XPK Start: Fri Apr 24 13:23:41 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 13:24:06.651086: 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 13:24:06.863385 134568525797184 max_utils.py:273] Attempting to initialize the jax distributed system... I0424 13:24:15.904182 134568525797184 distributed.py:149] Starting JAX distributed service on [::]:8482 I0424 13:24:15.906552 134568525797184 distributed.py:172] Connecting to JAX distributed service on mt-16-zarr3-false-sv-7vyxb-slice-job-0-0.mt-16-zarr3-false-sv-7vyxb:8482 I0424 13:24:17.005359 134568525797184 max_utils.py:284] Jax distributed system initialized! I0424 13:24:22.305374 134568525797184 max_utils.py:800] System Information: Jax Version: 0.9.2 I0424 13:24:22.305477 134568525797184 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0424 13:24:22.305519 134568525797184 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 13:24:22.305554 134568525797184 train_utils.py:391] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0424 13:24:23.007915 134568525797184 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 13:24:23.008532 134568525797184 maxtext_utils.py:1771] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0424 13:24:23.008718 134568525797184 checkpointing.py:688] Setting up checkpoint logger... I0424 13:24:23.008771 134568525797184 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=False I0424 13:24:23.008814 134568525797184 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0424 13:24:23.009179 134568525797184 base_pytree_checkpoint_handler.py:441] Created BasePyTreeCheckpointHandler: use_ocdbt=True, use_zarr3=False, pytree_metadata_options=PyTreeMetadataOptions(support_rich_types=False), array_metadata_store=<orbax.checkpoint._src.metadata.array_metadata_store.Store object at 0x7a630f31cf80>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0424 13:24:25.991713 134568525797184 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0424 13:24:25.991954 134568525797184 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 0x7a4f780c7dd0>}, handler_registry=None I0424 13:24:25.992254 134568525797184 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a4f780c7dd0>` 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 13:24:25.992320 134568525797184 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a4f780cf6e0>` 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 13:24:25.992358 134568525797184 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 0x7a4f780c7dd0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7a4f780c7dd0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a4f780cf6e0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7a4f780cf6e0>}). I0424 13:24:25.992733 134568525797184 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0424 13:24:25.992806 134568525797184 async_checkpointer.py:192] [process=6][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7a4e84375e40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0424 13:24:26.673088 134568525797184 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_16_zarr3_false/checkpoints I0424 13:24:27.019349 134568525797184 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_16_zarr3_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7a4f780c7d70> I0424 13:24:27.019521 134568525797184 checkpointing.py:302] Checkpoint manager created! I0424 13:24:27.956438 134568525797184 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0424 13:24:27.956546 134568525797184 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 13:24:28.346189 134568525797184 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 13:24:28.346303 134568525797184 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 13:24:28.365602 134568525797184 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch_attn', 'activation_length_attn', 'activation_embed_attn'). I0424 13:24:28.365663 134568525797184 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0424 13:24:28.389745 134568525797184 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 13:24:28.389814 134568525797184 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:24:28.406326 134568525797184 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 13:24:28.406389 134568525797184 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:24:28.422877 134568525797184 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 13:24:28.422939 134568525797184 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:24:28.447716 134568525797184 attentions.py:1198] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch_attn', 'activation_length_attn', 'activation_heads', 'activation_kv'). I0424 13:24:28.447786 134568525797184 attentions.py:1198] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0424 13:24:28.468949 134568525797184 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0424 13:24:28.469018 134568525797184 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0424 13:24:28.691287 134568525797184 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0424 13:24:28.691399 134568525797184 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0424 13:24:30.119390 134568525797184 maxtext_utils.py:1880] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0424 13:24:30.119519 134568525797184 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 13:24:30.119574 134568525797184 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 13:24:30.119632 134568525797184 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 13:24:30.119686 134568525797184 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 13:24:30.119727 134568525797184 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 13:24:30.119779 134568525797184 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 13:24:30.119834 134568525797184 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 13:24:30.119878 134568525797184 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 13:24:30.119920 134568525797184 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 13:24:30.119970 134568525797184 maxtext_utils.py:1880] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0424 13:24:30.120021 134568525797184 maxtext_utils.py:1880] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0424 13:24:30.622210 134568525797184 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 13:24:30.622307 134568525797184 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0424 13:24:30.637830 134568525797184 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0424 13:24:30.637890 134568525797184 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0424 13:24:41.719771 134568525797184 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 13:24:41.720683 134568525797184 metric_logger.py:301] number parameters: 1.104 billion I0424 13:24:53.333767 134568525797184 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0424 13:24:53.486973 134568525797184 checkpointing.py:798] Waited 0.15318846702575684 seconds for step 0 to finish before starting checkpointing. I0424 13:24:53.489397 134568525797184 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:24:53.491140 134568525797184 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 0 I0424 13:24:53.492428 134568525797184 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_16_zarr3_false/checkpoints/0. I0424 13:24:54.250982 134568525797184 signaling_client.py:364] Using JaxDistributedSignalingClient I0424 13:24:54.252158 134568525797184 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 13:24:54.252217 134568525797184 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 13:24:54.527810 134568525797184 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276911s I0424 13:24:54.527989 134568525797184 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 2.171 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.7105555534362793 s) (per-host) I0424 13:24:54.528043 134568525797184 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.710619s (batch_requests_ready=0.430301s, total_serialization_initiated=0.280246s, others=0.000072s) I0424 13:24:54.528154 134568525797184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.714740s (all_items=0.000019s, per_item={'items': '0.00001907'}, temp_paths=0.714721) I0424 13:24:54.528956 134568525797184 event_tracking.py:125] [process=6] [async] Finished blocking save in 1.04 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_16_zarr3_false/checkpoints/0. I0424 13:24:54.529265 134442835498752 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:44:54.529233 I0424 13:24:54.531279 134568525797184 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:24:54.531584 134440177633024 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:24:54.531715 134568525797184 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777037093.4893785, 'wait_for_prev_duration_secs': 6.079673767089844e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777037093.4911776, 'checkpointer_blocking_duration_secs': 1.038236379623413, 'get_old_steps_start_time': 1777037094.5294409, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1777037093.4875395, 'checkpoint_manager_blocking_duration_secs': 1.0441365242004395} I0424 13:24:54.531872 134568525797184 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0424 13:24:54.531925 134568525797184 max_utils.py:750] Memstats: After params initialized: I0424 13:24:54.531970 134568525797184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_24(process=6,(0,6,0,0)) I0424 13:24:54.532004 134568525797184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_25(process=6,(1,6,0,0)) I0424 13:24:54.532032 134568525797184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_28(process=6,(0,7,0,0)) I0424 13:24:54.532057 134568525797184 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_29(process=6,(1,7,0,0)) I0424 13:24:54.845474 134568525797184 metric_logger.py:196] completed step: 0, seconds: 11.613, TFLOP/s/device: 1.170, Tokens/s/device: 176.355, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0424 13:24:55.022513 134568525797184 metric_logger.py:196] completed step: 1, seconds: 1.510, TFLOP/s/device: 8.997, Tokens/s/device: 1356.125, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0424 13:24:55.447890 134568525797184 metric_logger.py:196] completed step: 2, seconds: 0.030, TFLOP/s/device: 456.158, Tokens/s/device: 68757.134, total_weights: 65536, loss: 10.560, lm_loss: 10.560, perplexity: 38547.719 I0424 13:24:55.595197 134568525797184 metric_logger.py:196] completed step: 3, seconds: 0.425, TFLOP/s/device: 31.932, Tokens/s/device: 4813.172, total_weights: 65536, loss: 9.980, lm_loss: 9.980, perplexity: 21590.506 I0424 13:24:55.599655 134568525797184 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0424 13:24:55.889347 134568525797184 checkpointing.py:798] Waited 0.2896573543548584 seconds for step 5 to finish before starting checkpointing. I0424 13:24:55.892489 134568525797184 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:24:57.158587 2774 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0424 13:24:59.202494 134440186025728 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_16_zarr3_false/checkpoints/0/items/array_metadatas/process_6 I0424 13:25:35.039243 134442835498752 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 38.319 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.22176766395569 s) (per-host) I0424 13:25:35.039373 134442835498752 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.509998s. I0424 13:25:43.851478 134442835498752 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 49.322085s. I0424 13:25:43.851770 134440177633024 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:25:43.851893 134440177633024 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:25:43.851945 134440177633024 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:25:43.853493 134440177633024 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0424 13:25:43.853682 134568525797184 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 13:25:43.853821 134568525797184 checkpoint_manager.py:1452] Waiting for previous save to complete took 47.961337 seconds. If this number is high, consider checkpointing less frequently. I0424 13:25:43.856075 134568525797184 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 5 I0424 13:25:43.858587 134568525797184 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_16_zarr3_false/checkpoints/5. I0424 13:25:44.158157 134568525797184 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 13:25:44.158251 134568525797184 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 13:25:44.193022 134568525797184 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.036052s I0424 13:25:44.193210 134568525797184 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 39.004 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.03954887390136719 s) (per-host) I0424 13:25:44.193272 134568525797184 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.039621s (batch_requests_ready=0.001793s, total_serialization_initiated=0.037747s, others=0.000082s) I0424 13:25:44.193382 134568525797184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.044682s (all_items=0.000017s, per_item={'items': '0.00001693'}, temp_paths=0.044665) I0424 13:25:44.194149 134568525797184 event_tracking.py:125] [process=6] [async] Finished blocking save in 0.34 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_16_zarr3_false/checkpoints/5. I0424 13:25:44.194498 134440177633024 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:45:44.194456 I0424 13:25:44.202858 134568525797184 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:25:44.203134 134440186025728 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:25:44.203306 134568525797184 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777037095.8924499, 'wait_for_prev_duration_secs': 47.961336851119995, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1777037143.856132, 'checkpointer_blocking_duration_secs': 0.33850717544555664, 'get_old_steps_start_time': 1777037144.1946642, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1777037095.890073, 'checkpoint_manager_blocking_duration_secs': 48.3131947517395} I0424 13:25:44.203423 134568525797184 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0424 13:25:44.204250 134568525797184 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.738, Tokens/s/device: 13375.567, total_weights: 65536, loss: 9.460, lm_loss: 9.460, perplexity: 12840.355 I0424 13:25:44.210329 134568525797184 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.304, Tokens/s/device: 13913.043, total_weights: 65536, loss: 8.959, lm_loss: 8.959, perplexity: 7776.467 I0424 13:25:49.404196 134420195440384 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_16_zarr3_false/checkpoints/5/items/array_metadatas/process_6 I0424 13:26:07.048861 134568525797184 metric_logger.py:196] completed step: 6, seconds: 48.609, TFLOP/s/device: 0.280, Tokens/s/device: 42.132, total_weights: 65536, loss: 8.469, lm_loss: 8.469, perplexity: 4762.740 I0424 13:26:07.196397 134568525797184 metric_logger.py:196] completed step: 7, seconds: 22.693, TFLOP/s/device: 0.599, Tokens/s/device: 90.250, total_weights: 65536, loss: 8.003, lm_loss: 8.003, perplexity: 2989.151 I0424 13:26:07.343500 134568525797184 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.294, Tokens/s/device: 13459.339, total_weights: 65536, loss: 7.572, lm_loss: 7.572, perplexity: 1943.098 I0424 13:26:07.490220 134568525797184 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0424 13:26:07.490955 134568525797184 checkpointing.py:798] Waited 0.000751495361328125 seconds for step 9 to finish before starting checkpointing. I0424 13:26:07.493117 134568525797184 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:26:26.623185 134440177633024 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.194 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.4694938659668 s) (per-host) I0424 13:26:26.623304 134440177633024 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.428693s. I0424 13:26:34.514003 134440177633024 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 50.319377s. I0424 13:26:34.514297 134440186025728 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:26:34.514418 134440186025728 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:26:34.514466 134440186025728 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:26:34.516029 134440186025728 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0424 13:26:34.516226 134568525797184 checkpoint_manager.py:2032] [process=6][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0424 13:26:34.516381 134568525797184 checkpoint_manager.py:1452] Waiting for previous save to complete took 27.023277 seconds. If this number is high, consider checkpointing less frequently. I0424 13:26:34.517856 134568525797184 checkpoint_manager.py:1512] [process=6] Saving checkpoint at step 9 I0424 13:26:34.519989 134568525797184 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_16_zarr3_false/checkpoints/9. I0424 13:26:34.831429 134568525797184 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0424 13:26:34.831528 134568525797184 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 13:26:34.865143 134568525797184 base_pytree_checkpoint_handler.py:154] [process=6][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034710s I0424 13:26:34.865310 134568525797184 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/blocking_gbytes_per_sec: 40.337 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.038242340087890625 s) (per-host) I0424 13:26:34.865360 134568525797184 base_pytree_checkpoint_handler.py:768] [process=6][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038302s (batch_requests_ready=0.001678s, total_serialization_initiated=0.036560s, others=0.000064s) I0424 13:26:34.865447 134568525797184 composite_checkpoint_handler.py:715] [process=6][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042474s (all_items=0.000011s, per_item={'items': '0.00001097'}, temp_paths=0.042463) I0424 13:26:34.866120 134568525797184 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_16_zarr3_false/checkpoints/9. I0424 13:26:34.866477 134420195440384 async_checkpointer.py:76] [process=6][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-24 13:46:34.866438 I0424 13:26:34.868458 134568525797184 checkpoint_manager.py:1560] [process=6][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0424 13:26:34.868739 134440186025728 async_checkpointer.py:280] [process=6][thread=save_finalize] Waiting for background save thread=async_save. I0424 13:26:34.868897 134568525797184 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_16_zarr3_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1777037167.4930744, 'wait_for_prev_duration_secs': 27.023277044296265, 'time_between_consecutive_saves_sec': 23.639541149139404, 'checkpointer_blocking_start_time': 1777037194.5178988, 'checkpointer_blocking_duration_secs': 0.34873366355895996, 'get_old_steps_start_time': 1777037194.866651, 'get_old_steps_duration_secs': 2.47955322265625e-05, 'checkpoint_manager_blocking_start_time': 1777037167.4912124, 'checkpoint_manager_blocking_duration_secs': 27.37765383720398} I0424 13:26:34.869006 134568525797184 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0424 13:26:34.869048 134568525797184 checkpoint_manager.py:2020] [process=6][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0424 13:26:40.066604 134439640225536 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_16_zarr3_false/checkpoints/9/items/array_metadatas/process_6 I0424 13:27:16.544256 134420195440384 base_pytree_checkpoint_handler.py:130] [process=6] /jax/orbax/write/gbytes_per_sec: 37.864 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.717148780822754 s) (per-host) I0424 13:27:16.544372 134420195440384 async_checkpointer.py:90] [process=6][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.677773s. I0424 13:27:26.239382 134420195440384 async_checkpointer.py:160] [process=6][thread=async_save] Background save thread done. Time taken: 51.372766s. I0424 13:27:26.239698 134440186025728 async_checkpointer.py:288] [process=6][thread=save_finalize] Done with waiting for background save thread=async_save. I0424 13:27:26.239817 134440186025728 async_checkpointer.py:298] [process=6][thread=save_finalize] No errors found in background save thread=async_save. I0424 13:27:26.239859 134440186025728 checkpoint_manager.py:2137] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0424 13:27:26.241398 134440186025728 checkpoint_manager.py:2146] [process=6][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0424 13:27:26.241571 134568525797184 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 13:27:26.241725 134568525797184 checkpoint_manager.py:2009] [process=6][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0424 13:27:26.242749 134568525797184 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.116, Tokens/s/device: 13884.746, 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 13:27:37 UTC 2026 EXIT_CODE=0