XPK Start: Fri Apr 17 04:00:15 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-17 04:00:39.565064: 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) I0417 04:00:40.104419 136417637324608 max_utils.py:273] Attempting to initialize the jax distributed system... I0417 04:00:49.146299 136417637324608 distributed.py:149] Starting JAX distributed service on [::]:8482 I0417 04:00:49.148725 136417637324608 distributed.py:172] Connecting to JAX distributed service on mt-15-ocdbt-true-sv-6d2fk-slice-job-0-0.mt-15-ocdbt-true-sv-6d2fk:8482 I0417 04:00:51.045853 136417637324608 max_utils.py:284] Jax distributed system initialized! I0417 04:00:57.056574 136417637324608 max_utils.py:800] System Information: Jax Version: 0.9.2 I0417 04:00:57.056673 136417637324608 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0417 04:00:57.056712 136417637324608 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 I0417 04:00:57.056746 136417637324608 train_utils.py:377] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0417 04:00:57.735879 136417637324608 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 04:00:57.736448 136417637324608 maxtext_utils.py:1687] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0417 04:00:57.736637 136417637324608 checkpointing.py:688] Setting up checkpoint logger... I0417 04:00:57.736690 136417637324608 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0417 04:00:57.736733 136417637324608 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0417 04:00:57.737074 136417637324608 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 0x7c117d250c50>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0417 04:01:00.743104 136417637324608 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0417 04:01:00.743332 136417637324608 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 0x7bfd483849b0>}, handler_registry=None I0417 04:01:00.743579 136417637324608 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bfd483849b0>` 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`. I0417 04:01:00.743630 136417637324608 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bfd48387b30>` 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`. I0417 04:01:00.743667 136417637324608 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 0x7bfd483849b0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7bfd483849b0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bfd48387b30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7bfd48387b30>}). I0417 04:01:00.743978 136417637324608 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0417 04:01:00.744052 136417637324608 async_checkpointer.py:192] [process=1][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7bfd48104680> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0417 04:01:02.169592 136417637324608 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints I0417 04:01:02.621797 136417637324608 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=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_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7bfd48385e20> I0417 04:01:02.621963 136417637324608 checkpointing.py:302] Checkpoint manager created! I0417 04:01:03.556607 136417637324608 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0417 04:01:03.556704 136417637324608 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:01:03.936373 136417637324608 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 04:01:03.936471 136417637324608 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:01:03.952591 136417637324608 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0417 04:01:03.952648 136417637324608 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0417 04:01:03.976129 136417637324608 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 04:01:03.976195 136417637324608 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:01:03.992372 136417637324608 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 04:01:03.992429 136417637324608 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:01:04.008611 136417637324608 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0417 04:01:04.008671 136417637324608 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:01:04.034348 136417637324608 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0417 04:01:04.034415 136417637324608 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0417 04:01:04.055202 136417637324608 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0417 04:01:04.055269 136417637324608 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0417 04:01:04.270995 136417637324608 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0417 04:01:04.271103 136417637324608 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0417 04:01:05.689985 136417637324608 maxtext_utils.py:1796] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0417 04:01:05.690113 136417637324608 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_0/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 04:01:05.690166 136417637324608 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wi_1/kernel Shape: float32[2048,16,7168] Logical: P('embed', 'layers', 'mlp') Physical: ('fsdp', None, None) I0417 04:01:05.690225 136417637324608 maxtext_utils.py:1796] params/params/decoder/layers/mlp/wo/kernel Shape: float32[7168,16,2048] Logical: P('mlp', 'layers', 'embed') Physical: (None, None, 'fsdp') I0417 04:01:05.690286 136417637324608 maxtext_utils.py:1796] params/params/decoder/layers/post_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 04:01:05.690327 136417637324608 maxtext_utils.py:1796] params/params/decoder/layers/pre_self_attention_layer_norm/scale Shape: float32[2048,16] Logical: P('norm', 'layers') Physical: (None, None) I0417 04:01:05.690378 136417637324608 maxtext_utils.py:1796] 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) I0417 04:01:05.690435 136417637324608 maxtext_utils.py:1796] 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') I0417 04:01:05.690647 136417637324608 maxtext_utils.py:1796] 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) I0417 04:01:05.690717 136417637324608 maxtext_utils.py:1796] 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) I0417 04:01:05.690783 136417637324608 maxtext_utils.py:1796] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0417 04:01:05.690845 136417637324608 maxtext_utils.py:1796] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0417 04:01:06.199694 136417637324608 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 04:01:06.199787 136417637324608 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0417 04:01:06.215319 136417637324608 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0417 04:01:06.215377 136417637324608 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0417 04:01:17.067318 136417637324608 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. I0417 04:01:17.068153 136417637324608 metric_logger.py:289] number parameters: 1.104 billion I0417 04:01:28.537139 136417637324608 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0417 04:01:28.690479 136417637324608 checkpointing.py:798] Waited 0.1533064842224121 seconds for step 0 to finish before starting checkpointing. I0417 04:01:28.692895 136417637324608 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 04:01:28.694632 136417637324608 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 0 I0417 04:01:28.696059 136417637324608 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/0. I0417 04:01:29.045392 136417637324608 signaling_client.py:364] Using JaxDistributedSignalingClient I0417 04:01:29.046298 136417637324608 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:01:29.046353 136417637324608 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 I0417 04:01:29.320343 136417637324608 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.275023s I0417 04:01:29.320532 136417637324608 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 5.496 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.28066444396972656 s) (per-host) I0417 04:01:29.320601 136417637324608 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.280744s (batch_requests_ready=0.002315s, total_serialization_initiated=0.278341s, others=0.000088s) I0417 04:01:29.320704 136417637324608 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.284906s (all_items=0.000017s, per_item={'items': '0.00001717'}, temp_paths=0.284889) I0417 04:01:29.321498 136417637324608 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.63 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/0. I0417 04:01:29.321827 136289143691008 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:21:29.321789 I0417 04:01:29.330183 136417637324608 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:01:29.330431 136287558297344 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:01:29.330571 136417637324608 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776398488.6928778, 'wait_for_prev_duration_secs': 5.936622619628906e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776398488.6946716, 'checkpointer_blocking_duration_secs': 0.6273107528686523, 'get_old_steps_start_time': 1776398489.3220072, 'get_old_steps_duration_secs': 3.075599670410156e-05, 'checkpoint_manager_blocking_start_time': 1776398488.6910098, 'checkpoint_manager_blocking_duration_secs': 0.6395225524902344} I0417 04:01:29.330673 136417637324608 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0417 04:01:29.330722 136417637324608 max_utils.py:750] Memstats: After params initialized: I0417 04:01:29.330768 136417637324608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_2(process=1,(2,0,0,0)) I0417 04:01:29.330801 136417637324608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_3(process=1,(3,0,0,0)) I0417 04:01:29.330827 136417637324608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_6(process=1,(2,1,0,0)) I0417 04:01:29.330852 136417637324608 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_7(process=1,(3,1,0,0)) I0417 04:01:29.645181 136417637324608 metric_logger.py:185] completed step: 0, seconds: 11.469, TFLOP/s/device: 1.185, Tokens/s/device: 178.570, total_weights: 65536, loss: 10.874 I0417 04:01:29.811423 136417637324608 metric_logger.py:185] completed step: 1, seconds: 1.107, TFLOP/s/device: 12.277, Tokens/s/device: 1850.460, total_weights: 65536, loss: 10.874 I0417 04:01:30.231394 136417637324608 metric_logger.py:185] completed step: 2, seconds: 0.019, TFLOP/s/device: 709.844, Tokens/s/device: 106995.455, total_weights: 65536, loss: 10.560 I0417 04:01:30.378535 136417637324608 metric_logger.py:185] completed step: 3, seconds: 0.420, TFLOP/s/device: 32.353, Tokens/s/device: 4876.527, total_weights: 65536, loss: 9.980 I0417 04:01:30.382897 136417637324608 checkpointing.py:794] Waiting for step 5 to finish before checkpoint... I0417 04:01:30.672919 136417637324608 checkpointing.py:798] Waited 0.289992094039917 seconds for step 5 to finish before starting checkpointing. I0417 04:01:30.675963 136417637324608 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:01:33.162167 2847 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0417 04:01:35.627288 136287566690048 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/0/items/array_metadatas/process_1 I0417 04:02:07.248474 136289143691008 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 41.341 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 38.20855736732483 s) (per-host) I0417 04:02:07.248582 136289143691008 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 37.926641s. I0417 04:02:15.994893 136289143691008 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 46.672937s. I0417 04:02:15.995164 136287558297344 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:02:15.995284 136287558297344 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:02:15.995346 136287558297344 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:02:15.998211 136287558297344 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0417 04:02:15.998406 136417637324608 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. W0417 04:02:15.998559 136417637324608 checkpoint_manager.py:1452] Waiting for previous save to complete took 45.322594 seconds. If this number is high, consider checkpointing less frequently. I0417 04:02:16.000555 136417637324608 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 5 I0417 04:02:16.002889 136417637324608 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/5. I0417 04:02:16.732321 136417637324608 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:02:16.732414 136417637324608 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 I0417 04:02:16.768344 136417637324608 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.037048s I0417 04:02:16.768527 136417637324608 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 37.879 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.040723562240600586 s) (per-host) I0417 04:02:16.768579 136417637324608 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.040786s (batch_requests_ready=0.001941s, total_serialization_initiated=0.038775s, others=0.000070s) I0417 04:02:16.768669 136417637324608 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045959s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.045943) I0417 04:02:16.769374 136417637324608 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.77 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/5. I0417 04:02:16.769686 136287558297344 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:22:16.769654 I0417 04:02:16.774762 136417637324608 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=5] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:02:16.775003 136267048138496 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:02:16.775130 136417637324608 standard_logger.py:34] {'step': 5, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776398490.6759353, 'wait_for_prev_duration_secs': 45.322594165802, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776398536.0005927, 'checkpointer_blocking_duration_secs': 0.7692379951477051, 'get_old_steps_start_time': 1776398536.7698553, 'get_old_steps_duration_secs': 3.147125244140625e-05, 'checkpoint_manager_blocking_start_time': 1776398490.673541, 'checkpoint_manager_blocking_duration_secs': 46.101553201675415} I0417 04:02:16.775233 136417637324608 checkpointing.py:409] Started an asynchronous checkpoint save for step 5 I0417 04:02:16.775824 136417637324608 metric_logger.py:185] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.965, Tokens/s/device: 13409.811, total_weights: 65536, loss: 9.460 I0417 04:02:16.781588 136417637324608 metric_logger.py:185] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.393, Tokens/s/device: 13926.478, total_weights: 65536, loss: 8.959 I0417 04:02:23.542072 136287566690048 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/5/items/array_metadatas/process_1 I0417 04:02:40.474440 136417637324608 metric_logger.py:185] completed step: 6, seconds: 46.397, TFLOP/s/device: 0.293, Tokens/s/device: 44.140, total_weights: 65536, loss: 8.469 I0417 04:02:40.621910 136417637324608 metric_logger.py:185] completed step: 7, seconds: 23.547, TFLOP/s/device: 0.577, Tokens/s/device: 86.976, total_weights: 65536, loss: 8.003 I0417 04:02:40.769378 136417637324608 metric_logger.py:185] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.431, Tokens/s/device: 13480.070, total_weights: 65536, loss: 7.572 I0417 04:02:40.916086 136417637324608 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0417 04:02:40.916764 136417637324608 checkpointing.py:798] Waited 0.0006973743438720703 seconds for step 9 to finish before starting checkpointing. I0417 04:02:40.918767 136417637324608 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=5][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:02:58.848822 136287558297344 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 37.501 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.120988845825195 s) (per-host) I0417 04:02:58.848928 136287558297344 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.079131s. I0417 04:03:07.628613 136287558297344 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 50.858801s. I0417 04:03:07.628820 136267048138496 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:03:07.628875 136267048138496 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:03:07.628921 136267048138496 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=5] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:03:07.630803 136267048138496 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=5] CheckpointManager Save Finalize is done on all hosts. I0417 04:03:07.630958 136417637324608 checkpoint_manager.py:2032] [process=1][thread=MainThread][step=5][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=5. W0417 04:03:07.631085 136417637324608 checkpoint_manager.py:1452] Waiting for previous save to complete took 26.712325 seconds. If this number is high, consider checkpointing less frequently. I0417 04:03:07.632569 136417637324608 checkpoint_manager.py:1512] [process=1] Saving checkpoint at step 9 I0417 04:03:07.634612 136417637324608 event_tracking.py:70] [process=1] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/9. I0417 04:03:08.345052 136417637324608 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0417 04:03:08.345145 136417637324608 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 I0417 04:03:08.378847 136417637324608 base_pytree_checkpoint_handler.py:154] [process=1][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.034605s I0417 04:03:08.379011 136417637324608 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/blocking_gbytes_per_sec: 40.453 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.0381321907043457 s) (per-host) I0417 04:03:08.379061 136417637324608 base_pytree_checkpoint_handler.py:768] [process=1][thread=MainThread] Initiated Pytree async_save. Time taken: 0.038193s (batch_requests_ready=0.001755s, total_serialization_initiated=0.036373s, others=0.000065s) I0417 04:03:08.379151 136417637324608 composite_checkpoint_handler.py:715] [process=1][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.042494s (all_items=0.000010s, per_item={'items': '0.00001001'}, temp_paths=0.042484) I0417 04:03:08.379880 136417637324608 event_tracking.py:125] [process=1] [async] Finished blocking save in 0.75 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/9. I0417 04:03:08.380192 136267048138496 async_checkpointer.py:76] [process=1][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-17 04:23:08.380157 I0417 04:03:08.382076 136417637324608 checkpoint_manager.py:1560] [process=1][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0417 04:03:08.382367 136287566690048 async_checkpointer.py:280] [process=1][thread=save_finalize] Waiting for background save thread=async_save. I0417 04:03:08.382513 136417637324608 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776398560.9187376, 'wait_for_prev_duration_secs': 26.71232509613037, 'time_between_consecutive_saves_sec': 24.92047953605652, 'checkpointer_blocking_start_time': 1776398587.6326082, 'checkpointer_blocking_duration_secs': 0.7477235794067383, 'get_old_steps_start_time': 1776398588.3803499, 'get_old_steps_duration_secs': 2.5272369384765625e-05, 'checkpoint_manager_blocking_start_time': 1776398560.9169893, 'checkpoint_manager_blocking_duration_secs': 27.465491771697998} I0417 04:03:08.382673 136417637324608 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0417 04:03:08.382715 136417637324608 checkpoint_manager.py:2020] [process=1][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0417 04:03:14.348124 136286987855616 array_metadata_store.py:203] [process=1][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260417_023710/linen_xpk_feat_nnx_post_train_fixes_20260417_023710_15_ocdbt_true/checkpoints/9/items/array_metadatas/process_1 I0417 04:03:50.077162 136267048138496 base_pytree_checkpoint_handler.py:130] [process=1] /jax/orbax/write/gbytes_per_sec: 37.847 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 41.73625946044922 s) (per-host) I0417 04:03:50.077265 136267048138496 async_checkpointer.py:90] [process=1][thread=async_save] 3 Handler Commit operations completed. Time taken: 41.696961s. I0417 04:03:58.683149 136267048138496 async_checkpointer.py:160] [process=1][thread=async_save] Background save thread done. Time taken: 50.302830s. I0417 04:03:58.683483 136287566690048 async_checkpointer.py:288] [process=1][thread=save_finalize] Done with waiting for background save thread=async_save. I0417 04:03:58.683606 136287566690048 async_checkpointer.py:298] [process=1][thread=save_finalize] No errors found in background save thread=async_save. I0417 04:03:58.683649 136287566690048 checkpoint_manager.py:2137] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0417 04:03:58.685144 136287566690048 checkpoint_manager.py:2146] [process=1][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0417 04:03:58.685318 136417637324608 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. I0417 04:03:58.685474 136417637324608 checkpoint_manager.py:2009] [process=1][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0417 04:03:58.686160 136417637324608 metric_logger.py:185] completed step: 9, seconds: 0.148, TFLOP/s/device: 92.099, Tokens/s/device: 13882.205, total_weights: 65536, loss: 7.181 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops error: stream error: stream ID 3; INTERNAL_ERROR; received from peer