{
  "pipeline_version": "v1_clean_full_chat",
  "timestamp": "2026-07-12T23:22:47.281297+00:00",
  "start_time": "2026-07-12T23:22:47.286550+00:00",
  "end_time": "2026-07-13T10:50:40.904588+00:00",
  "duration_seconds": 41273.61803887808,
  "duration_human_readable": "11:27:53",
  "trl_version": "1.4.0",
  "transformers_version": "5.9.0",
  "peft_version": "0.18.1",
  "torch_version": "2.12.1",
  "datasets_version": "4.8.5",
  "model_id": "Qwen/Qwen3.5-9B-Base",
  "attn_implementation_requested": "flash_attention_2",
  "attn_implementation_effective": "flash_attention_2",
  "use_flash_attention": true,
  "flash_attention_requested": true,
  "flash_attention_available": true,
  "flash_attention_availability": {
    "torch_cuda_version": "13.0",
    "cuda_available": true,
    "device_count": 1,
    "device_name_0": "NVIDIA L40S",
    "capability_0": [
      8,
      9
    ],
    "bf16_supported": true,
    "flash_attn_installed": true,
    "flash_attn_interface_installed": false,
    "kernels_installed": false,
    "flash_attention_2_available": true,
    "flash_attention_3_available": false
  },
  "flash_attention_installed": true,
  "flash_attn_installed": true,
  "model_device": "cuda:0",
  "torch_cuda_version": "13.0",
  "cuda_available": true,
  "device_count": 1,
  "device_name_0": "NVIDIA L40S",
  "capability_0": [
    8,
    9
  ],
  "bf16_supported": true,
  "llm": "qwen35_9b_base",
  "adapter": "lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5",
  "continued_lora_training": false,
  "continue_from_adapter": null,
  "additional_epochs": 5.0,
  "total_effective_epochs": null,
  "dataset_path": "data/sql_create_context/train_sft_qwen35_9b_base_full_chat_v1_clean_anti_overjoin_mix_spider_train_sqlcc_spider_schema_harmonized_complexity_enriched_25k_seed42_no_dev_overlap.jsonl",
  "eval_dataset_path": "data/sql_create_context/val_sft_qwen35_full_chat_v2_mixed_trainothers700_sqlcc1800_no_train_no_dev_overlap_seed42_schemaheaderfix.jsonl",
  "dataset_format": "full_chat_text",
  "dataset_text_field": "text",
  "loss_mode": "full_chat_loss",
  "completion_only_loss": false,
  "assistant_only_loss": false,
  "packing": true,
  "packing_strategy": "bfd",
  "max_length": 2048,
  "raw_dataset_len": 25000,
  "selected_dataset_len": 25000,
  "processed_train_dataset_len": 4011,
  "raw_eval_dataset_len": 2500,
  "selected_eval_dataset_len": 2500,
  "processed_eval_dataset_len": 444,
  "label_stats_verified": true,
  "label_stats": {
    "verified": true,
    "dataset_name": "train",
    "loss_mode": "full_chat_loss",
    "checked_samples": 4,
    "has_completion_mask": false,
    "total_tokens_checked": 7963,
    "trainable_tokens_checked": 7956,
    "masked_tokens_checked": 7,
    "padding_tokens_checked": 0,
    "sequence_start_masked_tokens_checked": 7,
    "trainable_ratio": 0.9991209343212357,
    "masked_ratio": 0.0008790656787642849
  },
  "eval_label_stats_verified": true,
  "eval_label_stats": {
    "verified": true,
    "dataset_name": "eval",
    "loss_mode": "full_chat_loss",
    "checked_samples": 4,
    "has_completion_mask": false,
    "total_tokens_checked": 8059,
    "trainable_tokens_checked": 8047,
    "masked_tokens_checked": 12,
    "padding_tokens_checked": 0,
    "sequence_start_masked_tokens_checked": 12,
    "trainable_ratio": 0.9985109815113538,
    "masked_ratio": 0.001489018488646234
  },
  "packing_verified": true,
  "packing_verification": {
    "verified": true,
    "raw_dataset_len": 25000,
    "processed_train_dataset_len": 4011,
    "has_seq_lengths": true,
    "sequence_length_stats": {
      "count": 4011,
      "avg": 1967.6325105958613,
      "p50": 1990,
      "p90": 2048,
      "p95": 2048,
      "p99": 2048,
      "max": 2048
    },
    "packing_efficiency": 0.9607580618143854,
    "plausible_reduction": true,
    "expected_processed_len_upper_bound": 24500
  },
  "trainable_params": 21639168,
  "total_params": 8975442432,
  "trainable_param_ratio": 0.0024109305099936046,
  "epochs": 5.0,
  "learning_rate": 0.0001,
  "per_device_train_batch_size": 2,
  "per_device_eval_batch_size": 1,
  "gradient_accumulation_steps": 4,
  "effective_batch_size": 8,
  "expected_steps_per_epoch_before_packing": 3125,
  "expected_total_steps_before_packing": 15625,
  "trainer_expected_steps_per_epoch": 502,
  "trainer_expected_total_steps": 2510,
  "eval_strategy": "epoch",
  "eval_steps": null,
  "eval_accumulation_steps": 1,
  "prediction_loss_only": true,
  "save_strategy": "epoch",
  "save_total_limit": 5,
  "load_best_model_at_end": true,
  "metric_for_best_model": "eval_loss",
  "greater_is_better": false,
  "save_best_model": true,
  "early_stopping": {
    "enabled": true,
    "early_stopping_patience": 2,
    "early_stopping_threshold": 0.001,
    "metric": "eval_loss"
  },
  "test_mode": {
    "enabled": false,
    "max_train_samples": null,
    "max_eval_samples": null,
    "disable_adapter_save": false
  },
  "auto_resume": true,
  "overwrite_output_dir": false,
  "gradient_checkpointing": true,
  "model_use_cache_before": true,
  "model_use_cache_after": false,
  "eval_cuda_empty_cache_enabled": true,
  "torch_compile": false,
  "torch_empty_cache_steps": 4,
  "fp16": true,
  "bf16": false,
  "warmup_ratio": 0.03,
  "lr_scheduler_type": "constant",
  "max_grad_norm": 0.3,
  "seed": 42,
  "lora": {
    "r": 8,
    "lora_alpha": 16,
    "lora_dropout": 0.05,
    "bias": "none",
    "task_type": "CAUSAL_LM",
    "use_dora": false,
    "target_modules": "all-linear"
  },
  "sft_config_applied_fields": {
    "output_dir": true,
    "dataset_text_field": true,
    "packing": true,
    "packing_strategy": true,
    "max_length": true,
    "completion_only_loss": true,
    "assistant_only_loss": true,
    "learning_rate": true,
    "num_train_epochs": true,
    "per_device_train_batch_size": true,
    "gradient_accumulation_steps": true,
    "save_strategy": true,
    "save_total_limit": true,
    "logging_steps": true,
    "fp16": true,
    "bf16": true,
    "gradient_checkpointing": true,
    "report_to": true,
    "warmup_ratio": true,
    "lr_scheduler_type": true,
    "max_grad_norm": true,
    "seed": true,
    "torch_compile": true,
    "torch_empty_cache_steps": true,
    "eval_strategy": true,
    "per_device_eval_batch_size": true,
    "prediction_loss_only": true,
    "load_best_model_at_end": true,
    "metric_for_best_model": true,
    "greater_is_better": true,
    "eval_accumulation_steps": true
  },
  "effective_sft_config_kwargs": {
    "output_dir": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5/checkpoints",
    "dataset_text_field": "text",
    "packing": true,
    "packing_strategy": "bfd",
    "max_length": 2048,
    "completion_only_loss": false,
    "assistant_only_loss": false,
    "learning_rate": 0.0001,
    "num_train_epochs": 5.0,
    "per_device_train_batch_size": 2,
    "gradient_accumulation_steps": 4,
    "save_strategy": "epoch",
    "save_total_limit": 5,
    "logging_steps": 10,
    "fp16": true,
    "bf16": false,
    "gradient_checkpointing": true,
    "report_to": "none",
    "warmup_ratio": 0.03,
    "lr_scheduler_type": "constant",
    "max_grad_norm": 0.3,
    "seed": 42,
    "torch_compile": false,
    "torch_empty_cache_steps": 4,
    "eval_strategy": "epoch",
    "per_device_eval_batch_size": 1,
    "prediction_loss_only": true,
    "load_best_model_at_end": true,
    "metric_for_best_model": "eval_loss",
    "greater_is_better": false,
    "eval_accumulation_steps": 1
  },
  "sft_config_effective": {
    "output_dir": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5/checkpoints",
    "per_device_train_batch_size": 2,
    "num_train_epochs": 5.0,
    "max_steps": -1,
    "learning_rate": 0.0001,
    "lr_scheduler_type": "constant",
    "lr_scheduler_kwargs": null,
    "warmup_steps": 0.03,
    "optim": "adamw_torch_fused",
    "optim_args": null,
    "weight_decay": 0.0,
    "adam_beta1": 0.9,
    "adam_beta2": 0.999,
    "adam_epsilon": 1e-08,
    "optim_target_modules": null,
    "gradient_accumulation_steps": 4,
    "average_tokens_across_devices": true,
    "max_grad_norm": 0.3,
    "label_smoothing_factor": 0.0,
    "bf16": false,
    "fp16": true,
    "bf16_full_eval": false,
    "fp16_full_eval": false,
    "tf32": null,
    "gradient_checkpointing": true,
    "gradient_checkpointing_kwargs": null,
    "torch_compile": false,
    "torch_compile_backend": null,
    "torch_compile_mode": null,
    "use_liger_kernel": false,
    "liger_kernel_config": null,
    "use_cache": false,
    "neftune_noise_alpha": null,
    "torch_empty_cache_steps": 4,
    "auto_find_batch_size": false,
    "logging_strategy": "steps",
    "logging_steps": 10,
    "logging_first_step": false,
    "log_on_each_node": true,
    "logging_nan_inf_filter": true,
    "include_num_input_tokens_seen": "no",
    "log_level": "passive",
    "log_level_replica": "warning",
    "disable_tqdm": false,
    "report_to": [],
    "run_name": null,
    "project": "huggingface",
    "trackio_space_id": null,
    "trackio_bucket_id": null,
    "trackio_static_space_id": null,
    "eval_strategy": "epoch",
    "eval_steps": null,
    "eval_delay": 0,
    "per_device_eval_batch_size": 1,
    "prediction_loss_only": true,
    "eval_on_start": false,
    "eval_do_concat_batches": true,
    "eval_use_gather_object": false,
    "eval_accumulation_steps": 1,
    "include_for_metrics": [],
    "batch_eval_metrics": false,
    "save_only_model": false,
    "save_strategy": "epoch",
    "save_steps": 500,
    "save_on_each_node": false,
    "save_total_limit": 5,
    "enable_jit_checkpoint": false,
    "push_to_hub": false,
    "hub_token": null,
    "hub_private_repo": null,
    "hub_model_id": null,
    "hub_strategy": "every_save",
    "hub_always_push": false,
    "hub_revision": null,
    "load_best_model_at_end": true,
    "metric_for_best_model": "eval_loss",
    "greater_is_better": false,
    "ignore_data_skip": false,
    "restore_callback_states_from_checkpoint": false,
    "full_determinism": false,
    "seed": 42,
    "data_seed": null,
    "use_cpu": false,
    "accelerator_config": "AcceleratorConfig(split_batches=False, dispatch_batches=None, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False)",
    "parallelism_config": null,
    "dataloader_drop_last": false,
    "dataloader_num_workers": 0,
    "dataloader_pin_memory": true,
    "dataloader_persistent_workers": false,
    "dataloader_prefetch_factor": null,
    "remove_unused_columns": true,
    "label_names": null,
    "train_sampling_strategy": "random",
    "length_column_name": "length",
    "ddp_find_unused_parameters": null,
    "ddp_bucket_cap_mb": null,
    "ddp_broadcast_buffers": null,
    "ddp_static_graph": null,
    "ddp_backend": null,
    "ddp_timeout": 1800,
    "fsdp": [],
    "fsdp_config": {
      "min_num_params": 0,
      "xla": false,
      "xla_fsdp_v2": false,
      "xla_fsdp_grad_ckpt": false
    },
    "deepspeed": null,
    "debug": [],
    "skip_memory_metrics": true,
    "do_train": false,
    "do_eval": true,
    "do_predict": false,
    "resume_from_checkpoint": null,
    "warmup_ratio": 0.03,
    "logging_dir": null,
    "local_rank": -1,
    "model_init_kwargs": null,
    "chat_template_path": null,
    "dataset_text_field": "text",
    "dataset_kwargs": null,
    "dataset_num_proc": null,
    "eos_token": null,
    "max_length": 2048,
    "truncation_mode": "keep_start",
    "shuffle_dataset": false,
    "packing": true,
    "packing_strategy": "bfd",
    "padding_free": false,
    "pad_to_multiple_of": null,
    "eval_packing": null,
    "completion_only_loss": false,
    "assistant_only_loss": false,
    "loss_type": "nll",
    "activation_offloading": false,
    "pad_token": null,
    "mixed_precision": "fp16",
    "distributed_state": "Distributed environment: DistributedType.NO\nNum processes: 1\nProcess index: 0\nLocal process index: 0\nDevice: cuda\n",
    "_n_gpu": 1,
    "_setup_devices": "cuda:0",
    "fsdp_plugin_args": null,
    "deepspeed_plugin": null
  },
  "output_dir": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5",
  "checkpoint_dir": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5/checkpoints",
  "training_history_csv": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5/training_history.csv",
  "training_history_jsonl": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5/training_history.jsonl",
  "central_training_history_csv": "results/training_metrics/qwen35_9b_base__lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5_training_history.csv",
  "central_training_history_jsonl": "results/training_metrics/qwen35_9b_base__lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5_training_history.jsonl",
  "final_train_loss": 0.1824839700107397,
  "latest_eval_loss": 0.4483620524406433,
  "best_model_checkpoint": "adapters/qwen35_9b_base/lora_v2_fullchat_old25k_r8_alpha16_mixedval2500_v2_schemaheaderfix_evalstop_maxlen2048_epochs5/checkpoints/checkpoint-502",
  "best_metric": 0.4077516198158264,
  "best_eval_loss": 0.4077516198158264,
  "stopped_epoch": 3.0,
  "gpu_name": "NVIDIA L40S",
  "cuda_version": "13.0",
  "peak_memory_allocated": 34693337600,
  "peak_memory_reserved": 42138075136,
  "trainer_max_steps": 2510,
  "trainer_num_train_epochs": 5.0,
  "train_begin_timestamp": "2026-07-12T23:22:47.579896+00:00",
  "adapter_save_enabled": true,
  "model_revision": "68c46c4b3498877f3ef123c856ecfde50c39f404",
  "adapter_repository_id": "mehmet1899/qwen35-9b-nl2sql-lora",
  "adapter_repository_revision": "e136b9c25ede3ee82210875d0db774089509b676"
}