py:185] completed step: 3, seconds: 0.117, TFLOP/s/device: 116.251, Tokens/s/device: 17522.545, total_weights: 16384, loss: 8.838
I0416 18:19:13.815491 3973479 google_auth_provider.cc:149] Using credentials at ~/.config/gcloud/application_default_credentials.json
I0416 18:19:13.815546 3973479 google_auth_provider.cc:156] Using OAuth2 AuthProvider
I0416 18:19:14.607139 124698646021696 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/items/array_metadatas/process_0
I0416 18:19:33.929286 124806203493504 metric_logger.py:185] completed step: 4, seconds: 0.118, TFLOP/s/device: 115.002, Tokens/s/device: 17334.338, total_weights: 16384, loss: 7.965
I0416 18:19:33.937270 124806203493504 metric_logger.py:185] completed step: 5, seconds: 0.118, TFLOP/s/device: 114.687, Tokens/s/device: 17286.931, total_weights: 16384, loss: 7.233
I0416 18:19:34.053824 124806203493504 metric_logger.py:185] completed step: 6, seconds: 20.145, TFLOP/s/device: 0.674, Tokens/s/device: 101.664, total_weights: 16384, loss: 6.676
I0416 18:19:34.171949 124806203493504 metric_logger.py:185] completed step: 7, seconds: 0.005, TFLOP/s/device: 2797.432, Tokens/s/device: 421659.461, total_weights: 16384, loss: 6.293
I0416 18:19:34.290212 124806203493504 metric_logger.py:185] completed step: 8, seconds: 0.117, TFLOP/s/device: 115.840, Tokens/s/device: 17460.696, total_weights: 16384, loss: 6.055
I0416 18:19:34.407968 124806203493504 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0416 18:19:34.408884 124806203493504 checkpointing.py:798] Waited 0.0009415149688720703 seconds for step 9 to finish before starting checkpointing.
I0416 18:19:34.409186 124806203493504 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 18:19:34.872161 124698666993216 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 21.052711s (commit=20.643771s, array_metadata_write=0.408939s)
I0416 18:19:34.873281 124698604078656 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 561.749 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 22 seconds) (per-host)
I0416 18:19:34.873383 124698604078656 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 21.773635s.
I0416 18:19:35.082415 124698604078656 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776363553039949308, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA
I0416 18:19:35.254137 124698604078656 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0416 18:19:35.451465 124698625050176 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776363553039949308, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/_CHECKPOINT_METADATA
I0416 18:19:35.697893 124698604078656 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 18:19:35.698536 124698604078656 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.578986s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/items
I0416 18:19:35.699187 124698604078656 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0/items
I0416 18:19:35.934717 124698604078656 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0
I0416 18:19:36.587855 124698604078656 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0`.
I0416 18:19:36.588499 124698604078656 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 24.290707s. directory=gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/0
I0416 18:19:36.588580 124698604078656 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 23.488834s.
I0416 18:19:36.588753 124698677478976 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 18:19:36.588866 124698677478976 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 18:19:36.588924 124698677478976 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0416 18:19:36.588969 124698677478976 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0416 18:19:36.589115 124806203493504 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0.
W0416 18:19:36.589243 124806203493504 checkpoint_manager.py:1441] Waiting for previous save to complete took 2.180062 seconds. If this number is high, consider checkpointing less frequently.
I0416 18:19:36.589981 124806203493504 checkpoint_manager.py:1501] [process=0] Saving checkpoint at step 9
I0416 18:19:36.590286 124806203493504 async_checkpointer.py:452] [process=0] Started async saving checkpoint to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9.
I0416 18:19:36.764835 124806203493504 jax_array_handlers.py:347] Scheduling D2H of 39 prioritized jax.Array.
I0416 18:19:36.764975 124697098323520 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9
I0416 18:19:36.765065 124806203493504 replica_slices.py:410] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False
I0416 18:19:37.474486 124697601640000 atomicity.py:137] Creating tmp directory gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9/items
I0416 18:19:48.233904 124806203493504 base_pytree_checkpoint_handler.py:153] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 11.469777s
I0416 18:19:48.238100 124806203493504 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/blocking_gbytes_per_sec: 1.067 GiB/s (total gbytes: 12.3 GiB) (time elapsed: 11 seconds) (per-host)
I0416 18:19:48.238365 124806203493504 base_pytree_checkpoint_handler.py:732] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 11.562331s (batch_requests_ready=0.084913s, total_serialization_initiated=11.473131s, others=0.004287s)
I0416 18:19:48.238438 124806203493504 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 11.562850s (all_items=0.000020s, per_item={'items': '0.00002003'}, temp_paths=11.562830)
I0416 18:19:48.239872 124697601640000 async_checkpointer.py:79] [process=0][thread=async_save] Background save thread started.
I0416 18:19:48.239952 124806203493504 async_checkpointer.py:561] Finished blocking save. Time taken: 11.649925s. Continuing background save to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9.
I0416 18:19:48.240153 124806203493504 checkpoint_manager.py:1549] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0416 18:19:48.240353 124698666993216 async_checkpointer.py:265] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0416 18:19:48.240468 124806203493504 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776363574.4091508, 'wait_for_prev_duration_secs': 2.1800618171691895, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776363576.5900083, 'checkpointer_blocking_duration_secs': 11.650059700012207, 'get_old_steps_start_time': 1776363588.240085, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776363574.40911, 'checkpoint_manager_blocking_duration_secs': 13.831331014633179}
I0416 18:19:48.240585 124806203493504 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0416 18:19:48.240624 124806203493504 checkpoint_manager.py:1994] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0416 18:19:48.991709 124697570182720 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9/items/array_metadatas/process_0
I0416 18:20:21.853287 124697035408960 base_pytree_checkpoint_handler.py:1217] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 33.614501s (commit=33.152267s, array_metadata_write=0.462234s)
I0416 18:20:21.854486 124697601640000 base_pytree_checkpoint_handler.py:128] [process=0] /jax/checkpoint/write/gbytes_per_sec: 279.708 MiB/s (total gbytes: 12.3 GiB) (time elapsed: 45 seconds) (per-host)
I0416 18:20:21.854589 124697601640000 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 33.614594s.
I0416 18:20:22.319246 124697601640000 array_metadata_store.py:367] [process=0][thread=async_save] Skipped cross-host ArrayMetadata validation because only one process is found: process_index=0.
I0416 18:20:22.739086 124697601640000 ocdbt_utils.py:56] Param validation support for Zarr3 will be added later (b/362328389).
I0416 18:20:22.739679 124697601640000 base_pytree_checkpoint_handler.py:1342] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 0.577012s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9/items
I0416 18:20:22.740365 124697601640000 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9/items
I0416 18:20:23.003082 124697601640000 atomicity.py:608] Finalizing gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9
I0416 18:20:23.696126 124697601640000 atomicity.py:794] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9`.
I0416 18:20:23.696756 124697601640000 async_checkpointer.py:420] Finished async_save (blocking + background). Time taken: 47.106735s. directory=gs://wanglance-maxtext/linen_ckpt_feat_nnx_post_train_fixes_20260416_181521/linen_feat_nnx_post_train_fixes_20260416_181521_02_synthetic/checkpoints/9
I0416 18:20:23.696848 124697601640000 async_checkpointer.py:144] [process=0][thread=async_save] Background save thread done. Time taken: 35.456853s.
I0416 18:20:23.697124 124698666993216 async_checkpointer.py:273] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0416 18:20:23.697234 124698666993216 async_checkpointer.py:283] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0416 18:20:23.697303 124698666993216 checkpoint_manager.py:2103] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0416 18:20:23.697345 124698666993216 checkpoint_manager.py:2112] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0416 18:20:23.697861 124806203493504 checkpoint_manager.py:2006] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0416 18:20:23.698004 124806203493504 checkpoint_manager.py:1983] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0416 18:20:23.698911 124806203493504 metric_logger.py:185] completed step: 9, seconds: 0.118, TFLOP/s/device: 114.902, Tokens/s/device: 17319.239, total_weights: 16384, loss: 5.915
Per train step:
Total TFLOPs: 13.59
split as 93.93% learnable weight flops and 6.07% attention flops