2025-10-21 00:40:53,334 - INFO: Problem Type: text_causal_language_modeling 2025-10-21 00:40:53,334 - INFO: Global random seed: 42 2025-10-21 00:40:53,334 - INFO: Preparing the data... 2025-10-21 00:40:53,334 - INFO: Setting up automatic validation split... 2025-10-21 00:40:53,364 - INFO: Preparing train and validation data 2025-10-21 00:40:53,364 - INFO: Loading train dataset... 2025-10-21 00:41:02,760 - INFO: Stop token ids: [tensor([256002]), tensor([256000]), tensor([256001])] 2025-10-21 00:41:02,811 - INFO: Loading validation dataset... 2025-10-21 00:41:05,316 - INFO: Stop token ids: [tensor([256002]), tensor([256000]), tensor([256001])] 2025-10-21 00:41:05,373 - INFO: Number of observations in train dataset: 1088 2025-10-21 00:41:05,373 - INFO: Number of observations in validation dataset: 58 2025-10-21 00:41:08,375 - INFO: Stop token ids: [tensor([256002], device='cuda:0'), tensor([256000], device='cuda:0'), tensor([256001], device='cuda:0')] 2025-10-21 00:41:08,467 - INFO: Using bfloat16 for backbone 2025-10-21 00:41:08,469 - INFO: Loading google/gemma-2-2b. This may take a while. 2025-10-21 00:55:04,688 - INFO: Loaded google/gemma-2-2b. 2025-10-21 00:55:04,688 - INFO: Resizing token embeddings to 256003 2025-10-21 00:55:05,339 - INFO: Attention implementation: eager 2025-10-21 00:55:05,342 - INFO: Lora module names: ['q_proj', 'k_proj', 'v_proj', 'o_proj', 'gate_proj', 'up_proj', 'down_proj'] 2025-10-21 00:55:05,518 - INFO: Unfreezing layer: base_model.model.model.embed_tokens.weight 2025-10-21 00:55:05,527 - INFO: Trainable parameters count: 755964672 2025-10-21 00:55:05,528 - INFO: Total parameters count: 2780482560 2025-10-21 00:55:05,528 - INFO: Trainable %: 27.1883% 2025-10-21 00:55:05,536 - INFO: Enough space available for saving model weights.Required space: 5788.83MB, Available space: 93745.59MB. 2025-10-21 00:55:06,168 - INFO: Starting validation inference 2025-10-21 00:55:06,169 - INFO: validation progress: 0%| | 0/6 [00:00