Hugging Face's logo

AmberYifan
/
Mistral-7B-v0.3-dpo-2k-hhrlhf

Safetensors
mistral
Generated from Trainer
Model card Files Community
AmberYifan commited on
Commit
f7913d6
•
1 Parent(s): 9923d24

Model save

Browse files
Files changed (7) hide show
  1. all_results.json +4 -4
  2. model-00001-of-00003.safetensors +1 -1
  3. model-00002-of-00003.safetensors +1 -1
  4. model-00003-of-00003.safetensors +1 -1
  5. train_results.json +4 -4
  6. trainer_state.json +70 -70
  7. training_args.bin +1 -1
all_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
- "train_loss": 0.7656974850162384,
5
- "train_runtime": 734.9736,
6
  "train_samples": 1999,
7
- "train_samples_per_second": 2.72,
8
- "train_steps_per_second": 0.084
9
  }
 
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
+ "train_loss": 0.7465078599991337,
5
+ "train_runtime": 792.7009,
6
  "train_samples": 1999,
7
+ "train_samples_per_second": 2.522,
8
+ "train_steps_per_second": 0.078
9
  }
model-00001-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5882b66b8fd21506d90f2d3d4ba8d168c5b96323e226a259bbc171a11d9f14fc
3
  size 4949453792
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b537c8fc3cbc4353b321fb0a15deeb9e90e1c350763638d17dd6ce46256baf8
3
  size 4949453792
model-00002-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8c997ea33b394db6af240abbed1873fa3143bb3d8f046c1e4f907a1472cb088e
3
  size 4999819336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a76bd3942d7aebeb4df4b7782e820f11f1c299e93d219172a68028e4ce689b66
3
  size 4999819336
model-00003-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c4bc0335b49d0f77223edcb3db5e163258455da12cf21522554a729b2e4f760b
3
  size 4546807800
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd49c2596f0a7887ab24843ac411f3166d5070bd9893211a4cf4dcc1516c9884
3
  size 4546807800
train_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
- "train_loss": 0.7656974850162384,
5
- "train_runtime": 734.9736,
6
  "train_samples": 1999,
7
- "train_samples_per_second": 2.72,
8
- "train_steps_per_second": 0.084
9
  }
 
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
+ "train_loss": 0.7465078599991337,
5
+ "train_runtime": 792.7009,
6
  "train_samples": 1999,
7
+ "train_samples_per_second": 2.522,
8
+ "train_steps_per_second": 0.078
9
  }
trainer_state.json CHANGED
@@ -10,13 +10,13 @@
10
  "log_history": [
11
  {
12
  "epoch": 0.016,
13
- "grad_norm": 48.370780385847006,
14
  "learning_rate": 7.142857142857142e-08,
15
- "logits/generated": -2.488245725631714,
16
- "logits/real": -2.398932933807373,
17
- "logps/generated": -95.94135284423828,
18
- "logps/real": -88.03836822509766,
19
- "loss": 0.788,
20
  "rewards/accuracies": 0.0,
21
  "rewards/generated": 0.0,
22
  "rewards/margins": 0.0,
@@ -25,102 +25,102 @@
25
  },
26
  {
27
  "epoch": 0.16,
28
- "grad_norm": 42.048520361743655,
29
  "learning_rate": 4.727272727272727e-07,
30
- "logits/generated": -2.621403694152832,
31
- "logits/real": -2.6067636013031006,
32
- "logps/generated": -116.70565032958984,
33
- "logps/real": -118.31328582763672,
34
- "loss": 0.8056,
35
- "rewards/accuracies": 0.2638888955116272,
36
- "rewards/generated": -0.05396932363510132,
37
- "rewards/margins": 0.051140956580638885,
38
- "rewards/real": -0.0028283733408898115,
39
  "step": 10
40
  },
41
  {
42
  "epoch": 0.32,
43
- "grad_norm": 55.87242539471649,
44
  "learning_rate": 3.818181818181818e-07,
45
- "logits/generated": -2.76572585105896,
46
- "logits/real": -2.746518611907959,
47
- "logps/generated": -107.32762145996094,
48
- "logps/real": -98.75493621826172,
49
- "loss": 0.7706,
50
- "rewards/accuracies": 0.22499999403953552,
51
- "rewards/generated": 0.1816125512123108,
52
- "rewards/margins": 0.10470892488956451,
53
- "rewards/real": 0.2863214612007141,
54
  "step": 20
55
  },
56
  {
57
  "epoch": 0.48,
58
- "grad_norm": 62.79136307414435,
59
  "learning_rate": 2.909090909090909e-07,
60
- "logits/generated": -2.794795036315918,
61
- "logits/real": -2.759561777114868,
62
- "logps/generated": -101.0676498413086,
63
- "logps/real": -92.65696716308594,
64
- "loss": 0.7564,
65
- "rewards/accuracies": 0.21250000596046448,
66
- "rewards/generated": 0.2623925805091858,
67
- "rewards/margins": 0.08424563705921173,
68
- "rewards/real": 0.3466382622718811,
69
  "step": 30
70
  },
71
  {
72
  "epoch": 0.64,
73
- "grad_norm": 68.38298514315854,
74
  "learning_rate": 2e-07,
75
- "logits/generated": -2.708467960357666,
76
- "logits/real": -2.7237162590026855,
77
- "logps/generated": -132.64068603515625,
78
- "logps/real": -121.7768325805664,
79
- "loss": 0.7587,
80
- "rewards/accuracies": 0.30000001192092896,
81
- "rewards/generated": 0.3463793098926544,
82
- "rewards/margins": 0.4203476905822754,
83
- "rewards/real": 0.7667269706726074,
84
  "step": 40
85
  },
86
  {
87
  "epoch": 0.8,
88
- "grad_norm": 59.06365958191896,
89
  "learning_rate": 1.0909090909090908e-07,
90
- "logits/generated": -2.716890573501587,
91
- "logits/real": -2.7266154289245605,
92
- "logps/generated": -113.8817367553711,
93
- "logps/real": -107.4018783569336,
94
- "loss": 0.7599,
95
- "rewards/accuracies": 0.30000001192092896,
96
- "rewards/generated": 0.9875177145004272,
97
- "rewards/margins": 0.26255884766578674,
98
- "rewards/real": 1.2500765323638916,
99
  "step": 50
100
  },
101
  {
102
  "epoch": 0.96,
103
- "grad_norm": 45.68197240709943,
104
  "learning_rate": 1.818181818181818e-08,
105
- "logits/generated": -2.747610569000244,
106
- "logits/real": -2.718883991241455,
107
- "logps/generated": -101.81327819824219,
108
- "logps/real": -84.88102722167969,
109
- "loss": 0.7545,
110
- "rewards/accuracies": 0.375,
111
- "rewards/generated": 1.2750626802444458,
112
- "rewards/margins": 0.2342238873243332,
113
- "rewards/real": 1.5092865228652954,
114
  "step": 60
115
  },
116
  {
117
  "epoch": 0.992,
118
  "step": 62,
119
  "total_flos": 0.0,
120
- "train_loss": 0.7656974850162384,
121
- "train_runtime": 734.9736,
122
- "train_samples_per_second": 2.72,
123
- "train_steps_per_second": 0.084
124
  }
125
  ],
126
  "logging_steps": 10,
 
10
  "log_history": [
11
  {
12
  "epoch": 0.016,
13
+ "grad_norm": 33.26401254999367,
14
  "learning_rate": 7.142857142857142e-08,
15
+ "logits/generated": -2.5860352516174316,
16
+ "logits/real": -2.4361655712127686,
17
+ "logps/generated": -109.8505859375,
18
+ "logps/real": -99.55709838867188,
19
+ "loss": 0.7803,
20
  "rewards/accuracies": 0.0,
21
  "rewards/generated": 0.0,
22
  "rewards/margins": 0.0,
 
25
  },
26
  {
27
  "epoch": 0.16,
28
+ "grad_norm": 40.749062783665345,
29
  "learning_rate": 4.727272727272727e-07,
30
+ "logits/generated": -2.6749789714813232,
31
+ "logits/real": -2.711810350418091,
32
+ "logps/generated": -103.34049987792969,
33
+ "logps/real": -90.66702270507812,
34
+ "loss": 0.7685,
35
+ "rewards/accuracies": 0.25,
36
+ "rewards/generated": 0.17951178550720215,
37
+ "rewards/margins": 0.16819272935390472,
38
+ "rewards/real": 0.34770452976226807,
39
  "step": 10
40
  },
41
  {
42
  "epoch": 0.32,
43
+ "grad_norm": 62.39690097528854,
44
  "learning_rate": 3.818181818181818e-07,
45
+ "logits/generated": -2.8889403343200684,
46
+ "logits/real": -2.8932297229766846,
47
+ "logps/generated": -112.69966125488281,
48
+ "logps/real": -104.17449951171875,
49
+ "loss": 0.7714,
50
+ "rewards/accuracies": 0.2750000059604645,
51
+ "rewards/generated": -0.4469524323940277,
52
+ "rewards/margins": 0.310979425907135,
53
+ "rewards/real": -0.1359730362892151,
54
  "step": 20
55
  },
56
  {
57
  "epoch": 0.48,
58
+ "grad_norm": 70.93062255354987,
59
  "learning_rate": 2.909090909090909e-07,
60
+ "logits/generated": -3.0333216190338135,
61
+ "logits/real": -3.0492300987243652,
62
+ "logps/generated": -98.00860595703125,
63
+ "logps/real": -83.57305908203125,
64
+ "loss": 0.7209,
65
+ "rewards/accuracies": 0.22499999403953552,
66
+ "rewards/generated": 0.3839380145072937,
67
+ "rewards/margins": 0.18280868232250214,
68
+ "rewards/real": 0.566746711730957,
69
  "step": 30
70
  },
71
  {
72
  "epoch": 0.64,
73
+ "grad_norm": 55.19518628250898,
74
  "learning_rate": 2e-07,
75
+ "logits/generated": -3.0203452110290527,
76
+ "logits/real": -3.0042731761932373,
77
+ "logps/generated": -129.9265594482422,
78
+ "logps/real": -99.64141845703125,
79
+ "loss": 0.7566,
80
+ "rewards/accuracies": 0.23749999701976776,
81
+ "rewards/generated": 0.8018711805343628,
82
+ "rewards/margins": 0.49393320083618164,
83
+ "rewards/real": 1.2958043813705444,
84
  "step": 40
85
  },
86
  {
87
  "epoch": 0.8,
88
+ "grad_norm": 21.78570359935915,
89
  "learning_rate": 1.0909090909090908e-07,
90
+ "logits/generated": -2.9519670009613037,
91
+ "logits/real": -2.9463257789611816,
92
+ "logps/generated": -100.23809814453125,
93
+ "logps/real": -82.25651550292969,
94
+ "loss": 0.7277,
95
+ "rewards/accuracies": 0.22499999403953552,
96
+ "rewards/generated": 1.315015435218811,
97
+ "rewards/margins": 0.4744125306606293,
98
+ "rewards/real": 1.7894279956817627,
99
  "step": 50
100
  },
101
  {
102
  "epoch": 0.96,
103
+ "grad_norm": 55.38490015693705,
104
  "learning_rate": 1.818181818181818e-08,
105
+ "logits/generated": -2.898479461669922,
106
+ "logits/real": -2.902143955230713,
107
+ "logps/generated": -86.09727478027344,
108
+ "logps/real": -77.5208511352539,
109
+ "loss": 0.7312,
110
+ "rewards/accuracies": 0.13750000298023224,
111
+ "rewards/generated": 1.5982306003570557,
112
+ "rewards/margins": 0.5042404532432556,
113
+ "rewards/real": 2.102470874786377,
114
  "step": 60
115
  },
116
  {
117
  "epoch": 0.992,
118
  "step": 62,
119
  "total_flos": 0.0,
120
+ "train_loss": 0.7465078599991337,
121
+ "train_runtime": 792.7009,
122
+ "train_samples_per_second": 2.522,
123
+ "train_steps_per_second": 0.078
124
  }
125
  ],
126
  "logging_steps": 10,
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9400c02dd38fcda38f13224f595871c3c246419370c5731f1241b03bcbe1e372
3
  size 6392
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc76299eccde211c3f8a9110323214c6d158f971238e01628c590b9aa6f2d524
3
  size 6392