XPK Start: Mon Apr 20 18:05:07 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-20 18:05:32.022382: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 18:05:32.203469 138098395113280 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 18:05:41.243903 138098395113280 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 18:05:41.246314 138098395113280 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-cbtk1-slice-job-0-0.mt-10-shardy-false-cbtk1:8482 I0420 18:05:42.528678 138098395113280 max_utils.py:284] Jax distributed system initialized! I0420 18:05:49.082336 138098395113280 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 18:05:49.082436 138098395113280 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 18:05:49.082478 138098395113280 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0420 18:05:49.082512 138098395113280 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 18:05:49.779372 138098395113280 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 18:05:49.890830 138098395113280 checkpointing.py:688] Setting up checkpoint logger... I0420 18:05:49.890942 138098395113280 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 18:05:49.890987 138098395113280 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 18:05:49.891184 138098395113280 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 0x7d98f0d15250>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 18:05:52.783627 138098395113280 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 18:05:52.783890 138098395113280 checkpoint_manager.py:708] [process=1][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d84880ff7d0>}, handler_registry=None I0420 18:05:52.784126 138098395113280 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d84880ff7d0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 18:05:52.784173 138098395113280 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d848852e030>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 18:05:52.784207 138098395113280 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 0x7d84880ff7d0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d84880ff7d0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d848852e030>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d848852e030>}). I0420 18:05:52.784538 138098395113280 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 18:05:52.784608 138098395113280 async_checkpointer.py:192] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d84883d4fe0> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 18:05:54.033355 138098395113280 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints I0420 18:05:54.477736 138098395113280 checkpoint_manager.py:929] [process=1][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d84880fb920> I0420 18:05:54.477905 138098395113280 checkpointing.py:302] Checkpoint manager created! I0420 18:05:55.442265 138098395113280 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 18:05:55.442381 138098395113280 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 18:05:56.803705 138098395113280 maxtext_utils.py:1836] decoder/decoder_norm/scale/value Shape: float32[2048] Physical: (None,) I0420 18:05:56.803848 138098395113280 maxtext_utils.py:1836] decoder/layers/mlp/wi_0/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 18:05:56.803916 138098395113280 maxtext_utils.py:1836] decoder/layers/mlp/wi_1/kernel/value Shape: float32[2048,16,7168] Physical: ('fsdp', None, None) I0420 18:05:56.803974 138098395113280 maxtext_utils.py:1836] decoder/layers/mlp/wo/kernel/value Shape: float32[7168,16,2048] Physical: (None, None, 'fsdp') I0420 18:05:56.804026 138098395113280 maxtext_utils.py:1836] decoder/layers/post_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 18:05:56.804070 138098395113280 maxtext_utils.py:1836] decoder/layers/pre_self_attention_layer_norm/scale/value Shape: float32[2048,16] Physical: (None, None) I0420 18:05:56.804108 138098395113280 maxtext_utils.py:1836] decoder/layers/self_attention/key/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:05:56.804144 138098395113280 maxtext_utils.py:1836] decoder/layers/self_attention/out/kernel/value Shape: float32[16,16,128,2048] Physical: (None, None, None, 'fsdp') I0420 18:05:56.804179 138098395113280 maxtext_utils.py:1836] decoder/layers/self_attention/query/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:05:56.804211 138098395113280 maxtext_utils.py:1836] decoder/layers/self_attention/value/kernel/value Shape: float32[2048,16,16,128] Physical: ('fsdp', None, None, None) I0420 18:05:56.804242 138098395113280 maxtext_utils.py:1836] decoder/logits_dense/kernel/value Shape: float32[2048,32000] Physical: ('fsdp', None) I0420 18:05:56.804273 138098395113280 maxtext_utils.py:1836] token_embedder/embedding/value Shape: float32[32000,2048] Physical: (None, 'fsdp') I0420 18:05:57.204878 138098395113280 nnx_decoders.py:465] nnx_decoders/carry Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 18:05:57.204970 138098395113280 nnx_decoders.py:465] nnx_decoders/carry Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:05:57.210281 138098395113280 nnx_decoders.py:465] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 18:05:57.210336 138098395113280 nnx_decoders.py:465] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:05:57.226441 138098395113280 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 18:05:57.226501 138098395113280 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:05:57.242014 138098395113280 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 18:05:57.242071 138098395113280 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 18:05:57.266880 138098395113280 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:05:57.266944 138098395113280 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:05:57.288923 138098395113280 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:05:57.289021 138098395113280 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:05:57.305118 138098395113280 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 18:05:57.305180 138098395113280 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:05:57.335376 138098395113280 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 18:05:57.335448 138098395113280 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 18:05:57.355221 138098395113280 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 18:05:57.355283 138098395113280 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 18:06:08.671521 138098395113280 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. I0420 18:06:08.673910 138098395113280 metric_logger.py:301] number parameters: 1.104 billion I0420 18:06:19.978621 138098395113280 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 18:06:20.133703 138098395113280 checkpointing.py:798] Waited 0.15505242347717285 seconds for step 0 to finish before starting checkpointing. I0420 18:06:20.136043 138098395113280 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:06:20.137701 138098395113280 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 0 I0420 18:06:20.139434 138098395113280 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0. I0420 18:06:20.464972 138098395113280 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 18:06:20.466022 138098395113280 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:06:20.466081 138098395113280 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 18:06:20.743116 138098395113280 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.278924s I0420 18:06:20.743285 138098395113280 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 5.349 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28838467597961426 s) (per-host) I0420 18:06:20.743341 138098395113280 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.288451s (batch_requests_ready=0.003385s, total_serialization_initiated=0.284992s, others=0.000074s) I0420 18:06:20.743441 138098395113280 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.292633s (all_items=0.000022s, per_item={'items': '0.00002217'}, temp_paths=0.292611) I0420 18:06:20.744203 138098395113280 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.61 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0. I0420 18:06:20.744527 137969616795392 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 18:26:20.744489 I0420 18:06:20.748131 138098395113280 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:06:20.748374 137967986300672 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:06:20.748494 138098395113280 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776708380.1360245, 'wait_for_prev_duration_secs': 6.031990051269531e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776708380.1377385, 'checkpointer_blocking_duration_secs': 0.6069555282592773, 'get_old_steps_start_time': 1776708380.7447166, 'get_old_steps_duration_secs': 2.956390380859375e-05, 'checkpoint_manager_blocking_start_time': 1776708380.1340866, 'checkpoint_manager_blocking_duration_secs': 0.614368200302124} I0420 18:06:20.748605 138098395113280 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 18:06:20.748687 138098395113280 max_utils.py:750] Memstats: After params initialized: I0420 18:06:20.748738 138098395113280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_2(process=1,(2,0,0,0)) I0420 18:06:20.748772 138098395113280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_3(process=1,(3,0,0,0)) I0420 18:06:20.748799 138098395113280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_6(process=1,(2,1,0,0)) I0420 18:06:20.748823 138098395113280 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_7(process=1,(3,1,0,0)) I0420 18:06:21.064612 138098395113280 metric_logger.py:196] completed step: 0, seconds: 11.304, TFLOP/s/device: 1.202, Tokens/s/device: 181.181, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0420 18:06:21.245436 138098395113280 metric_logger.py:196] completed step: 1, seconds: 1.084, TFLOP/s/device: 12.533, Tokens/s/device: 1889.107, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52786.469 I0420 18:06:21.674543 138098395113280 metric_logger.py:196] completed step: 2, seconds: 0.035, TFLOP/s/device: 392.295, Tokens/s/device: 59130.937, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28632.133 I0420 18:06:21.822178 138098395113280 metric_logger.py:196] completed step: 3, seconds: 0.419, TFLOP/s/device: 32.417, Tokens/s/device: 4886.289, total_weights: 65536, loss: 9.732, lm_loss: 9.732, perplexity: 16855.789 I0420 18:06:22.118379 138098395113280 metric_logger.py:196] completed step: 4, seconds: 0.165, TFLOP/s/device: 82.332, Tokens/s/device: 12410.015, total_weights: 65536, loss: 9.274, lm_loss: 9.274, perplexity: 10661.594 I0420 18:06:22.126075 138098395113280 metric_logger.py:196] completed step: 5, seconds: 0.146, TFLOP/s/device: 92.858, Tokens/s/device: 13996.624, total_weights: 65536, loss: 8.888, lm_loss: 8.888, perplexity: 7245.978 I0420 18:06:24.051186 2775 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 18:06:26.570337 137968003086080 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/0/items/array_metadatas/process_1 I0420 18:06:44.656513 138098395113280 metric_logger.py:196] completed step: 6, seconds: 0.297, TFLOP/s/device: 45.747, Tokens/s/device: 6895.484, total_weights: 65536, loss: 8.589, lm_loss: 8.589, perplexity: 5371.156 I0420 18:06:44.804049 138098395113280 metric_logger.py:196] completed step: 7, seconds: 22.384, TFLOP/s/device: 0.607, Tokens/s/device: 91.493, total_weights: 65536, loss: 8.380, lm_loss: 8.380, perplexity: 4360.465 I0420 18:06:44.951618 138098395113280 metric_logger.py:196] completed step: 8, seconds: 0.153, TFLOP/s/device: 88.666, Tokens/s/device: 13364.744, total_weights: 65536, loss: 8.251, lm_loss: 8.251, perplexity: 3831.256 I0420 18:06:45.099857 138098395113280 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0420 18:06:45.100512 138098395113280 checkpointing.py:798] Waited 0.0006687641143798828 seconds for step 9 to finish before starting checkpointing. I0420 18:06:45.102807 138098395113280 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 18:06:58.609452 137969616795392 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 41.400 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.15448546409607 s) (per-host) I0420 18:06:58.609593 137969616795392 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.864937s. I0420 18:07:05.708541 137969616795392 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 44.963883s. I0420 18:07:05.708782 137967986300672 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:07:05.708860 137967986300672 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:07:05.708925 137967986300672 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:07:05.711419 137967986300672 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 18:07:05.711536 138098395113280 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 18:07:05.711618 138098395113280 checkpoint_manager.py:1452] Waiting for previous save to complete took 20.608810 seconds. If this number is high, consider checkpointing less frequently. I0420 18:07:05.713613 138098395113280 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 9 I0420 18:07:05.717402 138098395113280 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9. I0420 18:07:06.438626 138098395113280 jax_array_handlers.py:360] Scheduling D2H of 69 prioritized jax.Array. I0420 18:07:06.438731 138098395113280 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 18:07:06.475262 138098395113280 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038264s I0420 18:07:06.475428 138098395113280 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 33.979 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.045397281646728516 s) (per-host) I0420 18:07:06.475478 138098395113280 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.045457s (batch_requests_ready=0.002770s, total_serialization_initiated=0.042619s, others=0.000068s) I0420 18:07:06.475576 138098395113280 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.049570s (all_items=0.000015s, per_item={'items': '0.00001454'}, temp_paths=0.049556) I0420 18:07:06.476304 138098395113280 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.76 seconds. Continuing save @ gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9. I0420 18:07:06.476648 137967986300672 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 18:27:06.476611 I0420 18:07:06.480101 138098395113280 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0420 18:07:06.480368 137968003086080 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0420 18:07:06.480539 138098395113280 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776708405.1027772, 'wait_for_prev_duration_secs': 20.60880970954895, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776708425.7136655, 'checkpointer_blocking_duration_secs': 0.7631418704986572, 'get_old_steps_start_time': 1776708426.4768302, 'get_old_steps_duration_secs': 3.24249267578125e-05, 'checkpoint_manager_blocking_start_time': 1776708405.1007705, 'checkpoint_manager_blocking_duration_secs': 21.379730701446533} I0420 18:07:06.480688 138098395113280 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0420 18:07:06.480743 138098395113280 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 18:07:11.672252 137944803268352 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 69 array_metadata.ArrayMetadata to gs://lance-maxtext/nnx_ckpt_xpk_feat_nnx_post_train_fixes_20260420_153543/nnx_xpk_feat_nnx_post_train_fixes_20260420_153543_10_shardy_false/checkpoints/9/items/array_metadatas/process_1 I0420 18:07:48.557804 137967986300672 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 37.495 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.12773418426514 s) (per-host) I0420 18:07:48.557929 137967986300672 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.081151s. I0420 18:07:55.801998 137967986300672 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 49.325203s. I0420 18:07:55.802277 137968003086080 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 18:07:55.802395 137968003086080 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0420 18:07:55.802444 137968003086080 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0420 18:07:55.804960 137968003086080 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0420 18:07:55.805140 138098395113280 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0420 18:07:55.805302 138098395113280 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 18:07:55.806499 138098395113280 metric_logger.py:196] completed step: 9, seconds: 0.148, TFLOP/s/device: 91.674, Tokens/s/device: 13818.138, total_weights: 65536, loss: 8.175, lm_loss: 8.175, perplexity: 3551.438 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 18:08:07 UTC 2026 EXIT_CODE=0