Serendipity
1c1cccd3f6
- 添加 evaluate_checkpoints.py 脚本，用于评估训练过程中的检查点模型 - 更新 generate_plots.py 以支持从真实评估结果生成图表 - 更新实验报告内容，包含具体实验结果数据和分析 - 添加中文支持并更新作者信息 - 生成评估结果JSON文件和相应图表
80 lines
1.7 KiB
JSON
80 lines
1.7 KiB
JSON
[
|
|
{
|
|
"model": "models/dqn_step_100000.pt",
|
|
"step": 100000,
|
|
"avg_reward": 17.8,
|
|
"std_reward": 5.2306787322488075
|
|
},
|
|
{
|
|
"model": "models/dqn_step_200000.pt",
|
|
"step": 200000,
|
|
"avg_reward": 14.0,
|
|
"std_reward": 6.603029607687671
|
|
},
|
|
{
|
|
"model": "models/dqn_step_400000.pt",
|
|
"step": 400000,
|
|
"avg_reward": 16.4,
|
|
"std_reward": 4.24735211631906
|
|
},
|
|
{
|
|
"model": "models/dqn_step_600000.pt",
|
|
"step": 600000,
|
|
"avg_reward": 19.0,
|
|
"std_reward": 4.123105625617661
|
|
},
|
|
{
|
|
"model": "models/dqn_step_800000.pt",
|
|
"step": 800000,
|
|
"avg_reward": 13.2,
|
|
"std_reward": 3.944616584663204
|
|
},
|
|
{
|
|
"model": "models/dqn_step_1000000.pt",
|
|
"step": 1000000,
|
|
"avg_reward": 15.7,
|
|
"std_reward": 4.960846701924985
|
|
},
|
|
{
|
|
"model": "models/dqn_step_1200000.pt",
|
|
"step": 1200000,
|
|
"avg_reward": 18.4,
|
|
"std_reward": 6.216108107168021
|
|
},
|
|
{
|
|
"model": "models/dqn_step_1400000.pt",
|
|
"step": 1400000,
|
|
"avg_reward": 14.2,
|
|
"std_reward": 4.467661580737736
|
|
},
|
|
{
|
|
"model": "models/dqn_step_1600000.pt",
|
|
"step": 1600000,
|
|
"avg_reward": 17.8,
|
|
"std_reward": 4.686149805543993
|
|
},
|
|
{
|
|
"model": "models/dqn_step_1800000.pt",
|
|
"step": 1800000,
|
|
"avg_reward": 21.5,
|
|
"std_reward": 4.984977432245807
|
|
},
|
|
{
|
|
"model": "models/dqn_step_2000000.pt",
|
|
"step": 2000000,
|
|
"avg_reward": 14.6,
|
|
"std_reward": 5.276362383309167
|
|
},
|
|
{
|
|
"model": "models/dqn_best.pt",
|
|
"step": -1,
|
|
"avg_reward": 19.9,
|
|
"std_reward": 6.920260110718383
|
|
},
|
|
{
|
|
"model": "models/dqn_final.pt",
|
|
"step": -2,
|
|
"avg_reward": 11.3,
|
|
"std_reward": 3.3778691508109073
|
|
}
|
|
] |