XPK Start: Sun Apr 19 00:23:10 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-19 00:23:34.426622: 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) I0419 00:23:34.606912 138465611859776 max_utils.py:273] Attempting to initialize the jax distributed system... I0419 00:23:43.648494 138465611859776 distributed.py:149] Starting JAX distributed service on [::]:8482 I0419 00:23:43.650886 138465611859776 distributed.py:172] Connecting to JAX distributed service on mt-10-shardy-false-t08h7-slice-job-0-0.mt-10-shardy-false-t08h7:8482 I0419 00:23:45.950943 138465611859776 max_utils.py:284] Jax distributed system initialized! I0419 00:23:51.914645 138465611859776 max_utils.py:800] System Information: Jax Version: 0.9.2 I0419 00:23:51.914748 138465611859776 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0419 00:23:51.914789 138465611859776 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 I0419 00:23:51.914841 138465611859776 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0419 00:23:52.618443 138465611859776 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 00:23:52.619030 138465611859776 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0419 00:23:52.619206 138465611859776 checkpointing.py:688] Setting up checkpoint logger... I0419 00:23:52.619256 138465611859776 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0419 00:23:52.619298 138465611859776 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0419 00:23:52.619628 138465611859776 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 0x7dee51f3bf80>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0419 00:23:56.352731 138465611859776 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0419 00:23:56.352979 138465611859776 checkpoint_manager.py:708] [process=0][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dda20687ef0>}, handler_registry=None I0419 00:23:56.353217 138465611859776 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dda20687ef0>` 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`. I0419 00:23:56.353266 138465611859776 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dda20689760>` 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`. I0419 00:23:56.353301 138465611859776 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 0x7dda20687ef0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7dda20687ef0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dda20689760>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7dda20689760>}). I0419 00:23:56.353609 138465611859776 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0419 00:23:56.353678 138465611859776 async_checkpointer.py:192] [process=0][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7dda181ff740> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0419 00:23:57.514549 138465611859776 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints I0419 00:23:57.953099 138465611859776 checkpoint_manager.py:929] [process=0][thread=MainThread] CheckpointManager created, primary_host=0, CheckpointManagerOptions=CheckpointManagerOptions(save_interval_steps=1, max_to_keep=None, keep_time_interval=None, keep_period=None, should_keep_fn=None, best_fn=None, best_mode='max', keep_checkpoints_without_metrics=True, step_prefix=None, step_format_fixed_length=None, step_name_format=None, create=True, cleanup_tmp_directories=False, save_on_steps=frozenset(), single_host_load_and_broadcast=False, todelete_subdir=None, todelete_full_path=None, enable_background_delete=False, read_only=False, enable_async_checkpointing=True, async_options=None, multiprocessing_options=MultiprocessingOptions(primary_host=0, active_processes=None, barrier_sync_key_prefix=None), should_save_fn=None, file_options=FileOptions(path_permission_mode=None), save_root_metadata=True, temporary_path_class=None, save_decision_policy=FixedIntervalPolicy(interval=10), preservation_policy=LatestN(n=None), prevent_write_metrics=False, enable_should_save_is_saving_in_progress_check=True, enable_per_process_directory_creation=False, lightweight_initialize=False), root_directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7dda202b3590> I0419 00:23:57.953280 138465611859776 checkpointing.py:302] Checkpoint manager created! I0419 00:23:58.864827 138465611859776 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0419 00:23:58.864929 138465611859776 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:23:59.248543 138465611859776 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 00:23:59.248634 138465611859776 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:23:59.264887 138465611859776 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0419 00:23:59.264947 138465611859776 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0419 00:23:59.288491 138465611859776 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:23:59.288558 138465611859776 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:23:59.304822 138465611859776 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:23:59.304883 138465611859776 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:23:59.321010 138465611859776 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0419 00:23:59.321070 138465611859776 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:23:59.345591 138465611859776 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0419 00:23:59.345658 138465611859776 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0419 00:23:59.366383 138465611859776 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0419 00:23:59.366448 138465611859776 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0419 00:23:59.584166 138465611859776 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0419 00:23:59.584272 138465611859776 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0419 00:24:00.958725 138465611859776 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0419 00:24:00.958859 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 00:24:00.958914 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0419 00:24:00.958973 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0419 00:24:00.959025 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 00:24:00.959063 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0419 00:24:00.959114 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/key/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 00:24:00.959165 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/out/kernel Shape: float32[16,16,128,2048] Logical: P('heads', 'layers', 'kv', 'embed') Physical: (None, None, None, 'fsdp') I0419 00:24:00.959210 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/query/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'q_heads', 'kv') Physical: ('fsdp', None, None, None) I0419 00:24:00.959247 138465611859776 maxtext_utils.py:1827] params/params/decoder/layers/self_attention/value/kernel Shape: float32[2048,16,16,128] Logical: P('embed', 'layers', 'kv_heads', 'kv_head_dim') Physical: ('fsdp', None, None, None) I0419 00:24:00.959293 138465611859776 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0419 00:24:00.959340 138465611859776 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0419 00:24:01.450482 138465611859776 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 00:24:01.450575 138465611859776 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0419 00:24:01.465927 138465611859776 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0419 00:24:01.465990 138465611859776 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0419 00:24:12.503466 138465611859776 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. I0419 00:24:12.511418 138465611859776 metric_logger.py:301] number parameters: 1.104 billion I0419 00:24:24.075468 138465611859776 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0419 00:24:24.228905 138465611859776 checkpointing.py:798] Waited 0.15341472625732422 seconds for step 0 to finish before starting checkpointing. I0419 00:24:24.231341 138465611859776 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 00:24:24.232834 138465611859776 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 0 I0419 00:24:24.234191 138465611859776 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0. I0419 00:24:25.388165 138465611859776 signaling_client.py:364] Using JaxDistributedSignalingClient I0419 00:24:25.404002 138465611859776 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:24:25.404095 138465611859776 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 I0419 00:24:25.679660 138465611859776 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.276401s I0419 00:24:25.680462 138465611859776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 5.445 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2833125591278076 s) (per-host) I0419 00:24:25.680534 138465611859776 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.283403s (batch_requests_ready=0.002748s, total_serialization_initiated=0.279953s, others=0.000703s) I0419 00:24:25.680613 138465611859776 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.287903s (all_items=0.000026s, per_item={'items': '0.00002599'}, temp_paths=0.287877) I0419 00:24:25.681379 138465611859776 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.45 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0. I0419 00:24:25.681694 138337289398016 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 00:44:25.681656 I0419 00:24:25.691317 138465611859776 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:24:25.691612 138332998715136 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:24:25.691766 138465611859776 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776558264.2313225, 'wait_for_prev_duration_secs': 6.0558319091796875e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776558264.232874, 'checkpointer_blocking_duration_secs': 1.4489834308624268, 'get_old_steps_start_time': 1776558265.6818824, 'get_old_steps_duration_secs': 2.6464462280273438e-05, 'checkpoint_manager_blocking_start_time': 1776558264.2293205, 'checkpoint_manager_blocking_duration_secs': 1.4624066352844238} I0419 00:24:25.691884 138465611859776 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0419 00:24:25.691935 138465611859776 max_utils.py:750] Memstats: After params initialized: I0419 00:24:25.691982 138465611859776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_0(process=0,(0,0,0,0)) I0419 00:24:25.692017 138465611859776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_1(process=0,(1,0,0,0)) I0419 00:24:25.692045 138465611859776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_4(process=0,(0,1,0,0)) I0419 00:24:25.692070 138465611859776 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_5(process=0,(1,1,0,0)) I0419 00:24:26.020121 138465611859776 metric_logger.py:196] completed step: 0, seconds: 11.263, TFLOP/s/device: 1.206, Tokens/s/device: 181.840, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0419 00:24:26.023558 138465611859776 metric_logger.py:281] To see full metrics 'tensorboard --logdir=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/tensorboard/' I0419 00:24:26.089220 138336802961152 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0 I0419 00:24:26.412263 138465611859776 metric_logger.py:196] completed step: 1, seconds: 1.943, TFLOP/s/device: 6.993, Tokens/s/device: 1054.124, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0419 00:24:26.558382 138465611859776 metric_logger.py:196] completed step: 2, seconds: 0.393, TFLOP/s/device: 34.586, Tokens/s/device: 5213.212, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0419 00:24:26.705844 138465611859776 metric_logger.py:196] completed step: 3, seconds: 0.009, TFLOP/s/device: 1466.501, Tokens/s/device: 221046.951, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0419 00:24:28.096763 138336786175744 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/items I0419 00:24:28.650680 138332451129088 checkpoint.py:188] Wrote Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776558267834759795, 'commit_timestamp_nsecs': None, 'custom_metadata': {}}, json={"item_handlers": null, "metrics": {}, "performance_metrics": {}, "init_timestamp_nsecs": 1776558267834759795, "commit_timestamp_nsecs": null, "custom_metadata": {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0419 00:24:29.639684 2841 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0419 00:24:31.458038 138333015500544 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/items/array_metadatas/process_0 I0419 00:24:48.836320 138465611859776 metric_logger.py:196] completed step: 4, seconds: 0.147, TFLOP/s/device: 92.375, Tokens/s/device: 13923.827, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0419 00:24:48.844906 138465611859776 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.211, Tokens/s/device: 13898.975, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0419 00:24:48.991075 138465611859776 metric_logger.py:196] completed step: 6, seconds: 22.129, TFLOP/s/device: 0.614, Tokens/s/device: 92.548, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0419 00:24:49.138467 138465611859776 metric_logger.py:196] completed step: 7, seconds: 0.007, TFLOP/s/device: 2061.154, Tokens/s/device: 310679.612, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0419 00:24:49.285629 138465611859776 metric_logger.py:196] completed step: 8, seconds: 0.147, TFLOP/s/device: 92.242, Tokens/s/device: 13903.787, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0419 00:24:49.432529 138465611859776 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0419 00:24:49.433230 138465611859776 checkpointing.py:798] Waited 0.0007228851318359375 seconds for step 9 to finish before starting checkpointing. I0419 00:24:49.435201 138465611859776 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:25:06.486758 138333007107840 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 36.851491s (commit=35.687267s, array_metadata_write=1.164224s) I0419 00:25:06.488057 138337289398016 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 38.442 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.09086585044861 s) (per-host) I0419 00:25:06.488208 138337289398016 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 40.806394s. I0419 00:25:07.152249 138337289398016 checkpoint.py:228] Read Metadata={'item_handlers': None, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776558267834759795, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} from gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0419 00:25:08.289620 138332451129088 checkpoint.py:247] Updated Metadata={'item_handlers': {'items': 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler'}, 'metrics': {}, 'performance_metrics': {}, 'init_timestamp_nsecs': 1776558267834759795, 'commit_timestamp_nsecs': None, 'custom_metadata': {}} to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/_CHECKPOINT_METADATA I0419 00:25:08.706258 138337289398016 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000322s. I0419 00:25:09.194518 138337289398016 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:25:10.131853 138337289398016 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.431699s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/items I0419 00:25:10.133436 138337289398016 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0/items I0419 00:25:10.667674 138337289398016 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0 I0419 00:25:12.908071 138337289398016 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0`. I0419 00:25:12.908956 138337289398016 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 48.68 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/0 I0419 00:25:12.910616 138337289398016 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 47.228810s. I0419 00:25:12.910849 138332998715136 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:25:12.910966 138332998715136 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:25:12.911037 138332998715136 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:25:12.913124 138332998715136 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0419 00:25:12.913315 138465611859776 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0419 00:25:12.913456 138465611859776 checkpoint_manager.py:1452] Waiting for previous save to complete took 23.478254 seconds. If this number is high, consider checkpointing less frequently. I0419 00:25:12.915148 138465611859776 checkpoint_manager.py:1512] [process=0] Saving checkpoint at step 9 I0419 00:25:12.917029 138465611859776 event_tracking.py:70] [process=0] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9. I0419 00:25:14.051255 138465611859776 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0419 00:25:14.051345 138465611859776 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 I0419 00:25:14.075898 138465611859776 base_pytree_checkpoint_handler.py:154] [process=0][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.025411s I0419 00:25:14.076665 138465611859776 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/blocking_gbytes_per_sec: 51.458 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.029977083206176758 s) (per-host) I0419 00:25:14.076730 138465611859776 base_pytree_checkpoint_handler.py:768] [process=0][thread=MainThread] Initiated Pytree async_save. Time taken: 0.030059s (batch_requests_ready=0.001949s, total_serialization_initiated=0.027398s, others=0.000712s) I0419 00:25:14.076799 138465611859776 composite_checkpoint_handler.py:715] [process=0][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.034071s (all_items=0.000015s, per_item={'items': '0.00001502'}, temp_paths=0.034056) I0419 00:25:14.077537 138465611859776 event_tracking.py:125] [process=0] [async] Finished blocking save in 1.16 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9. I0419 00:25:14.077881 138332998715136 async_checkpointer.py:76] [process=0][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-19 00:45:14.077843 I0419 00:25:14.095707 138465611859776 checkpoint_manager.py:1560] [process=0][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0419 00:25:14.095991 138336760997632 async_checkpointer.py:280] [process=0][thread=save_finalize] Waiting for background save thread=async_save. I0419 00:25:14.096137 138465611859776 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776558289.4351704, 'wait_for_prev_duration_secs': 23.478254079818726, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776558312.9151888, 'checkpointer_blocking_duration_secs': 1.1628291606903076, 'get_old_steps_start_time': 1776558314.0780454, 'get_old_steps_duration_secs': 3.0517578125e-05, 'checkpoint_manager_blocking_start_time': 1776558289.4334662, 'checkpoint_manager_blocking_duration_secs': 24.662635326385498} I0419 00:25:14.096249 138465611859776 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0419 00:25:14.096295 138465611859776 checkpoint_manager.py:2020] [process=0][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0419 00:25:14.723850 138337289398016 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9 I0419 00:25:16.439332 138333015500544 atomicity.py:140] Creating tmp directory gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9/items I0419 00:25:20.139599 138336786175744 array_metadata_store.py:203] [process=0][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9/items/array_metadatas/process_0 I0419 00:25:59.529862 138336769390336 base_pytree_checkpoint_handler.py:1282] [process=0][thread=write_metadata_after_commits] Commit + Array metadata written. Time taken: 41.551337s (commit=40.474518s, array_metadata_write=1.076818s) I0419 00:25:59.530775 138332998715136 base_pytree_checkpoint_handler.py:130] [process=0] /jax/orbax/write/gbytes_per_sec: 34.729 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 45.48405957221985 s) (per-host) I0419 00:25:59.530912 138332998715136 async_checkpointer.py:90] [process=0][thread=async_save] 3 Handler Commit operations completed. Time taken: 45.452919s. I0419 00:26:01.762163 138332998715136 array_metadata_store.py:411] [process=0][thread=async_save] Validated ArrayMetadata from all 8 hosts. Time taken: 0.000242s. I0419 00:26:02.150508 138332998715136 ocdbt_utils.py:49] Param validation support for Zarr3 will be added later (b/362328389). I0419 00:26:03.089175 138332998715136 base_pytree_checkpoint_handler.py:1406] [process=0][thread=async_save] Pytree save finalize (merge_ocdbt + ArrayMetadata validation) completed. Time taken: 2.740046s. use_zarr3=True, enable_post_merge_validation=True, directory=gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9/items I0419 00:26:03.090726 138332998715136 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9/items I0419 00:26:04.040901 138332998715136 atomicity.py:666] Finalizing gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9 I0419 00:26:05.880054 138332998715136 atomicity.py:847] [process=0][thread=async_save] Finished saving checkpoint (finalized tmp dir) to `gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9`. I0419 00:26:05.880939 138332998715136 event_tracking.py:138] [process=0] [async] Finished save (blocking + background) in 52.97 seconds @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260418_235007/linen_xpk_feat_nnx_post_train_fixes_20260418_235007_10_shardy_false/checkpoints/9 I0419 00:26:05.882440 138332998715136 async_checkpointer.py:160] [process=0][thread=async_save] Background save thread done. Time taken: 51.804446s. I0419 00:26:05.882654 138336760997632 async_checkpointer.py:288] [process=0][thread=save_finalize] Done with waiting for background save thread=async_save. I0419 00:26:05.882772 138336760997632 async_checkpointer.py:298] [process=0][thread=save_finalize] No errors found in background save thread=async_save. I0419 00:26:05.882849 138336760997632 checkpoint_manager.py:2137] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0419 00:26:05.884917 138336760997632 checkpoint_manager.py:2146] [process=0][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0419 00:26:05.885056 138465611859776 checkpoint_manager.py:2032] [process=0][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0419 00:26:05.885145 138465611859776 checkpoint_manager.py:2009] [process=0][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0419 00:26:05.886036 138465611859776 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.273, Tokens/s/device: 13908.414, total_weights: 65536, loss: 8.179, lm_loss: 8.179, perplexity: 3564.635 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Sun Apr 19 00:26:20 UTC 2026 EXIT_CODE=0