wuff-mann commited on
Commit
5c4ea4e
·
verified ·
1 Parent(s): bab0902

MHC-A1 durable progress index

Browse files
Files changed (1) hide show
  1. experiments/MHC_REAL_A1/progress.json +63 -49
experiments/MHC_REAL_A1/progress.json CHANGED
@@ -7,8 +7,8 @@
7
  "arm": "stateless",
8
  "validation": {
9
  "W3_MSE": 0.11103641552229722,
10
- "W4_MSE": 0.10791260438660781,
11
- "ctrl_max_rms": 0.00528129100954781
12
  },
13
  "checkpoint_step": 64
14
  },
@@ -18,7 +18,7 @@
18
  "validation": {
19
  "W3_MSE": 0.11103636150558789,
20
  "W4_MSE": 0.10791248766084512,
21
- "ctrl_max_rms": 0.005226058148158093
22
  },
23
  "checkpoint_step": 64
24
  },
@@ -26,9 +26,9 @@
26
  "seed": 17,
27
  "arm": "mhc_noaux",
28
  "validation": {
29
- "W3_MSE": 0.1110341747601827,
30
- "W4_MSE": 0.10790904238820076,
31
- "ctrl_max_rms": 0.0052098391655211644
32
  },
33
  "checkpoint_step": 64
34
  }
@@ -40,7 +40,7 @@
40
  "step": 0,
41
  "loss_train": 0.09119217842817307,
42
  "grad_norm": 0.013730722479522228,
43
- "head_grad_norm": 1.4221210449250066e-06,
44
  "q0_trials": [
45
  {
46
  "radius": 0.0003,
@@ -74,7 +74,7 @@
74
  "loss_train": 0.08606645464897156,
75
  "W3_MSE": 0.1147225337723891,
76
  "W4_MSE": 0.11396808549761772,
77
- "ctrl_max_rms": 0.00018752604228211567
78
  },
79
  {
80
  "seed": 17,
@@ -83,7 +83,7 @@
83
  "loss_train": 0.1242414116859436,
84
  "W3_MSE": 0.11465953538815181,
85
  "W4_MSE": 0.11387925470868747,
86
- "ctrl_max_rms": 0.00030438662846184644
87
  },
88
  {
89
  "seed": 17,
@@ -91,8 +91,8 @@
91
  "step": 24,
92
  "loss_train": 0.09662917256355286,
93
  "W3_MSE": 0.11453648408253987,
94
- "W4_MSE": 0.11368757920960586,
95
- "ctrl_max_rms": 0.0003923155600205064
96
  },
97
  {
98
  "seed": 17,
@@ -101,7 +101,7 @@
101
  "loss_train": 0.1582513302564621,
102
  "W3_MSE": 0.1143090259283781,
103
  "W4_MSE": 0.11331315711140633,
104
- "ctrl_max_rms": 0.0005296767340041697
105
  },
106
  {
107
  "seed": 17,
@@ -118,15 +118,15 @@
118
  "step": 48,
119
  "loss_train": 0.10963404178619385,
120
  "W3_MSE": 0.1130196029941241,
121
- "W4_MSE": 0.11111373826861382,
122
- "ctrl_max_rms": 0.00205805353471078
123
  },
124
  {
125
  "seed": 17,
126
  "arm": "stateless",
127
  "step": 56,
128
  "loss_train": 0.10562591254711151,
129
- "W3_MSE": 0.11191886352996032,
130
  "W4_MSE": 0.10920579296847184,
131
  "ctrl_max_rms": 0.004299115234365066
132
  },
@@ -136,8 +136,8 @@
136
  "step": 64,
137
  "loss_train": 0.1330445259809494,
138
  "W3_MSE": 0.11103641552229722,
139
- "W4_MSE": 0.10791260438660781,
140
- "ctrl_max_rms": 0.00528129100954781
141
  },
142
  {
143
  "seed": 17,
@@ -177,9 +177,9 @@
177
  "arm": "mhc",
178
  "step": 8,
179
  "loss_train": 0.08606645464897156,
180
- "W3_MSE": 0.11472253315150738,
181
  "W4_MSE": 0.113968084876736,
182
- "ctrl_max_rms": 0.00018755901813468276
183
  },
184
  {
185
  "seed": 17,
@@ -188,7 +188,7 @@
188
  "loss_train": 0.1242414116859436,
189
  "W3_MSE": 0.11465953538815181,
190
  "W4_MSE": 0.11387925657133262,
191
- "ctrl_max_rms": 0.0003044217810383998
192
  },
193
  {
194
  "seed": 17,
@@ -197,7 +197,7 @@
197
  "loss_train": 0.09662917256355286,
198
  "W3_MSE": 0.11453648408253987,
199
  "W4_MSE": 0.11368757796784242,
200
- "ctrl_max_rms": 0.00039232581184478477
201
  },
202
  {
203
  "seed": 17,
@@ -206,7 +206,7 @@
206
  "loss_train": 0.1582513302564621,
207
  "W3_MSE": 0.11430902406573296,
208
  "W4_MSE": 0.11331315897405148,
209
- "ctrl_max_rms": 0.0005288315539170677
210
  },
211
  {
212
  "seed": 17,
@@ -215,25 +215,25 @@
215
  "loss_train": 0.10655954480171204,
216
  "W3_MSE": 0.11380419631799062,
217
  "W4_MSE": 0.11249386705458164,
218
- "ctrl_max_rms": 0.0009341221496773263
219
  },
220
  {
221
  "seed": 17,
222
  "arm": "mhc",
223
  "step": 48,
224
  "loss_train": 0.109634168446064,
225
- "W3_MSE": 0.11301966508229573,
226
  "W4_MSE": 0.11111390093962352,
227
- "ctrl_max_rms": 0.0020173621790794036
228
  },
229
  {
230
  "seed": 17,
231
  "arm": "mhc",
232
  "step": 56,
233
  "loss_train": 0.10562613606452942,
234
- "W3_MSE": 0.1119188933322827,
235
- "W4_MSE": 0.10920589851836364,
236
- "ctrl_max_rms": 0.004241040248113374
237
  },
238
  {
239
  "seed": 17,
@@ -242,7 +242,7 @@
242
  "loss_train": 0.13304445147514343,
243
  "W3_MSE": 0.11103636150558789,
244
  "W4_MSE": 0.10791248766084512,
245
- "ctrl_max_rms": 0.005226058148158093
246
  },
247
  {
248
  "seed": 17,
@@ -284,7 +284,7 @@
284
  "loss_train": 0.08451367169618607,
285
  "W3_MSE": 0.11472250955800216,
286
  "W4_MSE": 0.11396805879970391,
287
- "ctrl_max_rms": 0.000187463211962798
288
  },
289
  {
290
  "seed": 17,
@@ -293,7 +293,7 @@
293
  "loss_train": 0.12192638218402863,
294
  "W3_MSE": 0.11465946398675442,
295
  "W4_MSE": 0.11387916964789231,
296
- "ctrl_max_rms": 0.00030426468583755195
297
  },
298
  {
299
  "seed": 17,
@@ -302,7 +302,7 @@
302
  "loss_train": 0.09486307948827744,
303
  "W3_MSE": 0.11453630837301414,
304
  "W4_MSE": 0.1136873426536719,
305
- "ctrl_max_rms": 0.00039202869569029036
306
  },
307
  {
308
  "seed": 17,
@@ -311,7 +311,7 @@
311
  "loss_train": 0.1553032398223877,
312
  "W3_MSE": 0.11430866395433743,
313
  "W4_MSE": 0.11331262377401193,
314
- "ctrl_max_rms": 0.0005277997018614163
315
  },
316
  {
317
  "seed": 17,
@@ -320,7 +320,7 @@
320
  "loss_train": 0.10460616648197174,
321
  "W3_MSE": 0.11380349472165108,
322
  "W4_MSE": 0.1124927494674921,
323
- "ctrl_max_rms": 0.0009292569302488118
324
  },
325
  {
326
  "seed": 17,
@@ -329,25 +329,25 @@
329
  "loss_train": 0.10761874914169312,
330
  "W3_MSE": 0.11301849285761516,
331
  "W4_MSE": 0.11111204698681831,
332
- "ctrl_max_rms": 0.002005132545794671
333
  },
334
  {
335
  "seed": 17,
336
  "arm": "mhc_noaux",
337
  "step": 56,
338
  "loss_train": 0.10368967801332474,
339
- "W3_MSE": 0.11191730387508869,
340
  "W4_MSE": 0.10920340940356255,
341
- "ctrl_max_rms": 0.004224800640561928
342
  },
343
  {
344
  "seed": 17,
345
  "arm": "mhc_noaux",
346
  "step": 64,
347
  "loss_train": 0.1305774450302124,
348
- "W3_MSE": 0.1110341747601827,
349
- "W4_MSE": 0.10790904238820076,
350
- "ctrl_max_rms": 0.0052098391655211644
351
  },
352
  {
353
  "seed": 41,
@@ -389,7 +389,16 @@
389
  "loss_train": 0.10454776883125305,
390
  "W3_MSE": 0.11467163388927777,
391
  "W4_MSE": 0.11390550496677558,
392
- "ctrl_max_rms": 9.996155737705219e-05
 
 
 
 
 
 
 
 
 
393
  }
394
  ],
395
  "checkpoints": [
@@ -399,42 +408,42 @@
399
  "bytes": 93186
400
  },
401
  {
402
- "sha256": "8564fa6b02b9004df33b139b15888e345f27bc9392db5b1bdd36918417abb5d3",
403
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0008.pt",
404
  "bytes": 268976
405
  },
406
  {
407
- "sha256": "42792020756dbc501ab231c407134a90b13db10de7d9dca0192c4da6eabafbae",
408
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0016.pt",
409
  "bytes": 268976
410
  },
411
  {
412
- "sha256": "c1297071cf2cd8ef2da147b116cd01823b41f9af63bc2a603e075a72ec9936ca",
413
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0024.pt",
414
  "bytes": 268976
415
  },
416
  {
417
- "sha256": "c95fcb3a196dfed881e2a01f5b54b8dea25fbfe13f5eae71336af1c671f71d14",
418
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0032.pt",
419
  "bytes": 268976
420
  },
421
  {
422
- "sha256": "84d1bb51dafe0dc9ab687d2bfaa6f2e590a6a257a72053a9d05610754a27e554",
423
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0040.pt",
424
  "bytes": 268976
425
  },
426
  {
427
- "sha256": "a2a4ffa22bf4208049e3b06accfb87699cb28b0987592bbbbdada6aea7410750",
428
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0048.pt",
429
  "bytes": 268976
430
  },
431
  {
432
- "sha256": "99d7cae2f26351d8183ceb0b2c152adf512324e147ffbce0c8b0a5fd4c34163a",
433
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0056.pt",
434
  "bytes": 268976
435
  },
436
  {
437
- "sha256": "e2a449c3025fe0223022be26a6a9c1ee8e6a60e46f146208f6a7a532f9dba37e",
438
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0064.pt",
439
  "bytes": 268976
440
  },
@@ -537,6 +546,11 @@
537
  "sha256": "d934489f5dcecd5ae7d831b52e1ca108c4866f2028f8b338745d3986bc4f8cd6",
538
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed41_stateless_step0008.pt",
539
  "bytes": 268976
 
 
 
 
 
540
  }
541
  ]
542
  }
 
7
  "arm": "stateless",
8
  "validation": {
9
  "W3_MSE": 0.11103641552229722,
10
+ "W4_MSE": 0.10791260376572609,
11
+ "ctrl_max_rms": 0.005281291048352917
12
  },
13
  "checkpoint_step": 64
14
  },
 
18
  "validation": {
19
  "W3_MSE": 0.11103636150558789,
20
  "W4_MSE": 0.10791248766084512,
21
+ "ctrl_max_rms": 0.005226058070547879
22
  },
23
  "checkpoint_step": 64
24
  },
 
26
  "seed": 17,
27
  "arm": "mhc_noaux",
28
  "validation": {
29
+ "W3_MSE": 0.11103417413930099,
30
+ "W4_MSE": 0.10790904176731904,
31
+ "ctrl_max_rms": 0.005209839281936486
32
  },
33
  "checkpoint_step": 64
34
  }
 
40
  "step": 0,
41
  "loss_train": 0.09119217842817307,
42
  "grad_norm": 0.013730722479522228,
43
+ "head_grad_norm": 1.4221209312381689e-06,
44
  "q0_trials": [
45
  {
46
  "radius": 0.0003,
 
74
  "loss_train": 0.08606645464897156,
75
  "W3_MSE": 0.1147225337723891,
76
  "W4_MSE": 0.11396808549761772,
77
+ "ctrl_max_rms": 0.00018752604106945606
78
  },
79
  {
80
  "seed": 17,
 
83
  "loss_train": 0.1242414116859436,
84
  "W3_MSE": 0.11465953538815181,
85
  "W4_MSE": 0.11387925470868747,
86
+ "ctrl_max_rms": 0.00030438662361120805
87
  },
88
  {
89
  "seed": 17,
 
91
  "step": 24,
92
  "loss_train": 0.09662917256355286,
93
  "W3_MSE": 0.11453648408253987,
94
+ "W4_MSE": 0.11368757858872414,
95
+ "ctrl_max_rms": 0.0003923155866990176
96
  },
97
  {
98
  "seed": 17,
 
101
  "loss_train": 0.1582513302564621,
102
  "W3_MSE": 0.1143090259283781,
103
  "W4_MSE": 0.11331315711140633,
104
+ "ctrl_max_rms": 0.0005296767437054465
105
  },
106
  {
107
  "seed": 17,
 
118
  "step": 48,
119
  "loss_train": 0.10963404178619385,
120
  "W3_MSE": 0.1130196029941241,
121
+ "W4_MSE": 0.1111137376477321,
122
+ "ctrl_max_rms": 0.002058053525009503
123
  },
124
  {
125
  "seed": 17,
126
  "arm": "stateless",
127
  "step": 56,
128
  "loss_train": 0.10562591254711151,
129
+ "W3_MSE": 0.1119188629090786,
130
  "W4_MSE": 0.10920579296847184,
131
  "ctrl_max_rms": 0.004299115234365066
132
  },
 
136
  "step": 64,
137
  "loss_train": 0.1330445259809494,
138
  "W3_MSE": 0.11103641552229722,
139
+ "W4_MSE": 0.10791260376572609,
140
+ "ctrl_max_rms": 0.005281291048352917
141
  },
142
  {
143
  "seed": 17,
 
177
  "arm": "mhc",
178
  "step": 8,
179
  "loss_train": 0.08606645464897156,
180
+ "W3_MSE": 0.1147225337723891,
181
  "W4_MSE": 0.113968084876736,
182
+ "ctrl_max_rms": 0.00018755902177266157
183
  },
184
  {
185
  "seed": 17,
 
188
  "loss_train": 0.1242414116859436,
189
  "W3_MSE": 0.11465953538815181,
190
  "W4_MSE": 0.11387925657133262,
191
+ "ctrl_max_rms": 0.00030442177133712295
192
  },
193
  {
194
  "seed": 17,
 
197
  "loss_train": 0.09662917256355286,
198
  "W3_MSE": 0.11453648408253987,
199
  "W4_MSE": 0.11368757796784242,
200
+ "ctrl_max_rms": 0.0003923258069941464
201
  },
202
  {
203
  "seed": 17,
 
206
  "loss_train": 0.1582513302564621,
207
  "W3_MSE": 0.11430902406573296,
208
  "W4_MSE": 0.11331315897405148,
209
+ "ctrl_max_rms": 0.0005288315490664294
210
  },
211
  {
212
  "seed": 17,
 
215
  "loss_train": 0.10655954480171204,
216
  "W3_MSE": 0.11380419631799062,
217
  "W4_MSE": 0.11249386705458164,
218
+ "ctrl_max_rms": 0.0009341221545279647
219
  },
220
  {
221
  "seed": 17,
222
  "arm": "mhc",
223
  "step": 48,
224
  "loss_train": 0.109634168446064,
225
+ "W3_MSE": 0.11301966632405917,
226
  "W4_MSE": 0.11111390093962352,
227
+ "ctrl_max_rms": 0.002017362198481957
228
  },
229
  {
230
  "seed": 17,
231
  "arm": "mhc",
232
  "step": 56,
233
  "loss_train": 0.10562613606452942,
234
+ "W3_MSE": 0.11191889395316441,
235
+ "W4_MSE": 0.1092058972766002,
236
+ "ctrl_max_rms": 0.004241040170503159
237
  },
238
  {
239
  "seed": 17,
 
242
  "loss_train": 0.13304445147514343,
243
  "W3_MSE": 0.11103636150558789,
244
  "W4_MSE": 0.10791248766084512,
245
+ "ctrl_max_rms": 0.005226058070547879
246
  },
247
  {
248
  "seed": 17,
 
284
  "loss_train": 0.08451367169618607,
285
  "W3_MSE": 0.11472250955800216,
286
  "W4_MSE": 0.11396805879970391,
287
+ "ctrl_max_rms": 0.0001874632143881172
288
  },
289
  {
290
  "seed": 17,
 
293
  "loss_train": 0.12192638218402863,
294
  "W3_MSE": 0.11465946398675442,
295
  "W4_MSE": 0.11387916964789231,
296
+ "ctrl_max_rms": 0.00030426468341223273
297
  },
298
  {
299
  "seed": 17,
 
302
  "loss_train": 0.09486307948827744,
303
  "W3_MSE": 0.11453630837301414,
304
  "W4_MSE": 0.1136873426536719,
305
+ "ctrl_max_rms": 0.0003920286981156096
306
  },
307
  {
308
  "seed": 17,
 
311
  "loss_train": 0.1553032398223877,
312
  "W3_MSE": 0.11430866395433743,
313
  "W4_MSE": 0.11331262377401193,
314
+ "ctrl_max_rms": 0.0005277996970107779
315
  },
316
  {
317
  "seed": 17,
 
320
  "loss_train": 0.10460616648197174,
321
  "W3_MSE": 0.11380349472165108,
322
  "W4_MSE": 0.1124927494674921,
323
+ "ctrl_max_rms": 0.0009292569399500886
324
  },
325
  {
326
  "seed": 17,
 
329
  "loss_train": 0.10761874914169312,
330
  "W3_MSE": 0.11301849285761516,
331
  "W4_MSE": 0.11111204698681831,
332
+ "ctrl_max_rms": 0.0020051325943010547
333
  },
334
  {
335
  "seed": 17,
336
  "arm": "mhc_noaux",
337
  "step": 56,
338
  "loss_train": 0.10368967801332474,
339
+ "W3_MSE": 0.11191730325420697,
340
  "W4_MSE": 0.10920340940356255,
341
+ "ctrl_max_rms": 0.004224800718172143
342
  },
343
  {
344
  "seed": 17,
345
  "arm": "mhc_noaux",
346
  "step": 64,
347
  "loss_train": 0.1305774450302124,
348
+ "W3_MSE": 0.11103417413930099,
349
+ "W4_MSE": 0.10790904176731904,
350
+ "ctrl_max_rms": 0.005209839281936486
351
  },
352
  {
353
  "seed": 41,
 
389
  "loss_train": 0.10454776883125305,
390
  "W3_MSE": 0.11467163388927777,
391
  "W4_MSE": 0.11390550496677558,
392
+ "ctrl_max_rms": 9.996155616439258e-05
393
+ },
394
+ {
395
+ "seed": 41,
396
+ "arm": "stateless",
397
+ "step": 16,
398
+ "loss_train": 0.09654892981052399,
399
+ "W3_MSE": 0.11463494785130024,
400
+ "W4_MSE": 0.1138407854984204,
401
+ "ctrl_max_rms": 0.0001616936715436168
402
  }
403
  ],
404
  "checkpoints": [
 
408
  "bytes": 93186
409
  },
410
  {
411
+ "sha256": "a13078dc643df7c8d28691c0174a8dd32d54f090f1bf45b927ff39b8651a156f",
412
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0008.pt",
413
  "bytes": 268976
414
  },
415
  {
416
+ "sha256": "931cf4a39db5a1dc51cd106afc874d157c15e0d61d775ab5d0cce63ac6638724",
417
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0016.pt",
418
  "bytes": 268976
419
  },
420
  {
421
+ "sha256": "e682bcc3339512ea4c8b6ca78b772ba77b523052f88d030f8205b3052a8b4e19",
422
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0024.pt",
423
  "bytes": 268976
424
  },
425
  {
426
+ "sha256": "f872e196b3e38155833df2fb4fab642093aa25f0ccff31ec6fe7f9001f3bed72",
427
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0032.pt",
428
  "bytes": 268976
429
  },
430
  {
431
+ "sha256": "a1f47b6860f76d86c226ee6b40207f80bda831c96079d9e8ed7e9df0785b320c",
432
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0040.pt",
433
  "bytes": 268976
434
  },
435
  {
436
+ "sha256": "5faa3709c19c3c8f5bf46b20a968d9edfe5379f02ffc4ff845bdc26f3f885e8c",
437
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0048.pt",
438
  "bytes": 268976
439
  },
440
  {
441
+ "sha256": "bd2526fced8e5ec09ad957fec56eb5138505edaacfd40c8ffc0335f0748051c6",
442
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0056.pt",
443
  "bytes": 268976
444
  },
445
  {
446
+ "sha256": "bcbf846ee11028b6446dd47f735a4eee5539c646622d22c71909db9fb5e156ec",
447
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed17_stateless_step0064.pt",
448
  "bytes": 268976
449
  },
 
546
  "sha256": "d934489f5dcecd5ae7d831b52e1ca108c4866f2028f8b338745d3986bc4f8cd6",
547
  "hf_path": "experiments/MHC_REAL_A1/pilot_seed41_stateless_step0008.pt",
548
  "bytes": 268976
549
+ },
550
+ {
551
+ "sha256": "e5d497263f3240df7bc9f8984dfdaea0c41a12763190413868ebf92da29fc2af",
552
+ "hf_path": "experiments/MHC_REAL_A1/pilot_seed41_stateless_step0016.pt",
553
+ "bytes": 268976
554
  }
555
  ]
556
  }