XPK Start: Fri Apr 24 13:06:48 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. `rope_parameters`'s factor field must be a float >= 1, got 40 `rope_parameters`'s beta_fast field must be a float, got 32 `rope_parameters`'s beta_slow field must be a float, got 1 DeepseekV32Config got `key=rope_scaling` in kwargs but hasn't set it as attribute. For RoPE standardization you need to set `self.rope_parameters` in model's config. 2026-04-24 13:07:13.310696: 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) I0424 13:07:13.527408 137020091410240 max_utils.py:273] Attempting to initialize the jax distributed system... I0424 13:07:22.567666 137020091410240 distributed.py:149] Starting JAX distributed service on [::]:8482 I0424 13:07:22.570539 137020091410240 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-false-sv-pn448-slice-job-0-0.mt-15-ocdbt-false-sv-pn448:8482 I0424 13:07:23.679529 137020091410240 max_utils.py:284] Jax distributed system initialized! I0424 13:07:29.791712 137020091410240 max_utils.py:800] System Information: Jax Version: 0.9.2 I0424 13:07:29.791818 137020091410240 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0424 13:07:29.791857 137020091410240 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 I0424 13:07:29.791893 137020091410240 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 13:07:30.490330 137020091410240 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 13:07:30.490914 137020091410240 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 13:07:30.491106 137020091410240 checkpointing.py:688] Setting up checkpoint logger... I0424 13:07:30.491158 137020091410240 checkpointing.py:234] Creating checkpoint manager with ocdbt=False and zarr3=True I0424 13:07:30.491201 137020091410240 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0424 13:07:30.491554 137020091410240 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=False, use_zarr3=True, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7c9dde114170>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 13:07:33.773710 137020091410240 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 13:07:33.773949 137020091410240 checkpoint_manager.py:708] [process=6][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c9cb5f76000>}, handler_registry=None I0424 13:07:33.774211 137020091410240 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c9cb5f76000>` 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`. I0424 13:07:33.774263 137020091410240 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8a44032f90>` 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`. I0424 13:07:33.774301 137020091410240 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 0x7c9cb5f76000>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7c9cb5f76000>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8a44032f90>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7c8a44032f90>}). I0424 13:07:33.774637 137020091410240 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0424 13:07:33.774704 137020091410240 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7c8a2c56de40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0424 13:07:34.519840 137020091410240 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints I0424 13:07:34.924564 137020091410240 checkpoint_manager.py:929] [process=6][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=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_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7c8a4402a2d0> I0424 13:07:34.924741 137020091410240 checkpointing.py:302] Checkpoint manager created! I0424 13:07:35.860077 137020091410240 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 13:07:35.860202 137020091410240 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 13:07:36.239240 137020091410240 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 13:07:36.239341 137020091410240 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 13:07:36.255935 137020091410240 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 13:07:36.255995 137020091410240 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 13:07:36.279865 137020091410240 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 13:07:36.279933 137020091410240 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:07:36.296458 137020091410240 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 13:07:36.296518 137020091410240 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:07:36.312984 137020091410240 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_length_attn', 'activation_kv_heads', 'activation_kv_head_dim'). I0424 13:07:36.313042 137020091410240 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:07:36.337754 137020091410240 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0424 13:07:36.337825 137020091410240 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:07:36.358749 137020091410240 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 13:07:36.358812 137020091410240 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 13:07:36.586324 137020091410240 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 13:07:36.586435 137020091410240 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 13:07:38.022127 137020091410240 maxtext_utils.py:1880] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0424 13:07:38.022259 137020091410240 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0424 13:07:38.022321 137020091410240 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0424 13:07:38.022381 137020091410240 maxtext_utils.py:1880] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0424 13:07:38.022435 137020091410240 maxtext_utils.py:1880] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0424 13:07:38.022474 137020091410240 maxtext_utils.py:1880] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0424 13:07:38.022543 137020091410240 maxtext_utils.py:1880] 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) I0424 13:07:38.022626 137020091410240 maxtext_utils.py:1880] 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') I0424 13:07:38.022687 137020091410240 maxtext_utils.py:1880] 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) I0424 13:07:38.022753 137020091410240 maxtext_utils.py:1880] 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) I0424 13:07:38.022845 137020091410240 maxtext_utils.py:1880] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0424 13:07:38.022940 137020091410240 maxtext_utils.py:1880] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0424 13:07:38.523625 137020091410240 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 13:07:38.523720 137020091410240 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 13:07:38.539353 137020091410240 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 13:07:38.539414 137020091410240 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 13:07:49.799895 137020091410240 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. I0424 13:07:49.800690 137020091410240 metric_logger.py:301] number parameters: 1.104 billion I0424 13:08:01.064015 137020091410240 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 13:08:01.217992 137020091410240 checkpointing.py:798] Waited 0.15395402908325195 seconds for step 0 to finish before starting checkpointing. I0424 13:08:01.220496 137020091410240 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:08:01.222186 137020091410240 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0424 13:08:01.223614 137020091410240 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/0. I0424 13:08:02.006793 137020091410240 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 13:08:02.007879 137020091410240 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 13:08:02.007940 137020091410240 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 I0424 13:08:02.285143 137020091410240 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.278430s I0424 13:08:02.285317 137020091410240 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 2.165 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7126028537750244 s) (per-host) I0424 13:08:02.285374 137020091410240 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.712669s (batch_requests_ready=0.430927s, total_serialization_initiated=0.281668s, others=0.000075s) I0424 13:08:02.285469 137020091410240 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.716673s (all_items=0.000018s, per_item={'items': '0.00001764'}, temp_paths=0.716655) I0424 13:08:02.286266 137020091410240 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.06 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/0. I0424 13:08:02.286606 136894901909248 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:28:02.286566 I0424 13:08:02.290779 137020091410240 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:08:02.291083 136893236168448 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:08:02.291253 137020091410240 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777036081.2204714, 'wait_for_prev_duration_secs': 6.747245788574219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777036081.2222235, 'checkpointer_blocking_duration_secs': 1.0645380020141602, 'get_old_steps_start_time': 1777036082.2867868, 'get_old_steps_duration_secs': 3.170967102050781e-05, 'checkpoint_manager_blocking_start_time': 1777036081.2185373, 'checkpoint_manager_blocking_duration_secs': 1.0726730823516846} I0424 13:08:02.291365 137020091410240 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 13:08:02.291415 137020091410240 max_utils.py:750] Memstats: After params initialized: I0424 13:08:02.291462 137020091410240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 13:08:02.291499 137020091410240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 13:08:02.291526 137020091410240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 13:08:02.291550 137020091410240 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 13:08:02.603886 137020091410240 metric_logger.py:196] completed step: 0, seconds: 11.263, TFLOP/s/device: 1.206, Tokens/s/device: 181.831, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0424 13:08:02.790352 137020091410240 metric_logger.py:196] completed step: 1, seconds: 1.538, TFLOP/s/device: 8.832, Tokens/s/device: 1331.264, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0424 13:08:03.218719 137020091410240 metric_logger.py:196] completed step: 2, seconds: 0.039, TFLOP/s/device: 347.027, Tokens/s/device: 52307.614, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0424 13:08:03.366135 137020091410240 metric_logger.py:196] completed step: 3, seconds: 0.429, TFLOP/s/device: 31.706, Tokens/s/device: 4779.017, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0424 13:08:03.370822 137020091410240 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 13:08:03.660452 137020091410240 checkpointing.py:798] Waited 0.2895958423614502 seconds for step 5 to finish before starting checkpointing. I0424 13:08:03.663409 137020091410240 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:08:04.960276 2781 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 13:08:06.969700 136893244561152 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/0/items/array_metadatas/process_6 I0424 13:08:12.047920 136894901909248 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 150.794 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.475164651870728 s) (per-host) I0424 13:08:12.048049 136894901909248 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 9.761325s. I0424 13:08:18.404632 136894901909248 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.117893s. I0424 13:08:18.404903 136893236168448 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:08:18.405019 136893236168448 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:08:18.405071 136893236168448 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:08:18.407152 136893236168448 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 13:08:18.407340 137020091410240 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0424 13:08:18.407474 137020091410240 checkpoint_manager.py:1452] Waiting for previous save to complete took 14.744064 seconds. If this number is high, consider checkpointing less frequently. I0424 13:08:18.409521 137020091410240 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0424 13:08:18.412065 137020091410240 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/5. I0424 13:08:19.027078 137020091410240 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 13:08:19.027190 137020091410240 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 I0424 13:08:19.064401 137020091410240 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038513s I0424 13:08:19.064570 137020091410240 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 36.762 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.041960716247558594 s) (per-host) I0424 13:08:19.064619 137020091410240 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.042019s (batch_requests_ready=0.001814s, total_serialization_initiated=0.040140s, others=0.000065s) I0424 13:08:19.064708 137020091410240 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.046965s (all_items=0.000022s, per_item={'items': '0.00002170'}, temp_paths=0.046943) I0424 13:08:19.066637 137020091410240 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.66 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/5. I0424 13:08:19.066981 136893236168448 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:28:19.066945 I0424 13:08:19.070532 137020091410240 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:08:19.070772 136894885123840 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:08:19.070895 137020091410240 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777036083.6633794, 'wait_for_prev_duration_secs': 14.744064092636108, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777036098.4095685, 'checkpointer_blocking_duration_secs': 0.6575684547424316, 'get_old_steps_start_time': 1777036099.067159, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1777036083.6610355, 'checkpoint_manager_blocking_duration_secs': 15.409824132919312} I0424 13:08:19.071013 137020091410240 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 13:08:19.071794 137020091410240 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.734, Tokens/s/device: 13374.956, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0424 13:08:19.077666 137020091410240 metric_logger.py:196] completed step: 5, seconds: 0.148, TFLOP/s/device: 92.105, Tokens/s/device: 13883.052, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0424 13:08:24.194372 136893244561152 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/5/items/array_metadatas/process_6 I0424 13:08:29.491740 136893236168448 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 150.882 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.469091176986694 s) (per-host) I0424 13:08:29.491870 136893236168448 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.424777s. I0424 13:08:35.389530 136893236168448 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 16.322423s. I0424 13:08:35.389761 136894885123840 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:08:35.389818 136894885123840 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:08:35.389865 136894885123840 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:08:35.392208 136894885123840 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 13:08:41.906220 137020091410240 metric_logger.py:196] completed step: 6, seconds: 15.705, TFLOP/s/device: 0.865, Tokens/s/device: 130.401, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0424 13:08:42.053867 137020091410240 metric_logger.py:196] completed step: 7, seconds: 22.682, TFLOP/s/device: 0.599, Tokens/s/device: 90.292, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0424 13:08:42.201132 137020091410240 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.953, Tokens/s/device: 13407.968, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0424 13:08:42.347931 137020091410240 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 13:08:42.348585 137020091410240 checkpointing.py:798] Waited 0.0006754398345947266 seconds for step 9 to finish before starting checkpointing. I0424 13:08:42.350649 137020091410240 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:08:42.352389 137020091410240 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0424 13:08:42.353822 137020091410240 event_tracking.py:70] [process=6] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/9. I0424 13:08:42.669714 137020091410240 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 13:08:42.669807 137020091410240 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 I0424 13:08:42.704790 137020091410240 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036155s I0424 13:08:42.704927 137020091410240 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 39.024 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03952908515930176 s) (per-host) I0424 13:08:42.704972 137020091410240 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039583s (batch_requests_ready=0.001772s, total_serialization_initiated=0.037753s, others=0.000058s) I0424 13:08:42.705053 137020091410240 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.043635s (all_items=0.000010s, per_item={'items': '0.00001001'}, temp_paths=0.043625) I0424 13:08:42.705783 137020091410240 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/9. I0424 13:08:42.706132 136866615506688 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:28:42.706080 I0424 13:08:42.709320 137020091410240 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:08:42.709619 136866649077504 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:08:42.709781 137020091410240 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777036122.3506162, 'wait_for_prev_duration_secs': 7.319450378417969e-05, 'time_between_consecutive_saves_sec': 6.958369016647339, 'checkpointer_blocking_start_time': 1777036122.3524275, 'checkpointer_blocking_duration_secs': 0.35385560989379883, 'get_old_steps_start_time': 1777036122.7063022, 'get_old_steps_duration_secs': 2.6226043701171875e-05, 'checkpoint_manager_blocking_start_time': 1777036122.3488176, 'checkpoint_manager_blocking_duration_secs': 0.3609304428100586} I0424 13:08:42.709899 137020091410240 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 13:08:42.709944 137020091410240 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:08:48.235630 136866632292096 array_metadata_store.py:203] [process=6][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260424_120657/linen_xpk_feat_nnx_post_train_fixes_20260424_120657_15_ocdbt_false/checkpoints/9/items/array_metadatas/process_6 I0424 13:08:53.298465 136866615506688 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 148.555 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 10.633046388626099 s) (per-host) I0424 13:08:53.298570 136866615506688 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 10.592323s. I0424 13:09:01.398482 136866615506688 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 18.692219s. I0424 13:09:01.398745 136866649077504 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:09:01.398862 136866649077504 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:09:01.398902 136866649077504 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:09:01.400508 136866649077504 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 13:09:01.400679 137020091410240 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0424 13:09:01.400830 137020091410240 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:09:01.401807 137020091410240 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.298, Tokens/s/device: 13912.098, 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: Fri Apr 24 13:09:13 UTC 2026 EXIT_CODE=0