jackf857/qwen3-8b-base-new-dpo-hh-harmless-4xh200-batch-64-q_t-0.45-s_star-0.4-eta-0.05-margin-log
收藏资源简介:
该数据集是从一个名为New-DPO的训练运行中导出的每一步边缘摘要统计。它包含了训练过程中的多个统计特征,如epoch(训练轮数)、step(步骤)、batch_size(批次大小)、mean(平均值)、std(标准差)、min(最小值)、p10(第10百分位数)、median(中位数)、p90(第90百分位数)、max(最大值)、pos_frac(正分数)、sample(每个示例的边缘值)以及npy(可选保存的完整边缘数组路径)。数据集来源于特定的模型训练运行,涉及特定的训练参数(如beta、f_divergence_type等)和数据集混合比例(如Anthropic/hh-rlhf)。
This dataset contains per-step margin summary statistics exported from a New-DPO training run. It includes various statistical features during training such as epoch, step, batch_size, mean, std, min, p10, median, p90, max, pos_frac, sample (per-example margins), and npy (optional path to the saved full margin array). The dataset originates from a specific model training run with particular training arguments (e.g., beta, f_divergence_type) and dataset mixer proportions (e.g., Anthropic/hh-rlhf).




