{
"episodes": 2,
"reward_count": 9,
"score_sum": 9.0,
"per_episode_spacing": {
"0": {
"steps": [
40,
89,
138,
174,
199,
222,
246
],
"count": 7,
"min_gap": 23,
"median_gap": 30.5,
"dense_gap_le_12_count": 0,
"gaps": [
49,
49,
36,
25,
23,
24
]
},
"1": {
"steps": [
40,
79
],
"count": 2,
"min_gap": 39,
"median_gap": 39.0,
"dense_gap_le_12_count": 0,
"gaps": [
39
]
}
},
"top_reward_window_ram": [
{
"byte": 70,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
44,
66,
44
]
},
{
"episode": 0,
"step": 89,
"values": [
66,
44,
22
]
},
{
"episode": 0,
"step": 138,
"values": [
44,
22,
0
]
},
{
"episode": 0,
"step": 174,
"values": [
184,
160,
138
]
},
{
"episode": 0,
"step": 199,
"values": [
68,
46,
46
]
},
{
"episode": 0,
"step": 222,
"values": [
160,
138,
138
]
},
{
"episode": 0,
"step": 246,
"values": [
68,
46,
68
]
},
{
"episode": 1,
"step": 40,
"values": [
158,
180,
180
]
}
]
},
{
"byte": 72,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
190,
179,
182
]
},
{
"episode": 0,
"step": 89,
"values": [
179,
182,
191
]
},
{
"episode": 0,
"step": 138,
"values": [
182,
191,
191
]
},
{
"episode": 0,
"step": 174,
"values": [
62,
64,
85
]
},
{
"episode": 0,
"step": 199,
"values": [
156,
177,
188
]
},
{
"episode": 0,
"step": 222,
"values": [
85,
86,
96
]
},
{
"episode": 0,
"step": 246,
"values": [
156,
177,
175
]
},
{
"episode": 1,
"step": 40,
"values": [
66,
65,
55
]
}
]
},
{
"byte": 77,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
0,
1,
1
]
},
{
"episode": 0,
"step": 89,
"values": [
1,
2,
2
]
},
{
"episode": 0,
"step": 138,
"values": [
2,
3,
3
]
},
{
"episode": 0,
"step": 174,
"values": [
3,
4,
4
]
},
{
"episode": 0,
"step": 199,
"values": [
4,
5,
5
]
},
{
"episode": 0,
"step": 222,
"values": [
5,
6,
6
]
},
{
"episode": 0,
"step": 246,
"values": [
6,
7,
7
]
},
{
"episode": 1,
"step": 40,
"values": [
0,
1,
1
]
}
]
},
{
"byte": 84,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
0,
5,
5
]
},
{
"episode": 0,
"step": 89,
"values": [
5,
10,
10
]
},
{
"episode": 0,
"step": 138,
"values": [
10,
15,
15
]
},
{
"episode": 0,
"step": 174,
"values": [
15,
20,
20
]
},
{
"episode": 0,
"step": 199,
"values": [
20,
25,
25
]
},
{
"episode": 0,
"step": 222,
"values": [
25,
30,
30
]
},
{
"episode": 0,
"step": 246,
"values": [
30,
35,
35
]
},
{
"episode": 1,
"step": 40,
"values": [
0,
5,
5
]
}
]
},
{
"byte": 90,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
173,
177,
181
]
},
{
"episode": 0,
"step": 89,
"values": [
113,
117,
121
]
},
{
"episode": 0,
"step": 138,
"values": [
53,
57,
61
]
},
{
"episode": 0,
"step": 174,
"values": [
197,
201,
205
]
},
{
"episode": 0,
"step": 199,
"values": [
41,
45,
49
]
},
{
"episode": 0,
"step": 222,
"values": [
133,
137,
141
]
},
{
"episode": 0,
"step": 246,
"values": [
229,
233,
237
]
},
{
"episode": 1,
"step": 40,
"values": [
174,
178,
182
]
}
]
},
{
"byte": 95,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
128,
5,
1
]
},
{
"episode": 0,
"step": 89,
"values": [
128,
5,
1
]
},
{
"episode": 0,
"step": 138,
"values": [
128,
3,
0
]
},
{
"episode": 0,
"step": 174,
"values": [
64,
5,
1
]
},
{
"episode": 0,
"step": 199,
"values": [
128,
3,
0
]
},
{
"episode": 0,
"step": 222,
"values": [
128,
5,
1
]
},
{
"episode": 0,
"step": 246,
"values": [
128,
3,
0
]
},
{
"episode": 1,
"step": 40,
"values": [
69,
3,
0
]
}
]
},
{
"byte": 99,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
186,
192,
198
]
},
{
"episode": 0,
"step": 89,
"values": [
192,
198,
195
]
},
{
"episode": 0,
"step": 138,
"values": [
198,
195,
189
]
},
{
"episode": 0,
"step": 174,
"values": [
74,
76,
78
]
},
{
"episode": 0,
"step": 199,
"values": [
172,
178,
184
]
},
{
"episode": 0,
"step": 222,
"values": [
91,
85,
79
]
},
{
"episode": 0,
"step": 246,
"values": [
163,
169,
175
]
},
{
"episode": 1,
"step": 40,
"values": [
58,
64,
70
]
}
]
},
{
"byte": 101,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
87,
85,
89
]
},
{
"episode": 0,
"step": 89,
"values": [
87,
85,
89
]
},
{
"episode": 0,
"step": 138,
"values": [
79,
81,
85
]
},
{
"episode": 0,
"step": 174,
"values": [
89,
85,
93
]
},
{
"episode": 0,
"step": 199,
"values": [
85,
89,
97
]
},
{
"episode": 0,
"step": 222,
"values": [
89,
85,
93
]
},
{
"episode": 0,
"step": 246,
"values": [
85,
89,
97
]
},
{
"episode": 1,
"step": 40,
"values": [
85,
87,
91
]
}
]
},
{
"byte": 103,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
255,
1,
1
]
},
{
"episode": 0,
"step": 89,
"values": [
255,
1,
1
]
},
{
"episode": 0,
"step": 138,
"values": [
255,
1,
1
]
},
{
"episode": 0,
"step": 174,
"values": [
254,
2,
2
]
},
{
"episode": 0,
"step": 199,
"values": [
254,
2,
2
]
},
{
"episode": 0,
"step": 222,
"values": [
254,
2,
2
]
},
{
"episode": 0,
"step": 246,
"values": [
254,
2,
2
]
},
{
"episode": 1,
"step": 40,
"values": [
255,
1,
1
]
}
]
},
{
"byte": 107,
"reward_window_changes": 9,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
65,
129,
129
]
},
{
"episode": 0,
"step": 89,
"values": [
66,
130,
130
]
},
{
"episode": 0,
"step": 138,
"values": [
67,
131,
131
]
},
{
"episode": 0,
"step": 174,
"values": [
68,
132,
132
]
},
{
"episode": 0,
"step": 199,
"values": [
69,
133,
133
]
},
{
"episode": 0,
"step": 222,
"values": [
70,
134,
134
]
},
{
"episode": 0,
"step": 246,
"values": [
71,
135,
135
]
},
{
"episode": 1,
"step": 40,
"values": [
65,
129,
129
]
}
]
},
{
"byte": 122,
"reward_window_changes": 5,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
246,
150,
150
]
},
{
"episode": 0,
"step": 89,
"values": [
246,
150,
150
]
},
{
"episode": 0,
"step": 138,
"values": [
246,
150,
150
]
},
{
"episode": 0,
"step": 174,
"values": [
246,
150,
150
]
},
{
"episode": 1,
"step": 40,
"values": [
246,
150,
150
]
}
]
},
{
"byte": 96,
"reward_window_changes": 4,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
0,
28,
28
]
},
{
"episode": 0,
"step": 138,
"values": [
28,
22,
22
]
},
{
"episode": 0,
"step": 174,
"values": [
22,
28,
28
]
},
{
"episode": 1,
"step": 40,
"values": [
0,
28,
28
]
}
]
},
{
"byte": 0,
"reward_window_changes": 2,
"examples": [
{
"episode": 0,
"step": 40,
"values": [
63,
51,
51
]
},
{
"episode": 0,
"step": 89,
"values": [
51,
3,
3
]
}
]
},
{
"byte": 6,
"reward_window_changes": 2,
"examples": [
{
"episode": 0,
"step": 199,
"values": [
255,
252,
252
]
},
{
"episode": 0,
"step": 246,
"values": [
252,
240,
240
]
}
]
},
{
"byte": 24,
"reward_window_changes": 2,
"examples": [
{
"episode": 0,
"step": 174,
"values": [
255,
207,
207
]
},
{
"episode": 0,
"step": 222,
"values": [
207,
195,
195
]
}
]
},
{
"byte": 105,
"reward_window_changes": 2,
"examples": [
{
"episode": 0,
"step": 89,
"values": [
254,
254,
1
]
},
{
"episode": 0,
"step": 138,
"values": [
254,
1,
1
]
}
]
},
{
"byte": 1,
"reward_window_changes": 1,
"examples": [
{
"episode": 0,
"step": 138,
"values": [
63,
15,
15
]
}
]
},
{
"byte": 12,
"reward_window_changes": 1,
"examples": [
{
"episode": 1,
"step": 79,
"values": [
255,
207,
207
]
}
]
},
{
"byte": 30,
"reward_window_changes": 1,
"examples": [
{
"episode": 1,
"step": 40,
"values": [
192,
0,
0
]
}
]
}
],
"visual_reward_examples": [
{
"episode": 0,
"step": 40,
"reward": 1.0,
"ball_before": [
137.5,
97.5
],
"ball_at_reward": [
143.5,
95.5
],
"ball_after": [
149.5,
99.5
],
"brick_area_before": 7800.0,
"brick_area_at_reward": 7752.0,
"brick_area_after": 7752.0
},
{
"episode": 0,
"step": 89,
"reward": 1.0,
"ball_before": [
143.5,
97.5
],
"ball_at_reward": [
149.5,
95.5
],
"ball_after": [
146.5,
99.5
],
"brick_area_before": 7752.0,
"brick_area_at_reward": 7704.0,
"brick_area_after": 7704.0
},
{
"episode": 0,
"step": 138,
"reward": 1.0,
"ball_before": null,
"ball_at_reward": [
146.5,
91.0
],
"ball_after": [
140.5,
95.5
],
"brick_area_before": 7704.0,
"brick_area_at_reward": 7658.0,
"brick_area_after": 7656.0
},
{
"episode": 0,
"step": 174,
"reward": 1.0,
"ball_before": [
25.5,
99.5
],
"ball_at_reward": [
27.5,
95.5
],
"ball_after": [
29.5,
103.5
],
"brick_area_before": 7656.0,
"brick_area_at_reward": 7608.0,
"brick_area_after": 7608.0
},
{
"episode": 0,
"step": 199,
"reward": 1.0,
"ball_before": [
123.5,
95.5
],
"ball_at_reward": [
129.5,
99.5
],
"ball_after": [
135.5,
107.5
],
"brick_area_before": 7608.0,
"brick_area_at_reward": 7560.0,
"brick_area_after": 7560.0
},
{
"episode": 0,
"step": 222,
"reward": 1.0,
"ball_before": [
42.5,
99.5
],
"ball_at_reward": [
36.5,
95.5
],
"ball_after": [
30.5,
103.5
],
"brick_area_before": 7560.0,
"brick_area_at_reward": 7512.0,
"brick_area_after": 7512.0
},
{
"episode": 0,
"step": 246,
"reward": 1.0,
"ball_before": [
114.5,
95.5
],
"ball_at_reward": [
120.5,
99.5
],
"ball_after": [
126.5,
107.5
],
"brick_area_before": 7512.0,
"brick_area_at_reward": 7464.0,
"brick_area_after": 7464.0
},
{
"episode": 1,
"step": 40,
"reward": 1.0,
"ball_before": [
9.5,
95.5
],
"ball_at_reward": [
15.5,
97.5
],
"ball_after": [
21.5,
101.5
],
"brick_area_before": 7800.0,
"brick_area_at_reward": 7752.0,
"brick_area_after": 7752.0
},
{
"episode": 1,
"step": 79,
"reward": 1.0,
"ball_before": null,
"ball_at_reward": [
80.5,
100.5
],
"ball_after": [
76.5,
106.5
],
"brick_area_before": 7752.0,
"brick_area_at_reward": 7704.0,
"brick_area_after": 7704.0
}
]
}
{
"episodes": 2,
"reward_count": 12,
"score_sum": 12.0,
"per_episode_spacing": {
"0": {
"steps": [
40,
128,
179,
236,
290,
346,
438
],
"count": 7,
"min_gap": 51,
"median_gap": 56.5,
"dense_gap_le_12_count": 0,
"gaps": [
88,
51,
57,
54,
56,
92
]
},
"1": {
"steps": [
40,
128,
216,
304,
392
],
"count": 5,
"min_gap": 88,
"median_gap": 88.0,
"dense_gap_le_12_count": 0,
"gaps": [
88,
88,
88,
88
]
}
},
"max_rewards_in_12_step_window": 1,
"brick_examples": [
{
"episode": 0,
"step": 40,
"reward": 1,
"ball_x": 56,
"ball_y": 79.75,
"ball_vx": 3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 4
}
]
},
{
"episode": 0,
"step": 128,
"reward": 1,
"ball_x": 35,
"ball_y": 79.75,
"ball_vx": 3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 2
}
]
},
{
"episode": 0,
"step": 179,
"reward": 1,
"ball_x": 131.25,
"ball_y": 78,
"ball_vx": -0.75,
"ball_vy": 10,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 11
}
]
},
{
"episode": 0,
"step": 236,
"reward": 1,
"ball_x": 48.5,
"ball_y": 71.5,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 4,
"col": 3
}
]
},
{
"episode": 0,
"step": 290,
"reward": 1,
"ball_x": 93.25,
"ball_y": 78,
"ball_vx": -0.75,
"ball_vy": 10,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 7
}
]
},
{
"episode": 0,
"step": 346,
"reward": 1,
"ball_x": 59.5,
"ball_y": 74.25,
"ball_vx": 3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 5
}
]
},
{
"episode": 0,
"step": 438,
"reward": 1,
"ball_x": 80,
"ball_y": 74.25,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 6
}
]
},
{
"episode": 1,
"step": 40,
"reward": 1,
"ball_x": 101,
"ball_y": 79.75,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 8
}
]
},
{
"episode": 1,
"step": 128,
"reward": 1,
"ball_x": 115,
"ball_y": 79.75,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 9
}
]
},
{
"episode": 1,
"step": 216,
"reward": 1,
"ball_x": 136,
"ball_y": 79.75,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 11
}
]
},
{
"episode": 1,
"step": 304,
"reward": 1,
"ball_x": 146.5,
"ball_y": 79.75,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 12
}
]
},
{
"episode": 1,
"step": 392,
"reward": 1,
"ball_x": 157,
"ball_y": 79.75,
"ball_vx": -3.5,
"ball_vy": 2.75,
"brick_events": [
{
"type": "brick",
"row": 5,
"col": 13
}
]
}
]
}
{
"before_ratio": 0.5,
"after_ratio": 0.75,
"ratio_delta": 0.25,
"before_go_mean": 9.0,
"after_go_mean": 6.0,
"before_atari_mean": 4.5,
"after_atari_mean": 4.5,
"after_verdict": "partial_transfer"
}
[ "Added brick_cooldown=12 to prevent one brick-region traversal from causing dense multi-brick reward bursts.", "The cooldown targets Go seed0's previous 7 rewards within 12 steps while preserving first_reward_step=40." ]
[ "https://file.aimusic.win/gemma/breakout/brick_score_v3_atari.mp4" ]