XPK Start: Mon Apr 20 12:26:58 UTC 2026 PyTorch was not found. Models won't be available and only tokenizers, configuration and file/data utilities can be used. 2026-04-20 12:27:23.110397: E external/local_xla/xla/stream_executor/cuda/cuda_platform.cc:51] failed call to cuInit: INTERNAL: CUDA error: Failed call to cuInit: UNKNOWN ERROR (303) I0420 12:27:23.290267 137741956515648 max_utils.py:273] Attempting to initialize the jax distributed system... I0420 12:27:32.329923 137741956515648 distributed.py:149] Starting JAX distributed service on [::]:8482 I0420 12:27:32.332256 137741956515648 distributed.py:172] Connecting to JAX distributed service on mt-05-fp8-mqgsj-slice-job-0-0.mt-05-fp8-mqgsj:8482 I0420 12:27:34.058945 137741956515648 max_utils.py:284] Jax distributed system initialized! I0420 12:27:40.294900 137741956515648 max_utils.py:800] System Information: Jax Version: 0.9.2 I0420 12:27:40.295000 137741956515648 max_utils.py:801] System Information: Jaxlib Version: 0.9.2 I0420 12:27:40.295039 137741956515648 max_utils.py:802] System Information: Jax Backend: PJRT C API TFRT TPU v6 lite Built on Mar 4 2026 11:32:08 (1772652728) cl/878335365 I0420 12:27:40.295073 137741956515648 train_utils.py:378] WARNING: Sequence packing is essentially ignored for synthetic data. Please use a real dataset to use sequence packing. I0420 12:27:41.342352 137741956515648 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 12:27:41.342927 137741956515648 maxtext_utils.py:1718] Num_devices: 32, shape (1, 1, 1, 32, 1, 1, 1, 1, 1, 1, 1, 1, 1) I0420 12:27:41.343106 137741956515648 checkpointing.py:688] Setting up checkpoint logger... I0420 12:27:41.343173 137741956515648 checkpointing.py:234] Creating checkpoint manager with ocdbt=True and zarr3=True I0420 12:27:41.343218 137741956515648 pytree_checkpoint_handler.py:592] save_device_host_concurrent_bytes=None I0420 12:27:41.343549 137741956515648 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 0x7d45f837a600>, enable_pinned_host_transfer=False, save_concurrent_bytes: 96000000000 (89.4 GiB), restore_concurrent_bytes: 96000000000 (89.4 GiB) I0420 12:27:43.932736 137741956515648 checkpointing.py:266] Enabling policy for fixed interval checkpointing. I0420 12:27:43.932925 137741956515648 checkpoint_manager.py:708] [process=3][thread=MainThread] CheckpointManager init: checkpointers=None, item_names=('items',), item_handlers={'items': <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d44c8773ec0>}, handler_registry=None I0420 12:27:43.933172 137741956515648 composite_checkpoint_handler.py:237] Deferred registration for item: "items". Adding handler `<orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d44c8773ec0>` for item "items" and save args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>` to `_handler_registry`. I0420 12:27:43.933221 137741956515648 composite_checkpoint_handler.py:237] Deferred registration for item: "metrics". Adding handler `<orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d44c8773770>` for item "metrics" and save args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>` and restore args `<class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>` to `_handler_registry`. I0420 12:27:43.933257 137741956515648 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 0x7d44c8773ec0>, ('items', <class 'orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeRestoreArgs'>): <orbax.checkpoint._src.handlers.pytree_checkpoint_handler.PyTreeCheckpointHandler object at 0x7d44c8773ec0>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonSaveArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d44c8773770>, ('metrics', <class 'orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonRestoreArgs'>): <orbax.checkpoint._src.handlers.json_checkpoint_handler.JsonCheckpointHandler object at 0x7d44c8773770>}). I0420 12:27:43.933574 137741956515648 abstract_checkpointer.py:35] orbax-checkpoint version: 0.11.34 I0420 12:27:43.933643 137741956515648 async_checkpointer.py:192] [process=3][thread=MainThread] Using barrier_sync_fn: <function get_barrier_sync_fn.<locals>._fn at 0x7d309c4dad40> timeout: 1200 secs and primary_host=0 for async checkpoint writes I0420 12:27:45.077701 137741956515648 checkpoint_manager.py:1812] Found 0 checkpoint steps in gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints I0420 12:27:45.500188 137741956515648 checkpoint_manager.py:929] [process=3][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_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints: <orbax.checkpoint.checkpoint_manager.CheckpointManager object at 0x7d44c87719a0> I0420 12:27:45.500361 137741956515648 checkpointing.py:302] Checkpoint manager created! I0420 12:27:46.590895 137741956515648 nnx_wrappers.py:437] Unknown Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_norm_length', 'activation_embed'). I0420 12:27:46.591004 137741956515648 nnx_wrappers.py:437] Unknown Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 12:27:46.972558 137741956515648 attentions.py:1088] attentions/inputs_q Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 12:27:46.972652 137741956515648 attentions.py:1088] attentions/inputs_q Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 12:27:46.989038 137741956515648 attentions.py:1089] attentions/inputs_kv Logical: bfloat16[32,2048,2048]...................................... ('activation_batch', 'activation_attn_length', 'activation_attn_embed'). I0420 12:27:46.989097 137741956515648 attentions.py:1089] attentions/inputs_kv Physical: bfloat16[32,2048,2048]...................................... ('fsdp', None, None). I0420 12:27:47.041624 137741956515648 attentions.py:1154] attentions/query Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 12:27:47.041714 137741956515648 attentions.py:1154] attentions/query Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:27:47.058457 137741956515648 attentions.py:1155] attentions/key Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 12:27:47.058519 137741956515648 attentions.py:1155] attentions/key Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:27:47.075327 137741956515648 attentions.py:1156] attentions/value Logical: bfloat16[32,2048,16,128].................................... ('activation_kv_batch', 'activation_attn_length', 'activation_kv_heads', 'activation_kv_head_dim'). I0420 12:27:47.075386 137741956515648 attentions.py:1156] attentions/value Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:27:47.100612 137741956515648 attentions.py:1197] attentions/out Logical: bfloat16[32,2048,16,128].................................... ('activation_batch', 'activation_attn_length', 'activation_heads', 'activation_kv'). I0420 12:27:47.100682 137741956515648 attentions.py:1197] attentions/out Physical: bfloat16[32,2048,16,128].................................... ('fsdp', None, None, None). I0420 12:27:47.152323 137741956515648 linears.py:525] linears/x Logical: bfloat16[32,2048,7168]...................................... ('activation_batch', 'activation_length', 'activation_mlp'). I0420 12:27:47.152402 137741956515648 linears.py:525] linears/x Physical: bfloat16[32,2048,7168]...................................... ('fsdp', None, None). I0420 12:27:47.594454 137741956515648 checkpointing.py:578] checkpoint manager exists so trying to load this run's existing checkpoint I0420 12:27:47.594565 137741956515648 checkpointing.py:676] No existing checkpoints found, not restoring checkpoint. fsdp: 32 I0420 12:27:49.515409 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515527 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515575 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515616 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515653 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515691 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_0/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515721 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515750 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515781 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515809 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515838 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515867 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wi_1/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515896 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515923 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.515951 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.515979 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516007 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516036 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/mlp/wo/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516063 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516090 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516117 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516157 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516187 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516214 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/key/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516240 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516267 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516293 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516319 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516345 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516372 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/out/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516399 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516425 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516451 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516477 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516503 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516530 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/query/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516556 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516582 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/input_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516608 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516634 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/kernel_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516660 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_amax_history Shape: float32[16,1024] Logical: P() Physical: () I0420 12:27:49.516691 137741956515648 maxtext_utils.py:1827] params/_overwrite_with_gradient/decoder/layers/self_attention/value/Fp8DirectDotGeneralOp_0/output_grad_scale Shape: float32[16,1] Logical: P() Physical: () I0420 12:27:49.516740 137741956515648 maxtext_utils.py:1827] params/params/decoder/decoder_norm/scale Shape: float32[2048] Logical: P('norm',) Physical: (None,) I0420 12:27:49.516792 137741956515648 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) I0420 12:27:49.516828 137741956515648 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) I0420 12:27:49.516871 137741956515648 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') I0420 12:27:49.516914 137741956515648 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) I0420 12:27:49.516945 137741956515648 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) I0420 12:27:49.516990 137741956515648 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) I0420 12:27:49.517035 137741956515648 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') I0420 12:27:49.517070 137741956515648 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) I0420 12:27:49.517102 137741956515648 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) I0420 12:27:49.517170 137741956515648 maxtext_utils.py:1827] params/params/decoder/logits_dense/kernel Shape: float32[2048,32000] Logical: P('embed_vocab', 'vocab') Physical: ('fsdp', None) I0420 12:27:49.517221 137741956515648 maxtext_utils.py:1827] params/params/token_embedder/embedding Shape: float32[32000,2048] Logical: P('vocab', 'embed_vocab') Physical: (None, 'fsdp') I0420 12:27:51.145033 137741956515648 train.py:157] train/xent Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 12:27:51.145131 137741956515648 train.py:157] train/xent Physical: float32[32,2048]............................................ ('fsdp', None). I0420 12:27:51.160989 137741956515648 train.py:164] train/z_loss Logical: float32[32,2048]............................................ ('activation_embed_and_logits_batch', 'activation_length'). I0420 12:27:51.161050 137741956515648 train.py:164] train/z_loss Physical: float32[32,2048]............................................ ('fsdp', None). I0420 12:28:04.816670 137741956515648 max_utils.py:791] Total memory size: 1.5 GB, Output size: 0.4 GB, Temp size: 1.1 GB, Argument size: 0.4 GB, Host temp size: 0.0 GB. I0420 12:28:04.817523 137741956515648 metric_logger.py:301] number parameters: 1.105 billion I0420 12:28:20.335092 137741956515648 checkpointing.py:794] Waiting for step 0 to finish before checkpoint... I0420 12:28:20.502082 137741956515648 checkpointing.py:798] Waited 0.16697120666503906 seconds for step 0 to finish before starting checkpointing. I0420 12:28:20.504466 137741956515648 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 12:28:20.506534 137741956515648 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 0 I0420 12:28:20.507956 137741956515648 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/0. I0420 12:28:21.284605 137741956515648 signaling_client.py:364] Using JaxDistributedSignalingClient I0420 12:28:21.285566 137741956515648 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array. I0420 12:28:21.285680 137741956515648 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 12:28:21.708012 137741956515648 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.423528s I0420 12:28:21.708197 137741956515648 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 3.573 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.4318065643310547 s) (per-host) I0420 12:28:21.708253 137741956515648 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.431872s (batch_requests_ready=0.003400s, total_serialization_initiated=0.428398s, others=0.000074s) I0420 12:28:21.708360 137741956515648 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.436128s (all_items=0.000019s, per_item={'items': '0.00001860'}, temp_paths=0.436109) I0420 12:28:21.709180 137741956515648 event_tracking.py:125] [process=3] [async] Finished blocking save in 1.20 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/0. I0420 12:28:21.709532 137611396523776 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 12:48:21.709494 I0420 12:28:21.712318 137741956515648 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=0] Starting CheckpointManager Save Finalize thread=save_finalize I0420 12:28:21.712654 137610862708480 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0420 12:28:21.712798 137741956515648 standard_logger.py:34] {'step': 0, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776688100.5044472, 'wait_for_prev_duration_secs': 6.4849853515625e-05, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776688100.506573, 'checkpointer_blocking_duration_secs': 1.2031233310699463, 'get_old_steps_start_time': 1776688101.7097232, 'get_old_steps_duration_secs': 2.9325485229492188e-05, 'checkpoint_manager_blocking_start_time': 1776688100.5027053, 'checkpoint_manager_blocking_duration_secs': 1.2100532054901123} I0420 12:28:21.712970 137741956515648 checkpointing.py:409] Started an asynchronous checkpoint save for step 0 I0420 12:28:21.713023 137741956515648 max_utils.py:750] Memstats: After params initialized: I0420 12:28:21.713073 137741956515648 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_10(process=3,(2,2,0,0)) I0420 12:28:21.713106 137741956515648 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_11(process=3,(3,2,0,0)) I0420 12:28:21.713132 137741956515648 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_14(process=3,(2,3,0,0)) I0420 12:28:21.713170 137741956515648 max_utils.py:756] Using (GB) 0.45 / 31.25 (1.440000%) on TPU_15(process=3,(3,3,0,0)) I0420 12:28:22.022707 137741956515648 metric_logger.py:196] completed step: 0, seconds: 15.517, TFLOP/s/device: 0.876, Tokens/s/device: 131.980, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52776.805 I0420 12:28:22.230832 137741956515648 metric_logger.py:196] completed step: 1, seconds: 1.686, TFLOP/s/device: 8.061, Tokens/s/device: 1215.032, total_weights: 65536, loss: 10.874, lm_loss: 10.874, perplexity: 52761.840 I0420 12:28:22.639661 137741956515648 metric_logger.py:196] completed step: 2, seconds: 0.051, TFLOP/s/device: 267.532, Tokens/s/device: 40325.280, total_weights: 65536, loss: 10.816, lm_loss: 10.816, perplexity: 49832.398 I0420 12:28:22.798040 137741956515648 metric_logger.py:196] completed step: 3, seconds: 0.409, TFLOP/s/device: 33.239, Tokens/s/device: 5010.140, total_weights: 65536, loss: 10.431, lm_loss: 10.431, perplexity: 33901.820 I0420 12:28:23.115325 137741956515648 metric_logger.py:196] completed step: 4, seconds: 0.164, TFLOP/s/device: 82.853, Tokens/s/device: 12488.566, total_weights: 65536, loss: 9.992, lm_loss: 9.992, perplexity: 21847.639 I0420 12:28:23.122093 137741956515648 metric_logger.py:196] completed step: 5, seconds: 0.158, TFLOP/s/device: 85.949, Tokens/s/device: 12955.220, total_weights: 65536, loss: 9.549, lm_loss: 9.549, perplexity: 14035.006 I0420 12:28:24.599373 2827 google_auth_provider.cc:181] Running on GCE, using service account 562977990677-compute@developer.gserviceaccount.com I0420 12:28:26.710280 137610871101184 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/0/items/array_metadatas/process_3 I0420 12:28:41.493805 137611396523776 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 78.139 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 20.217388153076172 s) (per-host) I0420 12:28:41.493911 137611396523776 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 19.784266s. I0420 12:28:47.300932 137741956515648 metric_logger.py:196] completed step: 6, seconds: 0.318, TFLOP/s/device: 42.667, Tokens/s/device: 6431.171, total_weights: 65536, loss: 9.111, lm_loss: 9.111, perplexity: 9058.688 I0420 12:28:47.459362 137741956515648 metric_logger.py:196] completed step: 7, seconds: 24.022, TFLOP/s/device: 0.566, Tokens/s/device: 85.256, total_weights: 65536, loss: 8.685, lm_loss: 8.685, perplexity: 5914.720 I0420 12:28:47.617876 137741956515648 metric_logger.py:196] completed step: 8, seconds: 0.163, TFLOP/s/device: 83.213, Tokens/s/device: 12542.718, total_weights: 65536, loss: 8.281, lm_loss: 8.281, perplexity: 3950.010 I0420 12:28:47.622708 137741956515648 checkpointing.py:794] Waiting for step 10 to finish before checkpoint... I0420 12:28:47.935034 137741956515648 checkpointing.py:798] Waited 0.3122901916503906 seconds for step 10 to finish before starting checkpointing. I0420 12:28:47.937951 137741956515648 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=0][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 12:28:50.172399 137611396523776 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 28.462737s. I0420 12:28:50.172698 137610862708480 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 12:28:50.172825 137610862708480 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0420 12:28:50.172876 137610862708480 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is syncing with other hosts... I0420 12:28:50.175566 137610862708480 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=0] CheckpointManager Save Finalize is done on all hosts. I0420 12:28:50.175694 137741956515648 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=0][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=0. W0420 12:28:50.175790 137741956515648 checkpoint_manager.py:1452] Waiting for previous save to complete took 2.237842 seconds. If this number is high, consider checkpointing less frequently. I0420 12:28:50.177567 137741956515648 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 10 I0420 12:28:50.179559 137741956515648 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/10. I0420 12:28:50.470172 137741956515648 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array. I0420 12:28:50.470268 137741956515648 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 12:28:50.527590 137741956515648 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.058362s I0420 12:28:50.527756 137741956515648 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 23.932 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.06446385383605957 s) (per-host) I0420 12:28:50.527804 137741956515648 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.064523s (batch_requests_ready=0.003034s, total_serialization_initiated=0.061423s, others=0.000066s) I0420 12:28:50.527908 137741956515648 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.069396s (all_items=0.000016s, per_item={'items': '0.00001645'}, temp_paths=0.069379) I0420 12:28:50.528680 137741956515648 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/10. I0420 12:28:50.529011 137610862708480 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 12:48:50.528978 I0420 12:28:50.530987 137741956515648 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=10] Starting CheckpointManager Save Finalize thread=save_finalize I0420 12:28:50.531302 137609762281216 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0420 12:28:50.531434 137741956515648 standard_logger.py:34] {'step': 10, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776688127.9379213, 'wait_for_prev_duration_secs': 2.2378416061401367, 'time_between_consecutive_saves_sec': None, 'checkpointer_blocking_start_time': 1776688130.1776056, 'checkpointer_blocking_duration_secs': 0.35155582427978516, 'get_old_steps_start_time': 1776688130.5291862, 'get_old_steps_duration_secs': 3.0994415283203125e-05, 'checkpoint_manager_blocking_start_time': 1776688127.935756, 'checkpoint_manager_blocking_duration_secs': 2.595644950866699} I0420 12:28:50.531608 137741956515648 checkpointing.py:409] Started an asynchronous checkpoint save for step 10 I0420 12:28:50.532217 137741956515648 metric_logger.py:196] completed step: 9, seconds: 0.158, TFLOP/s/device: 85.841, Tokens/s/device: 12938.850, total_weights: 65536, loss: 7.908, lm_loss: 7.908, perplexity: 2720.116 I0420 12:28:50.539426 137741956515648 metric_logger.py:196] completed step: 10, seconds: 0.159, TFLOP/s/device: 85.719, Tokens/s/device: 12920.565, total_weights: 65536, loss: 7.568, lm_loss: 7.568, perplexity: 1936.198 I0420 12:28:50.697938 137741956515648 metric_logger.py:196] completed step: 11, seconds: 2.914, TFLOP/s/device: 4.662, Tokens/s/device: 702.696, total_weights: 65536, loss: 7.288, lm_loss: 7.288, perplexity: 1463.272 I0420 12:28:50.856235 137741956515648 metric_logger.py:196] completed step: 12, seconds: 0.006, TFLOP/s/device: 2147.823, Tokens/s/device: 323743.282, total_weights: 65536, loss: 7.032, lm_loss: 7.032, perplexity: 1132.061 I0420 12:28:51.015181 137741956515648 metric_logger.py:196] completed step: 13, seconds: 0.159, TFLOP/s/device: 85.277, Tokens/s/device: 12853.906, total_weights: 65536, loss: 6.815, lm_loss: 6.815, perplexity: 911.031 I0420 12:28:51.173397 137741956515648 metric_logger.py:196] completed step: 14, seconds: 0.159, TFLOP/s/device: 85.707, Tokens/s/device: 12918.690, total_weights: 65536, loss: 6.635, lm_loss: 6.635, perplexity: 761.080 I0420 12:28:51.723648 137741956515648 metric_logger.py:196] completed step: 15, seconds: 0.159, TFLOP/s/device: 85.657, Tokens/s/device: 12911.116, total_weights: 65536, loss: 6.492, lm_loss: 6.492, perplexity: 659.915 I0420 12:28:51.881975 137741956515648 metric_logger.py:196] completed step: 16, seconds: 0.545, TFLOP/s/device: 24.917, Tokens/s/device: 3755.724, total_weights: 65536, loss: 6.380, lm_loss: 6.380, perplexity: 589.948 I0420 12:28:52.040321 137741956515648 metric_logger.py:196] completed step: 17, seconds: 0.163, TFLOP/s/device: 83.173, Tokens/s/device: 12536.805, total_weights: 65536, loss: 6.293, lm_loss: 6.293, perplexity: 540.941 I0420 12:28:52.198441 137741956515648 metric_logger.py:196] completed step: 18, seconds: 0.158, TFLOP/s/device: 85.941, Tokens/s/device: 12953.909, total_weights: 65536, loss: 6.223, lm_loss: 6.223, perplexity: 504.108 I0420 12:28:52.356230 137741956515648 checkpointing.py:794] Waiting for step 19 to finish before checkpoint... I0420 12:28:52.357221 137741956515648 checkpointing.py:798] Waited 0.0010120868682861328 seconds for step 19 to finish before starting checkpointing. I0420 12:28:52.359272 137741956515648 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=10][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 12:28:56.450215 137608193632000 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/10/items/array_metadatas/process_3 I0420 12:29:33.306313 137610862708480 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.873 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.84298300743103 s) (per-host) I0420 12:29:33.306453 137610862708480 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.777330s. I0420 12:29:41.709659 137610862708480 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 51.180520s. I0420 12:29:41.709930 137609762281216 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 12:29:41.710066 137609762281216 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0420 12:29:41.710115 137609762281216 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is syncing with other hosts... I0420 12:29:41.711786 137609762281216 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=10] CheckpointManager Save Finalize is done on all hosts. I0420 12:29:41.711960 137741956515648 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=10][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=10. W0420 12:29:41.712083 137741956515648 checkpoint_manager.py:1452] Waiting for previous save to complete took 49.352818 seconds. If this number is high, consider checkpointing less frequently. I0420 12:29:41.716386 137741956515648 checkpoint_manager.py:1512] [process=3] Saving checkpoint at step 19 I0420 12:29:41.718412 137741956515648 event_tracking.py:70] [process=3] [async] Started save checkpoint @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/19. I0420 12:29:42.008405 137741956515648 jax_array_handlers.py:360] Scheduling D2H of 81 prioritized jax.Array. I0420 12:29:42.008497 137741956515648 replica_slices.py:424] Transferring arrays to host memory with options: use_replica_parallel=True, min_slice_bytes_for_replica_parallel=None, max_replicas_for_replica_parallel=None, enable_pinned_host_transfer=False I0420 12:29:42.064344 137741956515648 base_pytree_checkpoint_handler.py:154] [process=3][thread=MainThread] Initiated "orbax.checkpoint._src.serialization.jax_array_handlers.ArrayHandler".serialize. Time taken: 0.056927s I0420 12:29:42.064527 137741956515648 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/blocking_gbytes_per_sec: 24.422 GiB/s (total gbytes: 1.5 GiB) (time elapsed: 0.06316876411437988 s) (per-host) I0420 12:29:42.064591 137741956515648 base_pytree_checkpoint_handler.py:768] [process=3][thread=MainThread] Initiated Pytree async_save. Time taken: 0.063242s (batch_requests_ready=0.003010s, total_serialization_initiated=0.060154s, others=0.000078s) I0420 12:29:42.064715 137741956515648 composite_checkpoint_handler.py:715] [process=3][thread=MainThread] Initiated CompositeCheckpointHandler.async_save. Time taken: 0.067443s (all_items=0.000012s, per_item={'items': '0.00001168'}, temp_paths=0.067431) I0420 12:29:42.065508 137741956515648 event_tracking.py:125] [process=3] [async] Finished blocking save in 0.35 seconds. Continuing save @ gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/19. I0420 12:29:42.065816 137609762281216 async_checkpointer.py:76] [process=3][thread=async_save] Background save thread started. Deadline for this save operation is 2026-04-20 12:49:42.065785 I0420 12:29:42.067835 137741956515648 checkpoint_manager.py:1560] [process=3][thread=MainThread][step=19] Starting CheckpointManager Save Finalize thread=save_finalize I0420 12:29:42.068104 137608193632000 async_checkpointer.py:280] [process=3][thread=save_finalize] Waiting for background save thread=async_save. I0420 12:29:42.068255 137741956515648 standard_logger.py:34] {'step': 19, 'event_type': 'save', 'directory': 'gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints', 'reached_preemption': False, 'preemption_received_at': None, 'synchronous': False, 'wait_for_prev_start_time': 1776688132.359243, 'wait_for_prev_duration_secs': 49.35281777381897, 'time_between_consecutive_saves_sec': 2.18363618850708, 'checkpointer_blocking_start_time': 1776688181.7164266, 'checkpointer_blocking_duration_secs': 0.349531888961792, 'get_old_steps_start_time': 1776688182.0659764, 'get_old_steps_duration_secs': 2.384185791015625e-05, 'checkpoint_manager_blocking_start_time': 1776688132.3574994, 'checkpoint_manager_blocking_duration_secs': 49.71072459220886} I0420 12:29:42.068436 137741956515648 checkpointing.py:409] Started an asynchronous checkpoint save for step 19 I0420 12:29:42.068481 137741956515648 checkpoint_manager.py:2020] [process=3][thread=MainThread][step=19][wait_until_finished] Waiting for Save Finalize thread (save_finalize) to complete. I0420 12:29:47.935306 137609770673920 array_metadata_store.py:203] [process=3][thread=array_type_handler] Wrote 81 array_metadata.ArrayMetadata to gs://lance-maxtext/linen_ckpt_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158/linen_xpk_feat_nnx_linen_converter_and_sharding_tools_20260420_112158_05_fp8/checkpoints/19/items/array_metadatas/process_3 I0420 12:30:24.764974 137609762281216 base_pytree_checkpoint_handler.py:130] [process=3] /jax/orbax/write/gbytes_per_sec: 36.942 MiB/s (total gbytes: 1.5 GiB) (time elapsed: 42.76358962059021 s) (per-host) I0420 12:30:24.765100 137609762281216 async_checkpointer.py:90] [process=3][thread=async_save] 3 Handler Commit operations completed. Time taken: 42.699174s. I0420 12:30:32.939180 137609762281216 async_checkpointer.py:160] [process=3][thread=async_save] Background save thread done. Time taken: 50.873238s. I0420 12:30:32.939503 137608193632000 async_checkpointer.py:288] [process=3][thread=save_finalize] Done with waiting for background save thread=async_save. I0420 12:30:32.939630 137608193632000 async_checkpointer.py:298] [process=3][thread=save_finalize] No errors found in background save thread=async_save. I0420 12:30:32.939672 137608193632000 checkpoint_manager.py:2137] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is syncing with other hosts... I0420 12:30:32.942431 137608193632000 checkpoint_manager.py:2146] [process=3][thread=save_finalize][step=19] CheckpointManager Save Finalize is done on all hosts. I0420 12:30:32.942605 137741956515648 checkpoint_manager.py:2032] [process=3][thread=MainThread][step=19][wait_until_finished] Done waiting for Save Finalize thread (save_finalize) running at step=19. I0420 12:30:32.942785 137741956515648 checkpoint_manager.py:2009] [process=3][thread=MainThread][wait_until_finished] No Save Finalize thread to wait for. Returning. I0420 12:30:32.943704 137741956515648 metric_logger.py:196] completed step: 19, seconds: 0.158, TFLOP/s/device: 85.780, Tokens/s/device: 12929.701, total_weights: 65536, loss: 6.167, lm_loss: 6.167, perplexity: 476.947 Per train step: Total TFLOPs: 13.59 split as 93.93% learnable weight flops and 6.07% attention flops XPK End: Mon Apr 20 12:30:44 UTC 2026 EXIT_CODE=0