XPK Start: Tue Apr 21 12:41:22 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-21 12:41:46.362321: 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) I0421 12:41:46.893483 138947006310208 max_utils.py:273] Attempting to initialize the jax distributed system... I0421 12:41:55.935094 138947006310208 distributed.py:149] Starting JAX distributed service on [::]:8482 I0421 12:41:55.937491 138947006310208 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-i4jzi-slice-job-0-0.mt-15-ocdbt-true-sv-i4jzi:8482 I0421 12:41:57.007182 138947006310208 max_utils.py:284] Jax distributed system initialized! I0421 12:42:03.179248 138947006310208 max_utils.py:800] System Information: Jax Version: 0.9.2 I0421 12:42:03.179348 138947006310208 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0421 12:42:03.179388 138947006310208 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 I0421 12:42:03.179424 138947006310208 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 12:42:03.879825 138947006310208 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 12:42:03.880408 138947006310208 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 12:42:03.880637 138947006310208 checkpointing.py:688] Setting up checkpoint logger... I0421 12:42:03.880693 138947006310208 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0421 12:42:03.880752 138947006310208 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0421 12:42:03.881088 138947006310208 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 0x7e5e82f17020>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 12:42:06.739602 138947006310208 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0421 12:42:06.739853 138947006310208 checkpoint_manager.py:708] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e5d58585070>}, handler_registry=None I0421 12:42:06.740097 138947006310208 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e5d58585070>` 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`. I0421 12:42:06.740145 138947006310208 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e4a10697aa0>` 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`. I0421 12:42:06.740180 138947006310208 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 0x7e5d58585070>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7e5d58585070>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e4a10697aa0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7e4a10697aa0>}). I0421 12:42:06.740497 138947006310208 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0421 12:42:06.740568 138947006310208 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7e4a102c8180> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0421 12:42:08.001038 138947006310208 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints I0421 12:42:08.036883 138947006310208 checkpoint_manager.py:929] [process=7][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=5), 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_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7e4a10697710> I0421 12:42:08.037031 138947006310208 checkpointing.py:302] Checkpoint manager created! I0421 12:42:08.979076 138947006310208 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 12:42:08.979184 138947006310208 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 12:42:09.362888 138947006310208 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 12:42:09.362980 138947006310208 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 12:42:09.379682 138947006310208 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 12:42:09.379752 138947006310208 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 12:42:09.404492 138947006310208 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 12:42:09.404558 138947006310208 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:42:09.421159 138947006310208 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 12:42:09.421220 138947006310208 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:42:09.437814 138947006310208 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 12:42:09.437878 138947006310208 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:42:09.462998 138947006310208 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 12:42:09.463072 138947006310208 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:42:09.484312 138947006310208 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 12:42:09.484376 138947006310208 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 12:42:09.702483 138947006310208 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0421 12:42:09.702591 138947006310208 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 12:42:11.145310 138947006310208 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0421 12:42:11.145433 138947006310208 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) I0421 12:42:11.145486 138947006310208 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) I0421 12:42:11.145548 138947006310208 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') I0421 12:42:11.145607 138947006310208 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) I0421 12:42:11.145676 138947006310208 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) I0421 12:42:11.145768 138947006310208 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) I0421 12:42:11.145835 138947006310208 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') I0421 12:42:11.145878 138947006310208 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) I0421 12:42:11.145914 138947006310208 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) I0421 12:42:11.145961 138947006310208 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 12:42:11.146008 138947006310208 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 12:42:11.641260 138947006310208 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 12:42:11.641352 138947006310208 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 12:42:11.656959 138947006310208 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 12:42:11.657020 138947006310208 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 12:42:22.564537 138947006310208 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0421 12:42:22.565418 138947006310208 metric_logger.py:301] number parameters: 1.104 billion I0421 12:42:33.569764 138947006310208 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0421 12:42:34.010938 138947006310208 checkpointing.py:798] Waited 0.4411652088165283 seconds for step 0 to finish before starting checkpointing. I0421 12:42:34.013571 138947006310208 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 12:42:34.015163 138947006310208 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 0 I0421 12:42:34.016659 138947006310208 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/0. I0421 12:42:34.335959 138947006310208 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 12:42:34.336912 138947006310208 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 12:42:34.336970 138947006310208 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 I0421 12:42:34.611673 138947006310208 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275786s I0421 12:42:34.611866 138947006310208 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.487 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2811553478240967 s) (per-host) I0421 12:42:34.611924 138947006310208 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.281223s (batch_requests_ready=0.002194s, total_serialization_initiated=0.278952s, others=0.000076s) I0421 12:42:34.612036 138947006310208 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.285396s (all_items=0.000017s, per_item={'items': '0.00001740'}, temp_paths=0.285378) I0421 12:42:34.612877 138947006310208 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.60 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/0. I0421 12:42:34.613213 138818339542784 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 13:02:34.613177 I0421 12:42:34.623483 138947006310208 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 12:42:34.623796 138817263621888 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 12:42:34.623954 138947006310208 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776775354.0135531, 'wait_for_prev_duration_secs': 6.246566772460938e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776775354.0152025, 'checkpointer_blocking_duration_secs': 0.5981645584106445, 'get_old_steps_start_time': 1776775354.6133928, 'get_old_steps_duration_secs': 2.8848648071289062e-05, 'checkpoint_manager_blocking_start_time': 1776775354.01146, 'checkpoint_manager_blocking_duration_secs': 0.6124567985534668} I0421 12:42:34.624060 138947006310208 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0421 12:42:34.624111 138947006310208 max_utils.py:750] Memstats: After params initialized: I0421 12:42:34.624156 138947006310208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0421 12:42:34.624186 138947006310208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0421 12:42:34.624212 138947006310208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0421 12:42:34.624237 138947006310208 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0421 12:42:34.941329 138947006310208 metric_logger.py:196] completed step: 0, seconds: 11.004, TFLOP/s/device: 1.235, Tokens/s/device: 186.110, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 12:42:35.111754 138947006310208 metric_logger.py:196] completed step: 1, seconds: 1.370, TFLOP/s/device: 9.918, Tokens/s/device: 1494.992, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 12:42:35.547310 138947006310208 metric_logger.py:196] completed step: 2, seconds: 0.023, TFLOP/s/device: 586.512, Tokens/s/device: 88405.422, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0421 12:42:35.694959 138947006310208 metric_logger.py:196] completed step: 3, seconds: 0.436, TFLOP/s/device: 31.163, Tokens/s/device: 4697.215, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0421 12:42:35.699971 138947006310208 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0421 12:42:35.989081 138947006310208 checkpointing.py:798] Waited 0.2890744209289551 seconds for step 5 to finish before starting checkpointing. I0421 12:42:35.992177 138947006310208 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 12:42:38.524370 2789 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 12:42:40.489041 138817273063168 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_7 I0421 12:43:12.860544 138818339542784 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 40.997 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.529794454574585 s) (per-host) I0421 12:43:12.860671 138818339542784 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 38.247344s. I0421 12:43:20.533323 138818339542784 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 45.919979s. I0421 12:43:20.533626 138817263621888 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 12:43:20.533781 138817263621888 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 12:43:20.533841 138817263621888 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 12:43:20.535492 138817263621888 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 12:43:20.535694 138947006310208 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 12:43:20.535856 138947006310208 checkpoint_manager.py:1452] Waiting for previous save to complete took 44.543678 seconds. If this number is high, consider checkpointing less frequently. I0421 12:43:20.537832 138947006310208 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 5 I0421 12:43:20.540408 138947006310208 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/5. I0421 12:43:21.067064 138947006310208 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 12:43:21.067158 138947006310208 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 I0421 12:43:21.103071 138947006310208 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037105s I0421 12:43:21.103238 138947006310208 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 38.061 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04052925109863281 s) (per-host) I0421 12:43:21.103290 138947006310208 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040589s (batch_requests_ready=0.001769s, total_serialization_initiated=0.038753s, others=0.000068s) I0421 12:43:21.103378 138947006310208 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045637s (all_items=0.000016s, per_item={'items': '0.00001621'}, temp_paths=0.045621) I0421 12:43:21.104051 138947006310208 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.57 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/5. I0421 12:43:21.104383 138817263621888 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 13:03:21.104348 I0421 12:43:21.116167 138947006310208 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0421 12:43:21.116457 138796381214464 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 12:43:21.116617 138947006310208 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776775355.9921474, 'wait_for_prev_duration_secs': 44.54367756843567, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776775400.5378706, 'checkpointer_blocking_duration_secs': 0.5666582584381104, 'get_old_steps_start_time': 1776775401.1045525, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776775355.9897053, 'checkpoint_manager_blocking_duration_secs': 45.12687706947327} I0421 12:43:21.116727 138947006310208 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0421 12:43:21.117518 138947006310208 metric_logger.py:196] completed step: 4, seconds: 0.154, TFLOP/s/device: 88.486, Tokens/s/device: 13337.588, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0421 12:43:21.123458 138947006310208 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.127, Tokens/s/device: 13886.346, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0421 12:43:26.299484 138817273063168 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_7 I0421 12:43:43.648252 138947006310208 metric_logger.py:196] completed step: 6, seconds: 45.422, TFLOP/s/device: 0.299, Tokens/s/device: 45.088, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0421 12:43:43.795413 138947006310208 metric_logger.py:196] completed step: 7, seconds: 22.378, TFLOP/s/device: 0.607, Tokens/s/device: 91.517, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0421 12:43:43.942764 138947006310208 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.257, Tokens/s/device: 13453.769, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0421 12:43:44.089327 138947006310208 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0421 12:43:44.090036 138947006310208 checkpointing.py:798] Waited 0.0007290840148925781 seconds for step 9 to finish before starting checkpointing. I0421 12:43:44.091960 138947006310208 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 12:44:03.263406 138817263621888 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.431 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.20065879821777 s) (per-host) I0421 12:44:03.263527 138817263621888 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.159032s. I0421 12:44:09.928578 138817263621888 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 48.824066s. I0421 12:44:09.928892 138796381214464 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 12:44:09.929016 138796381214464 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 12:44:09.929062 138796381214464 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0421 12:44:09.930514 138796381214464 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0421 12:44:09.930674 138947006310208 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0421 12:44:09.930812 138947006310208 checkpoint_manager.py:1452] Waiting for previous save to complete took 25.838858 seconds. If this number is high, consider checkpointing less frequently. I0421 12:44:09.932277 138947006310208 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 9 I0421 12:44:09.934270 138947006310208 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/9. I0421 12:44:10.639420 138947006310208 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 12:44:10.639512 138947006310208 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 I0421 12:44:10.676108 138947006310208 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037596s I0421 12:44:10.676275 138947006310208 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 37.625 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.04099917411804199 s) (per-host) I0421 12:44:10.676326 138947006310208 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.041060s (batch_requests_ready=0.001647s, total_serialization_initiated=0.039348s, others=0.000066s) I0421 12:44:10.676414 138947006310208 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045435s (all_items=0.000011s, per_item={'items': '0.00001121'}, temp_paths=0.045424) I0421 12:44:10.677166 138947006310208 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.74 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/9. I0421 12:44:10.677446 138796381214464 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 13:04:10.677414 I0421 12:44:10.683754 138947006310208 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 12:44:10.683966 138817273063168 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 12:44:10.684092 138947006310208 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776775424.0919292, 'wait_for_prev_duration_secs': 25.838857889175415, 'time_between_consecutive_saves_sec': 23.556398391723633, 'checkpointer_blocking_start_time': 1776775449.9323277, 'checkpointer_blocking_duration_secs': 0.7452576160430908, 'get_old_steps_start_time': 1776775450.6776042, 'get_old_steps_duration_secs': 2.5033950805664062e-05, 'checkpoint_manager_blocking_start_time': 1776775424.0902658, 'checkpoint_manager_blocking_duration_secs': 26.593794584274292} I0421 12:44:10.684198 138947006310208 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0421 12:44:10.684239 138947006310208 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 12:44:15.691492 138815129765632 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_7 I0421 12:44:52.981402 138796381214464 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.302 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.34608697891235 s) (per-host) I0421 12:44:52.981531 138796381214464 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.303977s. I0421 12:44:58.139269 138796381214464 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 47.461700s. I0421 12:44:58.139492 138817273063168 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 12:44:58.139548 138817273063168 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 12:44:58.139587 138817273063168 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 12:44:58.141077 138817273063168 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 12:44:58.141245 138947006310208 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 12:44:58.141400 138947006310208 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 12:44:58.142380 138947006310208 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.350, Tokens/s/device: 13920.041, total_weights: 65536, loss: 7.181, lm_loss: 7.181, perplexity: 1314.495 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Tue Apr 21 12:45:07 UTC 2026 EXIT_CODE=0