[2026-04-23 22:52:29,657] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:2655295] baseline 0.000GB () [2026-04-23 22:52:29,657] [INFO] [axolotl.cli.config.load_cfg:259] [PID:2655295] config: { "activation_offloading": false, "adapter": "lora", "axolotl_config_path": "./axolotl_configs/Qwen2.5-coder-7b-instruct/stage_1_2/lora-stage2-secure-token-diff-ctx0.yaml", "base_model": "Qwen/Qwen2.5-Coder-7B-Instruct", "base_model_config": "Qwen/Qwen2.5-Coder-7B-Instruct", "batch_size": 64, "bf16": true, "capabilities": { "bf16": true, "compute_capability": "sm_90", "fp8": true, "n_gpu": 4, "n_node": 1 }, "context_parallel_size": 1, "dataloader_num_workers": 4, "dataloader_pin_memory": true, "dataloader_prefetch_factor": 256, "dataset_num_proc": 96, "dataset_prepared_path": "/home/tkwang/links/scratch/SecSteer-v2/axolotl-datasets/lora/Qwen2.5-Coder-7B/stage_2_secure_token_diff_mask_skip_indent_ctx0_chat", "datasets": [ { "message_property_mappings": { "content": "content", "role": "role" }, "path": "felixwangg/stage_2_secure_token_diff_mask_skip_indent_ctx0_chat", "split": "train", "trust_remote_code": false, "type": "pretokenized" } ], "ddp": true, "device": "cuda:0", "device_map": { "": 0 }, "dion_rank_fraction": 1.0, "dion_rank_multiple_of": 1, "early_stopping_patience": 1000, "env_capabilities": { "torch_version": "2.10.0" }, "eval_batch_size": 4, "eval_causal_lm_metrics": [ "sacrebleu", "comet", "ter", "chrf" ], "eval_max_new_tokens": 128, "eval_sample_packing": false, "eval_steps": 15, "eval_table_size": 0, "experimental_skip_move_to_device": true, "flash_attention": true, "fp16": false, "gradient_accumulation_steps": 4, "gradient_checkpointing": true, "gradient_checkpointing_kwargs": { "use_reentrant": true }, "include_tkps": true, "is_falcon_derived_model": false, "is_llama_derived_model": false, "is_mistral_derived_model": false, "learning_rate": 4e-05, "lisa_layers_attribute": "model.layers", "load_best_model_at_end": true, "load_in_4bit": false, "load_in_8bit": false, "local_rank": 0, "logging_steps": 1, "lora_alpha": 16, "lora_dropout": 0.05, "lora_model_dir": "/home/tkwang/links/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-stage1-combined/checkpoint-6", "lora_r": 16, "lora_target_linear": true, "loraplus_lr_embedding": 1e-06, "lr_scheduler": "cosine", "mean_resizing_embeddings": false, "merge_lora": false, "micro_batch_size": 4, "model_config_type": "qwen2", "num_epochs": 2.0, "optimizer": "adamw_torch", "otel_metrics_host": "localhost", "otel_metrics_port": 8000, "output_dir": "/home/tkwang/links/scratch/SecSteer-v2/axolotl-outputs/lora/Qwen2.5-Coder-7B-stage2-secure-token-diff-ctx0", "pad_to_sequence_len": true, "plugins": [ "diff_mask_trainer.plugin.DiffMaskPlugin" ], "pretrain_multipack_attn": true, "profiler_steps_start": 0, "qlora_sharded_model_loading": false, "ray_num_workers": 1, "resources_per_worker": { "GPU": 1 }, "sample_packing": false, "sample_packing_bin_size": 200, "sample_packing_group_size": 100000, "save_only_model": false, "save_safetensors": true, "save_steps": 15, "save_total_limit": 1000, "sequence_len": 4096, "shuffle_before_merging_datasets": false, "shuffle_merged_datasets": true, "skip_prepare_dataset": false, "streaming_multipack_buffer_size": 10000, "strict": false, "tensor_parallel_size": 1, "test_datasets": [ { "message_property_mappings": { "content": "content", "role": "role" }, "path": "felixwangg/stage_2_secure_token_diff_mask_skip_indent_ctx0_chat", "split": "validation", "trust_remote_code": false, "type": "pretokenized" } ], "tf32": false, "tiled_mlp_use_original_mlp": true, "tokenizer_config": "Qwen/Qwen2.5-Coder-7B-Instruct", "tokenizer_save_jinja_files": true, "tokenizer_type": "AutoTokenizer", "torch_dtype": "torch.bfloat16", "train_on_inputs": false, "trl": { "log_completions": false, "mask_truncated_completions": false, "ref_model_mixup_alpha": 0.9, "ref_model_sync_steps": 64, "scale_rewards": true, "sync_ref_model": false, "use_vllm": false, "vllm_server_host": "0.0.0.0", "vllm_server_port": 8000 }, "type_of_model": "Qwen2ForCausalLM", "use_otel_metrics": false, "use_ray": false, "use_wandb": true, "val_set_size": 0.0, "vllm": { "device": "auto", "dtype": "auto", "gpu_memory_utilization": 0.9, "host": "0.0.0.0", "port": 8000 }, "wandb_entity": "wtkuan", "wandb_log_model": "false", "wandb_name": "Qwen2.5-Coder-7B-stage2-secure-token-diff-ctx0", "wandb_project": "diff-mask-stage1-2-ctx-0", "wandb_watch": "false", "warmup_ratio": 0.1, "weight_decay": 0.02, "world_size": 4 } [2026-04-23 22:52:30,639] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:285] [PID:2655295] EOS: 151645 / <|im_end|> [2026-04-23 22:52:30,640] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:286] [PID:2655295] BOS: None / None [2026-04-23 22:52:30,640] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:287] [PID:2655295] PAD: 151643 / <|endoftext|> [2026-04-23 22:52:30,640] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:288] [PID:2655295] UNK: None / None [2026-04-23 22:52:30,647] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:475] [PID:2655295] Loading prepared dataset from disk at /home/tkwang/links/scratch/SecSteer-v2/axolotl-datasets/lora/Qwen2.5-Coder-7B/stage_2_secure_token_diff_mask_skip_indent_ctx0_chat/21d42ece2ab8cc786c50d82a4c015221... [2026-04-23 22:52:30,660] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:475] [PID:2655295] Loading prepared dataset from disk at /home/tkwang/links/scratch/SecSteer-v2/axolotl-datasets/lora/Qwen2.5-Coder-7B/stage_2_secure_token_diff_mask_skip_indent_ctx0_chat/8ede901bf97cb6e2000275c274728568... [2026-04-23 22:52:30,676] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:417] [PID:2655295] total_num_tokens: 4_377_675 [2026-04-23 22:52:30,761] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:435] [PID:2655295] `total_supervised_tokens: 3_580_214` [2026-04-23 22:52:30,761] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:533] [PID:2655295] total_num_steps: 115 [2026-04-23 22:52:30,761] [INFO] [axolotl.utils.data.sft._prepare_standard_dataset:121] [PID:2655295] Maximum number of steps set at 115 [2026-04-23 22:52:30,781] [DEBUG] [axolotl.train.setup_model_and_tokenizer:70] [PID:2655295] loading tokenizer... Qwen/Qwen2.5-Coder-7B-Instruct [2026-04-23 22:52:31,189] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:285] [PID:2655295] EOS: 151645 / <|im_end|> [2026-04-23 22:52:31,189] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:286] [PID:2655295] BOS: None / None [2026-04-23 22:52:31,189] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:287] [PID:2655295] PAD: 151643 / <|endoftext|> [2026-04-23 22:52:31,189] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:288] [PID:2655295] UNK: None / None [2026-04-23 22:52:31,189] [DEBUG] [axolotl.train.setup_model_and_tokenizer:82] [PID:2655295] Loading model [2026-04-23 22:52:31,245] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:87] [PID:2655295] Patched Trainer.evaluation_loop with nanmean loss calculation [2026-04-23 22:52:31,246] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:138] [PID:2655295] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation Loading checkpoint shards: 0%| | 0/4 [00:00