XPK Start: Tue Apr 21 12:00:15 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-21 12:00:40.106225: 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) I0421 12:00:40.287274 140311354169152 max_utils.py:273] Attempting to initialize the jax distributed system... I0421 12:00:49.329053 140311354169152 distributed.py:149] Starting JAX distributed service on [::]:8482 I0421 12:00:49.331414 140311354169152 distributed.py:172] Connecting to JAX distributed service on mt-07-eval-5ezt6-slice-job-0-0.mt-07-eval-5ezt6:8482 I0421 12:00:50.574501 140311354169152 max_utils.py:284] Jax distributed system initialized! I0421 12:00:55.825539 140311354169152 max_utils.py:800] System Information: Jax Version: 0.9.2 I0421 12:00:55.825642 140311354169152 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0421 12:00:55.825681 140311354169152 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 I0421 12:00:55.825714 140311354169152 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0421 12:00:56.522783 140311354169152 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 12:00:56.523368 140311354169152 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0421 12:00:56.523545 140311354169152 checkpointing.py:688] Setting up checkpoint logger... I0421 12:00:56.523593 140311354169152 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0421 12:00:56.523636 140311354169152 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0421 12:00:56.523984 140311354169152 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 0x7f9c2cf1cb90>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0421 12:00:59.510885 140311354169152 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0421 12:00:59.511125 140311354169152 checkpoint_manager.py:708] [process=7][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f9b056c0e30>}, handler_registry=None I0421 12:00:59.511365 140311354169152 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f9b056c0e30>` 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`. I0421 12:00:59.511413 140311354169152 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f889045d190>` 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`. I0421 12:00:59.511448 140311354169152 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 0x7f9b056c0e30>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7f9b056c0e30>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f889045d190>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7f889045d190>}). I0421 12:00:59.511772 140311354169152 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0421 12:00:59.511844 140311354169152 async_checkpointer.py:192] [process=7][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7f878c774540> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0421 12:01:00.658938 140311354169152 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints I0421 12:01:01.114401 140311354169152 checkpoint_manager.py:929] [process=7][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_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7f889045f7a0> I0421 12:01:01.114569 140311354169152 checkpointing.py:302] Checkpoint manager created! I0421 12:01:02.068813 140311354169152 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0421 12:01:02.068916 140311354169152 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 12:01:02.446369 140311354169152 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 12:01:02.446458 140311354169152 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 12:01:02.462826 140311354169152 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0421 12:01:02.462884 140311354169152 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0421 12:01:02.486636 140311354169152 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 12:01:02.486706 140311354169152 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:01:02.503198 140311354169152 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 12:01:02.503260 140311354169152 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:01:02.519610 140311354169152 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0421 12:01:02.519672 140311354169152 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:01:02.544318 140311354169152 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0421 12:01:02.544391 140311354169152 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0421 12:01:02.565217 140311354169152 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0421 12:01:02.565280 140311354169152 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0421 12:01:02.779965 140311354169152 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0421 12:01:02.780071 140311354169152 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0421 12:01:04.202472 140311354169152 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0421 12:01:04.202601 140311354169152 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) I0421 12:01:04.202653 140311354169152 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) I0421 12:01:04.202709 140311354169152 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') I0421 12:01:04.202775 140311354169152 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) I0421 12:01:04.202815 140311354169152 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) I0421 12:01:04.202867 140311354169152 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) I0421 12:01:04.202917 140311354169152 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') I0421 12:01:04.202955 140311354169152 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) I0421 12:01:04.202995 140311354169152 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) I0421 12:01:04.203044 140311354169152 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0421 12:01:04.203091 140311354169152 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0421 12:01:04.695873 140311354169152 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 12:01:04.695967 140311354169152 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0421 12:01:04.711403 140311354169152 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0421 12:01:04.711463 140311354169152 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0421 12:01:15.720212 140311354169152 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. I0421 12:01:15.721064 140311354169152 metric_logger.py:301] number parameters: 1.104 billion I0421 12:01:26.989279 140311354169152 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0421 12:01:27.142927 140311354169152 checkpointing.py:798] Waited 0.1536252498626709 seconds for step 0 to finish before starting checkpointing. I0421 12:01:27.145345 140311354169152 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 12:01:27.147137 140311354169152 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 0 I0421 12:01:27.148715 140311354169152 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints/0. I0421 12:01:27.922751 140311354169152 signaling_client.py:364] Using JaxDistributedSignalingClient I0421 12:01:27.923720 140311354169152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 12:01:27.923801 140311354169152 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 I0421 12:01:28.210330 140311354169152 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.287666s I0421 12:01:28.210505 140311354169152 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 5.257 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.2934138774871826 s) (per-host) I0421 12:01:28.210561 140311354169152 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.293478s (batch_requests_ready=0.002332s, total_serialization_initiated=0.291074s, others=0.000072s) I0421 12:01:28.210656 140311354169152 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.297730s (all_items=0.000019s, per_item={'items': '0.00001884'}, temp_paths=0.297712) I0421 12:01:28.211464 140311354169152 event_tracking.py:125] [process=7] [async] Finished blocking save in 1.06 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints/0. I0421 12:01:28.211772 140185400403712 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 12:21:28.211725 I0421 12:01:28.213990 140311354169152 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0421 12:01:28.214216 140183831832320 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 12:01:28.214335 140311354169152 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776772887.1453273, 'wait_for_prev_duration_secs': 5.984306335449219e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776772887.1471753, 'checkpointer_blocking_duration_secs': 1.0647516250610352, 'get_old_steps_start_time': 1776772888.211952, 'get_old_steps_duration_secs': 3.123283386230469e-05, 'checkpoint_manager_blocking_start_time': 1776772887.1434314, 'checkpoint_manager_blocking_duration_secs': 1.0708670616149902} I0421 12:01:28.214449 140311354169152 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0421 12:01:28.214501 140311354169152 max_utils.py:750] Memstats: After params initialized: I0421 12:01:28.214544 140311354169152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_26(process=7,(2,6,0,0)) I0421 12:01:28.214575 140311354169152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_27(process=7,(3,6,0,0)) I0421 12:01:28.214602 140311354169152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_30(process=7,(2,7,0,0)) I0421 12:01:28.214625 140311354169152 max_utils.py:756] Using (GB) 0.43 / 31.25 (1.376000%) on TPU_31(process=7,(3,7,0,0)) I0421 12:01:28.524392 140311354169152 metric_logger.py:196] completed step: 0, seconds: 11.268, TFLOP/s/device: 1.206, Tokens/s/device: 181.752, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 12:01:28.708169 140311354169152 metric_logger.py:196] completed step: 1, seconds: 1.534, TFLOP/s/device: 8.860, Tokens/s/device: 1335.463, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52779.875 I0421 12:01:29.140673 140311354169152 metric_logger.py:196] completed step: 2, seconds: 0.036, TFLOP/s/device: 372.648, Tokens/s/device: 56169.606, total_weights: 65536, loss: 10.262, lm_loss: 10.262, perplexity: 28637.129 I0421 12:01:29.288123 140311354169152 metric_logger.py:196] completed step: 3, seconds: 0.433, TFLOP/s/device: 31.371, Tokens/s/device: 4728.569, total_weights: 65536, loss: 9.734, lm_loss: 9.734, perplexity: 16889.203 I0421 12:01:29.583133 140311354169152 metric_logger.py:196] completed step: 4, seconds: 0.153, TFLOP/s/device: 88.830, Tokens/s/device: 13389.384, total_weights: 65536, loss: 9.277, lm_loss: 9.277, perplexity: 10694.614 I0421 12:01:29.588993 140311354169152 metric_logger.py:196] completed step: 5, seconds: 0.147, TFLOP/s/device: 92.322, Tokens/s/device: 13915.785, total_weights: 65536, loss: 8.892, lm_loss: 8.892, perplexity: 7270.717 I0421 12:01:31.220103 2738 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0421 12:01:33.292897 140183840225024 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints/0/items/array_metadatas/process_7 I0421 12:01:51.348116 140311354169152 metric_logger.py:196] completed step: 6, seconds: 0.295, TFLOP/s/device: 45.987, Tokens/s/device: 6931.611, total_weights: 65536, loss: 8.592, lm_loss: 8.592, perplexity: 5390.759 I0421 12:01:51.495396 140311354169152 metric_logger.py:196] completed step: 7, seconds: 21.612, TFLOP/s/device: 0.629, Tokens/s/device: 94.760, total_weights: 65536, loss: 8.384, lm_loss: 8.384, perplexity: 4376.905 I0421 12:01:51.642832 140311354169152 metric_logger.py:196] completed step: 8, seconds: 0.152, TFLOP/s/device: 89.126, Tokens/s/device: 13434.089, total_weights: 65536, loss: 8.255, lm_loss: 8.255, perplexity: 3846.406 I0421 12:01:51.789504 140311354169152 checkpointing.py:794] Waiting for step 9 to finish before checkpoint... I0421 12:01:51.790194 140311354169152 checkpointing.py:798] Waited 0.0007114410400390625 seconds for step 9 to finish before starting checkpointing. I0421 12:01:51.792258 140311354169152 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 12:02:04.848594 140185400403712 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 42.771 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 36.93146347999573 s) (per-host) I0421 12:02:04.848721 140185400403712 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 36.636833s. I0421 12:02:12.808436 140185400403712 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 44.596533s. I0421 12:02:12.808675 140183831832320 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 12:02:12.808758 140183831832320 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 12:02:12.808814 140183831832320 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0421 12:02:12.810362 140183831832320 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0421 12:02:12.810543 140311354169152 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0421 12:02:12.810683 140311354169152 checkpoint_manager.py:1452] Waiting for previous save to complete took 21.018426 seconds. If this number is high, consider checkpointing less frequently. I0421 12:02:12.812727 140311354169152 checkpoint_manager.py:1512] [process=7] Saving checkpoint at step 9 I0421 12:02:12.814890 140311354169152 event_tracking.py:70] [process=7] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints/9. I0421 12:02:13.098700 140311354169152 jax_array_handlers.py:360] Scheduling D2H of 39 prioritized jax.Array. I0421 12:02:13.098808 140311354169152 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 I0421 12:02:13.135831 140311354169152 base_pytree_checkpoint_handler.py:154] [process=7][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.038237s I0421 12:02:13.135978 140311354169152 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/blocking_gbytes_per_sec: 36.956 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.041741371154785156 s) (per-host) I0421 12:02:13.136026 140311354169152 base_pytree_checkpoint_handler.py:768] [process=7][thread=MainThread] Initiated Pytree async_save. Time taken: 0.041798s (batch_requests_ready=0.001889s, total_serialization_initiated=0.039845s, others=0.000065s) I0421 12:02:13.136110 140311354169152 composite_checkpoint_handler.py:715] [process=7][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.045955s (all_items=0.000016s, per_item={'items': '0.00001597'}, temp_paths=0.045939) I0421 12:02:13.136835 140311354169152 event_tracking.py:125] [process=7] [async] Finished blocking save in 0.32 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints/9. I0421 12:02:13.137173 140183831832320 async_checkpointer.py:76] [process=7][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-21 12:22:13.137139 I0421 12:02:13.139169 140311354169152 checkpoint_manager.py:1560] [process=7][thread=MainThread][step=9] Starting CheckpointManager Save Finalize thread=save_finalize I0421 12:02:13.139473 140183840225024 async_checkpointer.py:280] [process=7][thread=save_finalize] Waiting for background save thread=async_save. I0421 12:02:13.139609 140311354169152 standard_logger.py:34] {'step': 9, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776772911.7922273, 'wait_for_prev_duration_secs': 21.018426179885864, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776772932.8127856, 'checkpointer_blocking_duration_secs': 0.32453131675720215, 'get_old_steps_start_time': 1776772933.1373417, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776772911.790417, 'checkpoint_manager_blocking_duration_secs': 21.349158763885498} I0421 12:02:13.139792 140311354169152 checkpointing.py:409] Started an asynchronous checkpoint save for step 9 I0421 12:02:13.139842 140311354169152 checkpoint_manager.py:2020] [process=7][thread=MainThread][step=9][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0421 12:02:18.891640 140180082128640 array_metadata_store.py:203] [process=7][thread=array_type_handler] Wrote 39 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_post_train_fixes_20260421_114057/linen_xpk_feat_nnx_post_train_fixes_20260421_114057_07_eval/checkpoints/9/items/array_metadatas/process_7 I0421 12:02:55.510540 140183831832320 base_pytree_checkpoint_handler.py:130] [process=7] /jax/orbax/write/gbytes_per_sec: 37.240 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.416260957717896 s) (per-host) I0421 12:02:55.510663 140183831832320 async_checkpointer.py:90] [process=7][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.373379s. I0421 12:03:03.946642 140183831832320 async_checkpointer.py:160] [process=7][thread=async_save] Background save thread done. Time taken: 50.809343s. I0421 12:03:03.946923 140183840225024 async_checkpointer.py:288] [process=7][thread=save_finalize] Done with waiting for background save thread=async_save. I0421 12:03:03.947062 140183840225024 async_checkpointer.py:298] [process=7][thread=save_finalize] No errors found in background save thread=async_save. I0421 12:03:03.947111 140183840225024 checkpoint_manager.py:2137] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is syncing with other hosts... I0421 12:03:03.948774 140183840225024 checkpoint_manager.py:2146] [process=7][thread=save_finalize][step=9] CheckpointManager Save Finalize is done on all hosts. I0421 12:03:03.948958 140311354169152 checkpoint_manager.py:2032] [process=7][thread=MainThread][step=9][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=9. I0421 12:03:03.949120 140311354169152 checkpoint_manager.py:2009] [process=7][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0421 12:03:03.950119 140311354169152 metric_logger.py:196] completed step: 9, seconds: 0.147, TFLOP/s/device: 92.382, Tokens/s/device: 13924.774, 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: Tue Apr 21 12:03:16 UTC 2026 EXIT_CODE=0