XPK Start: Mon Apr 20 15:57:01 UTC 2026
PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
2026-04-20 15:57:26.247157: 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)
I0420 15:57:26.424816 136024784041792 max_utils.py:273] Attempting to initialize the jax distributed system...
I0420 15:57:35.466860 136024784041792 distributed.py:149] Starting JAX distributed service on [::]:8482
I0420 15:57:35.469202 136024784041792 distributed.py:172] Connecting to JAX distributed service on mt-06-grad-accum-zu8nd-slice-job-0-0.mt-06-grad-accum-zu8nd:8482
I0420 15:57:36.039056 136024784041792 max_utils.py:284] Jax distributed system initialized!
I0420 15:57:41.901652 136024784041792 max_utils.py:800] System Information: Jax Version: 0.9.2
I0420 15:57:41.901756 136024784041792 max_utils.py:801] System Information: Jaxlib Version: 0.9.2
I0420 15:57:41.901796 136024784041792 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
I0420 15:57:41.901831 136024784041792 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing.
I0420 15:57:42.607983 136024784041792 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 15:57:42.608572 136024784041792 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1)
I0420 15:57:42.608747 136024784041792 checkpointing.py:688] Setting up checkpoint logger...
I0420 15:57:42.608797 136024784041792 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True
I0420 15:57:42.608839 136024784041792 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None
I0420 15:57:42.609187 136024784041792 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 0x7bb6239ef8c0>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB)
I0420 15:57:45.585971 136024784041792 checkpointing.py:266] Enabling policy for fixed interval checkpointing.
I0420 15:57:45.586221 136024784041792 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ba2703ea150>}, handler_registry=None
I0420 15:57:45.586453 136024784041792 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ba2703ea150>` 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`.
I0420 15:57:45.586502 136024784041792 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ba2703ec380>` 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`.
I0420 15:57:45.586538 136024784041792 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 0x7ba2703ea150>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7ba2703ea150>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ba2703ec380>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7ba2703ec380>}).
I0420 15:57:45.586852 136024784041792 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34
I0420 15:57:45.586925 136024784041792 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7ba198797740> timeout: 1200 secs and primary_host=0 for async checkpoint writes
I0420 15:57:46.725913 136024784041792 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints
I0420 15:57:47.178944 136024784041792 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7ba2703e92e0>
I0420 15:57:47.179147 136024784041792 checkpointing.py:302] Checkpoint manager created!
I0420 15:57:48.872205 136024784041792 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed').
I0420 15:57:48.872307 136024784041792 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 15:57:49.256559 136024784041792 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 15:57:49.256647 136024784041792 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 15:57:49.273092 136024784041792 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed').
I0420 15:57:49.273149 136024784041792 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None).
I0420 15:57:49.296781 136024784041792 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 15:57:49.296854 136024784041792 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 15:57:49.313148 136024784041792 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 15:57:49.313208 136024784041792 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 15:57:49.329525 136024784041792 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim').
I0420 15:57:49.329588 136024784041792 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 15:57:49.354018 136024784041792 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv').
I0420 15:57:49.354093 136024784041792 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None).
I0420 15:57:49.374836 136024784041792 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp').
I0420 15:57:49.374901 136024784041792 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None).
I0420 15:57:49.589155 136024784041792 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint
I0420 15:57:49.589261 136024784041792 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint.
fsdp: 32
I0420 15:57:51.008124 136024784041792 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale
Shape: float32[2048]
Logical: P('norm',)
Physical: (None,)
I0420 15:57:51.008249 136024784041792 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0420 15:57:51.008303 136024784041792 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel
Shape: float32[2048,16,7168]
Logical: P('embed', 'layers', 'mlp')
Physical: ('fsdp', None, None)
I0420 15:57:51.008369 136024784041792 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel
Shape: float32[7168,16,2048]
Logical: P('mlp', 'layers', 'embed')
Physical: (None, None, 'fsdp')
I0420 15:57:51.008443 136024784041792 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0420 15:57:51.008498 136024784041792 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale
Shape: float32[2048,16]
Logical: P('norm', 'layers')
Physical: (None, None)
I0420 15:57:51.008556 136024784041792 maxtext_utils.py:1827] 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)
I0420 15:57:51.008610 136024784041792 maxtext_utils.py:1827] 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')
I0420 15:57:51.008654 136024784041792 maxtext_utils.py:1827] 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)
I0420 15:57:51.008691 136024784041792 maxtext_utils.py:1827] 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)
I0420 15:57:51.008740 136024784041792 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel
Shape: float32[2048,32000]
Logical: P('embed_vocab', 'vocab')
Physical: ('fsdp', None)
I0420 15:57:51.008789 136024784041792 maxtext_utils.py:1827] params/params/token_embedder/embedding
Shape: float32[32000,2048]
Logical: P('vocab', 'embed_vocab')
Physical: (None, 'fsdp')
I0420 15:57:51.032152 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048]............................................... Unknown.
I0420 15:57:51.032214 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048]............................................... (None,).
I0420 15:57:51.046846 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,7168]....................................... Unknown.
I0420 15:57:51.046908 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,7168]....................................... ('fsdp', None, None).
I0420 15:57:51.075989 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[7168,16,2048]....................................... Unknown.
I0420 15:57:51.076053 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[7168,16,2048]....................................... (None, None, 'fsdp').
I0420 15:57:51.090693 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16]............................................ Unknown.
I0420 15:57:51.090750 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16]............................................ (None, None).
I0420 15:57:51.119805 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,16,16,128]..................................... Unknown.
I0420 15:57:51.119866 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,16,16,128]..................................... ('fsdp', None, None, None).
I0420 15:57:51.134461 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[16,16,128,2048]..................................... Unknown.
I0420 15:57:51.134518 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[16,16,128,2048]..................................... (None, None, None, 'fsdp').
I0420 15:57:51.178054 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[2048,32000]......................................... Unknown.
I0420 15:57:51.178137 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[2048,32000]......................................... ('fsdp', None).
I0420 15:57:51.192720 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Logical: float32[32000,2048]......................................... Unknown.
I0420 15:57:51.192775 136024784041792 gradient_accumulation.py:70] gradient_accumulation/inputs Physical: float32[32000,2048]......................................... (None, 'fsdp').
I0420 15:57:51.859240 136024784041792 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0420 15:57:51.859336 136024784041792 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None).
I0420 15:57:51.874758 136024784041792 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length').
I0420 15:57:51.874817 136024784041792 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None).
I0420 15:58:03.200105 136024784041792 max_utils.py:791] Total memory size: 1.7 GB, Output size: 0.4 GB, Temp size: 1.3 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB.
I0420 15:58:03.208175 136024784041792 metric_logger.py:301] number parameters: 1.104 billion
I0420 15:58:05.390221 136024784041792 checkpointing.py:794] Waiting for step 0 to finish before checkpoint...
I0420 15:58:15.736169 136024784041792 checkpointing.py:798] Waited 10.345925092697144 seconds for step 0 to finish before starting checkpointing.
I0420 15:58:15.738611 136024784041792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 15:58:15.740438 136024784041792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0
I0420 15:58:15.741794 136024784041792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0.
I0420 15:58:16.462203 136024784041792 signaling_client.py:364] Using JaxDistributedSignalingClient
I0420 15:58:16.476873 136024784041792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0420 15:58:16.476939 136024784041792 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
I0420 15:58:16.725398 135900306581248 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0
I0420 15:58:16.745729 136024784041792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.269739s
I0420 15:58:16.746689 136024784041792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.588 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.27602696418762207 s) (per-host)
I0420 15:58:16.746762 136024784041792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.276119s (batch_requests_ready=0.002038s, total_serialization_initiated=0.273206s, others=0.000875s)
I0420 15:58:16.746838 136024784041792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.280190s (all_items=0.000018s, per_item={'items': '0.00001788'}, temp_paths=0.280172)
I0420 15:58:16.747647 136024784041792 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.01 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0.
I0420 15:58:16.747985 135900314973952 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 16:18:16.747947
I0420 15:58:16.760960 136024784041792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 15:58:16.761243 135898729539328 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 15:58:16.761371 136024784041792 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776700695.7385929, 'wait_for_prev_duration_secs': 6.29425048828125e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776700695.7404773, 'checkpointer_blocking_duration_secs': 1.0076713562011719, 'get_old_steps_start_time': 1776700696.7481735, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776700695.7366993, 'checkpoint_manager_blocking_duration_secs': 1.024634599685669}
I0420 15:58:16.761475 136024784041792 checkpointing.py:409] Started an asynchronous checkpoint save for step 0
I0420 15:58:16.761526 136024784041792 max_utils.py:750]
Memstats: After params initialized:
I0420 15:58:16.761580 136024784041792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0))
I0420 15:58:16.761614 136024784041792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0))
I0420 15:58:16.761643 136024784041792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0))
I0420 15:58:16.761668 136024784041792 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0))
I0420 15:58:17.115790 136024784041792 metric_logger.py:196] completed step: 0, seconds: 1.856, TFLOP/s/device: 29.290, Tokens/s/device: 4414.854, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0420 15:58:17.119605 136024784041792 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/tensorboard/'
I0420 15:58:17.691872 136024784041792 metric_logger.py:196] completed step: 1, seconds: 11.724, TFLOP/s/device: 4.636, Tokens/s/device: 698.756, total_weights: 262144, loss: 10.877, lm_loss: 10.877, perplexity: 52959.059
I0420 15:58:18.269946 136024784041792 metric_logger.py:196] completed step: 2, seconds: 0.417, TFLOP/s/device: 130.312, Tokens/s/device: 19642.116, total_weights: 262144, loss: 10.563, lm_loss: 10.563, perplexity: 38662.707
I0420 15:58:18.598283 135899811673856 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/items
I0420 15:58:18.753278 135896545310464 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776700698342381200, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776700698342381200, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0420 15:58:18.847902 136024784041792 metric_logger.py:196] completed step: 3, seconds: 0.170, TFLOP/s/device: 318.842, Tokens/s/device: 48059.323, total_weights: 262144, loss: 10.272, lm_loss: 10.272, perplexity: 28909.668
I0420 15:58:19.784197 2552 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com
I0420 15:58:21.583806 135899756111616 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/items/array_metadatas/process_0
I0420 15:58:58.485038 135898737932032 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 38.713327s (commit=37.016516s, array_metadata_write=1.696811s)
I0420 15:58:58.487342 135900314973952 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 37.595 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.016650915145874 s) (per-host)
I0420 15:58:58.487467 135900314973952 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.739353s.
I0420 15:58:59.289196 135900314973952 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776700698342381200, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0420 15:59:00.029394 135896545310464 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776700698342381200, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/_CHECKPOINT_METADATA
I0420 15:59:00.273551 135900314973952 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 15:59:00.448291 135900314973952 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000261s.
I0420 15:59:01.362276 135900314973952 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.947539s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/items
I0420 15:59:01.363956 135900314973952 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0/items
I0420 15:59:01.972269 135900314973952 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0
I0420 15:59:03.601832 135900314973952 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0`.
I0420 15:59:03.602753 135900314973952 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 47.86 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/0
I0420 15:59:03.604292 135900314973952 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 46.856176s.
I0420 15:59:03.604439 135898729539328 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 15:59:03.604503 135898729539328 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 15:59:03.604571 135898729539328 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts...
I0420 15:59:03.606618 135898729539328 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts.
I0420 15:59:05.957680 136024784041792 metric_logger.py:196] completed step: 4, seconds: 0.578, TFLOP/s/device: 94.005, Tokens/s/device: 14169.480, total_weights: 262144, loss: 10.022, lm_loss: 10.022, perplexity: 22524.992
I0420 15:59:05.966848 136024784041792 metric_logger.py:196] completed step: 5, seconds: 0.578, TFLOP/s/device: 93.959, Tokens/s/device: 14162.572, total_weights: 262144, loss: 9.820, lm_loss: 9.820, perplexity: 18401.865
I0420 15:59:06.543902 136024784041792 metric_logger.py:196] completed step: 6, seconds: 47.108, TFLOP/s/device: 1.154, Tokens/s/device: 173.899, total_weights: 262144, loss: 9.667, lm_loss: 9.667, perplexity: 15787.604
I0420 15:59:07.121686 136024784041792 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 8093.599, Tokens/s/device: 1219955.324, total_weights: 262144, loss: 9.561, lm_loss: 9.561, perplexity: 14203.827
I0420 15:59:07.699814 136024784041792 metric_logger.py:196] completed step: 8, seconds: 0.578, TFLOP/s/device: 93.992, Tokens/s/device: 14167.446, total_weights: 262144, loss: 9.496, lm_loss: 9.496, perplexity: 13302.920
I0420 15:59:08.277259 136024784041792 checkpointing.py:794] Waiting for step 9 to finish before checkpoint...
I0420 15:59:08.277951 136024784041792 checkpointing.py:798] Waited 0.0007145404815673828 seconds for step 9 to finish before starting checkpointing.
I0420 15:59:08.280033 136024784041792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 15:59:08.281725 136024784041792 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9
I0420 15:59:08.283243 136024784041792 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9.
I0420 15:59:09.069005 136024784041792 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array.
I0420 15:59:09.069114 136024784041792 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
I0420 15:59:09.095710 136024784041792 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.027441s
I0420 15:59:09.096723 136024784041792 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 48.107 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.032065629959106445 s) (per-host)
I0420 15:59:09.096793 136024784041792 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.032153s (batch_requests_ready=0.001825s, total_serialization_initiated=0.029402s, others=0.000926s)
I0420 15:59:09.096878 136024784041792 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.036314s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.036297)
I0420 15:59:09.097649 136024784041792 event_tracking.py:125] [process=0] [async] Finished blocking save in 0.82 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9.
I0420 15:59:09.098002 135898729539328 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 16:19:09.097967
I0420 15:59:09.109052 136024784041792 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize
I0420 15:59:09.109328 135899258017536 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save.
I0420 15:59:09.109470 136024784041792 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776700748.279997, 'wait_for_prev_duration_secs': 9.1552734375e-05, 'time_between_consecutive_saves_sec': 4.673340082168579, 'checkpointer_blocking_start_time': 1776700748.2817633, 'checkpointer_blocking_duration_secs': 0.816354513168335, 'get_old_steps_start_time': 1776700749.0981433, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776700748.2781959, 'checkpoint_manager_blocking_duration_secs': 0.8312339782714844}
I0420 15:59:09.109586 136024784041792 checkpointing.py:409] Started an asynchronous checkpoint save for step 9
I0420 15:59:09.109641 136024784041792 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete.
I0420 15:59:09.315216 135900314973952 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9
I0420 15:59:11.511396 135899274802944 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9/items
I0420 15:59:14.889500 135899756111616 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9/items/array_metadatas/process_0
I0420 15:59:55.668159 135899266410240 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 42.847013s (commit=41.303699s, array_metadata_write=1.543314s)
I0420 15:59:55.669287 135898729539328 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 33.894 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 46.604592084884644 s) (per-host)
I0420 15:59:55.669416 135898729539328 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 46.570774s.
I0420 15:59:57.613289 135898729539328 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389).
I0420 15:59:57.785032 135898729539328 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000242s.
I0420 15:59:58.657714 135898729539328 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 1.936156s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9/items
I0420 15:59:58.659584 135898729539328 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9/items
I0420 15:59:59.270281 135898729539328 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9
I0420 16:00:01.505182 135898729539328 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9`.
I0420 16:00:01.506017 135898729539328 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 53.22 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/linen_xpk_feat_nnx_post_train_fixes_20260420_153543_06_grad_accum/checkpoints/9
I0420 16:00:01.507789 135898729539328 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 52.409143s.
I0420 16:00:01.508023 135899258017536 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save.
I0420 16:00:01.508158 135899258017536 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save.
I0420 16:00:01.508225 135899258017536 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts...
I0420 16:00:01.510112 135899258017536 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts.
I0420 16:00:01.510290 136024784041792 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9.
I0420 16:00:01.510446 136024784041792 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning.
I0420 16:00:01.511483 136024784041792 metric_logger.py:196] completed step: 9, seconds: 0.578, TFLOP/s/device: 94.046, Tokens/s/device: 14175.684, total_weights: 262144, loss: 9.457, lm_loss: 9.457, perplexity: 12802.546
Per train step:
Total TFLOPs: 54.35
split as 93.93% learnable weight flops and 6.07% attention flops
XPK End: Mon Apr 20 16:00:12 UTC 2026
EXIT_CODE=0