Hugging Face's logo

AmberYifan
/
Mistral-7B-v0.1-dpo-2k-hhrlhf

Safetensors
mistral
Generated from Trainer
Model card Files Community
AmberYifan commited on
Commit
776a901
•
1 Parent(s): 814eeb4

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 +69 -69
  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": 0.7687068062443887,
5
- "train_runtime": 747.5876,
6
  "train_samples": 1999,
7
- "train_samples_per_second": 2.674,
8
- "train_steps_per_second": 0.083
9
  }
 
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
+ "train_loss": 0.7473330401605175,
5
+ "train_runtime": 796.1787,
6
  "train_samples": 1999,
7
+ "train_samples_per_second": 2.511,
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 sha256:1c2b9bdb866fe09e4f9fb5e0f4cdc1c7f7a11f2068dfd2938eac0baaa24199c3
3
  size 4943162336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e3a19f872e7fb0e19870aba76ae3f5c9223b90ddf54797691e9076f6b9ab12e
3
  size 4943162336
model-00002-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6b2b14331f50809e47b874a6e162fa925e63750dd557297660ad89f89dc3f7d6
3
  size 4999819336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:59af0b2cca6f97a2cee171268613b799dfd21cfeca1485da4630ca5d2eac0ec5
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:f4ecb43cb776e8ce67c7b7134790207215a0e51655651fbc28f0546750b0c333
3
  size 4540516344
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01242609eebd7a9d5c5bacafa974af69e25c4af122fcd83365210d342ac88bd7
3
  size 4540516344
train_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
- "train_loss": 0.7687068062443887,
5
- "train_runtime": 747.5876,
6
  "train_samples": 1999,
7
- "train_samples_per_second": 2.674,
8
- "train_steps_per_second": 0.083
9
  }
 
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
+ "train_loss": 0.7473330401605175,
5
+ "train_runtime": 796.1787,
6
  "train_samples": 1999,
7
+ "train_samples_per_second": 2.511,
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": 52.54359963110853,
14
  "learning_rate": 7.142857142857142e-08,
15
- "logits/generated": -2.6779448986053467,
16
- "logits/real": -2.691202163696289,
17
- "logps/generated": -74.53842163085938,
18
- "logps/real": -93.44010925292969,
19
- "loss": 0.7913,
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": 28.122526927341976,
29
  "learning_rate": 4.727272727272727e-07,
30
- "logits/generated": -2.5826447010040283,
31
- "logits/real": -2.572615385055542,
32
- "logps/generated": -123.79043579101562,
33
- "logps/real": -108.88796997070312,
34
- "loss": 0.7974,
35
- "rewards/accuracies": 0.25,
36
- "rewards/generated": -0.017271483317017555,
37
- "rewards/margins": 0.0596686527132988,
38
- "rewards/real": 0.04239716753363609,
39
  "step": 10
40
  },
41
  {
42
  "epoch": 0.32,
43
- "grad_norm": 44.434349792005534,
44
  "learning_rate": 3.818181818181818e-07,
45
- "logits/generated": -2.709092378616333,
46
- "logits/real": -2.6981873512268066,
47
- "logps/generated": -99.3179702758789,
48
- "logps/real": -96.93788146972656,
49
- "loss": 0.7883,
50
- "rewards/accuracies": 0.25,
51
- "rewards/generated": 1.0143389701843262,
52
- "rewards/margins": 0.029613429680466652,
53
- "rewards/real": 1.043952226638794,
54
  "step": 20
55
  },
56
  {
57
  "epoch": 0.48,
58
- "grad_norm": 45.15314438688675,
59
  "learning_rate": 2.909090909090909e-07,
60
- "logits/generated": -2.633875846862793,
61
- "logits/real": -2.6102986335754395,
62
- "logps/generated": -110.6008071899414,
63
- "logps/real": -97.6747817993164,
64
- "loss": 0.7362,
65
- "rewards/accuracies": 0.30000001192092896,
66
- "rewards/generated": 0.7923309206962585,
67
- "rewards/margins": 0.15161481499671936,
68
- "rewards/real": 0.9439458847045898,
69
  "step": 30
70
  },
71
  {
72
  "epoch": 0.64,
73
- "grad_norm": 65.12954197470013,
74
  "learning_rate": 2e-07,
75
- "logits/generated": -2.7233123779296875,
76
- "logits/real": -2.6725680828094482,
77
- "logps/generated": -126.6224594116211,
78
- "logps/real": -109.69065856933594,
79
- "loss": 0.7946,
80
- "rewards/accuracies": 0.32499998807907104,
81
- "rewards/generated": 0.30916541814804077,
82
- "rewards/margins": 0.30192217230796814,
83
- "rewards/real": 0.6110875010490417,
84
  "step": 40
85
  },
86
  {
87
  "epoch": 0.8,
88
- "grad_norm": 54.78449313424441,
89
  "learning_rate": 1.0909090909090908e-07,
90
- "logits/generated": -2.696377992630005,
91
- "logits/real": -2.7040600776672363,
92
- "logps/generated": -105.97816467285156,
93
- "logps/real": -98.66194915771484,
94
- "loss": 0.7688,
95
- "rewards/accuracies": 0.16249999403953552,
96
- "rewards/generated": 1.6170318126678467,
97
- "rewards/margins": 0.025405455380678177,
98
- "rewards/real": 1.6424373388290405,
99
  "step": 50
100
  },
101
  {
102
  "epoch": 0.96,
103
- "grad_norm": 34.743799756626075,
104
  "learning_rate": 1.818181818181818e-08,
105
- "logits/generated": -2.609443187713623,
106
- "logits/real": -2.639867067337036,
107
- "logps/generated": -116.21678161621094,
108
- "logps/real": -96.13401794433594,
109
- "loss": 0.7323,
110
  "rewards/accuracies": 0.2750000059604645,
111
- "rewards/generated": 1.5233509540557861,
112
- "rewards/margins": 0.1510307937860489,
113
- "rewards/real": 1.6743818521499634,
114
  "step": 60
115
  },
116
  {
117
  "epoch": 0.992,
118
  "step": 62,
119
  "total_flos": 0.0,
120
- "train_loss": 0.7687068062443887,
121
- "train_runtime": 747.5876,
122
- "train_samples_per_second": 2.674,
123
- "train_steps_per_second": 0.083
124
  }
125
  ],
126
  "logging_steps": 10,
 
10
  "log_history": [
11
  {
12
  "epoch": 0.016,
13
+ "grad_norm": 30.93577209985028,
14
  "learning_rate": 7.142857142857142e-08,
15
+ "logits/generated": -2.6151156425476074,
16
+ "logits/real": -2.56895112991333,
17
+ "logps/generated": -100.78079223632812,
18
+ "logps/real": -84.83761596679688,
19
+ "loss": 0.7771,
20
  "rewards/accuracies": 0.0,
21
  "rewards/generated": 0.0,
22
  "rewards/margins": 0.0,
 
25
  },
26
  {
27
  "epoch": 0.16,
28
+ "grad_norm": 53.91196703984471,
29
  "learning_rate": 4.727272727272727e-07,
30
+ "logits/generated": -2.4759128093719482,
31
+ "logits/real": -2.5104005336761475,
32
+ "logps/generated": -110.86439514160156,
33
+ "logps/real": -96.45692443847656,
34
+ "loss": 0.7744,
35
+ "rewards/accuracies": 0.1944444477558136,
36
+ "rewards/generated": 0.4696773886680603,
37
+ "rewards/margins": 0.184379443526268,
38
+ "rewards/real": 0.6540568470954895,
39
  "step": 10
40
  },
41
  {
42
  "epoch": 0.32,
43
+ "grad_norm": 41.533419926208836,
44
  "learning_rate": 3.818181818181818e-07,
45
+ "logits/generated": -2.7200608253479004,
46
+ "logits/real": -2.7094995975494385,
47
+ "logps/generated": -111.71916198730469,
48
+ "logps/real": -97.46357727050781,
49
+ "loss": 0.7549,
50
+ "rewards/accuracies": 0.16249999403953552,
51
+ "rewards/generated": 1.2888102531433105,
52
+ "rewards/margins": 0.058855485171079636,
53
+ "rewards/real": 1.347665548324585,
54
  "step": 20
55
  },
56
  {
57
  "epoch": 0.48,
58
+ "grad_norm": 28.929547003454825,
59
  "learning_rate": 2.909090909090909e-07,
60
+ "logits/generated": -2.762415885925293,
61
+ "logits/real": -2.7689743041992188,
62
+ "logps/generated": -84.8138656616211,
63
+ "logps/real": -86.74251556396484,
64
+ "loss": 0.7391,
65
+ "rewards/accuracies": 0.23749999701976776,
66
+ "rewards/generated": 1.5304011106491089,
67
+ "rewards/margins": 0.12500552833080292,
68
+ "rewards/real": 1.6554067134857178,
69
  "step": 30
70
  },
71
  {
72
  "epoch": 0.64,
73
+ "grad_norm": 27.557827905589537,
74
  "learning_rate": 2e-07,
75
+ "logits/generated": -2.7791144847869873,
76
+ "logits/real": -2.817704916000366,
77
+ "logps/generated": -101.5577163696289,
78
+ "logps/real": -89.77957153320312,
79
+ "loss": 0.7543,
80
+ "rewards/accuracies": 0.20000000298023224,
81
+ "rewards/generated": 1.3332226276397705,
82
+ "rewards/margins": 0.20582261681556702,
83
+ "rewards/real": 1.5390453338623047,
84
  "step": 40
85
  },
86
  {
87
  "epoch": 0.8,
88
+ "grad_norm": 38.68815181578532,
89
  "learning_rate": 1.0909090909090908e-07,
90
+ "logits/generated": -2.7200920581817627,
91
+ "logits/real": -2.7181735038757324,
92
+ "logps/generated": -101.0867691040039,
93
+ "logps/real": -86.67424011230469,
94
+ "loss": 0.7238,
95
+ "rewards/accuracies": 0.23749999701976776,
96
+ "rewards/generated": 1.2783713340759277,
97
+ "rewards/margins": 0.4123215675354004,
98
+ "rewards/real": 1.6906929016113281,
99
  "step": 50
100
  },
101
  {
102
  "epoch": 0.96,
103
+ "grad_norm": 27.90637817594034,
104
  "learning_rate": 1.818181818181818e-08,
105
+ "logits/generated": -2.6357321739196777,
106
+ "logits/real": -2.649296522140503,
107
+ "logps/generated": -121.07977294921875,
108
+ "logps/real": -98.04679107666016,
109
+ "loss": 0.732,
110
  "rewards/accuracies": 0.2750000059604645,
111
+ "rewards/generated": 0.9751936197280884,
112
+ "rewards/margins": 0.5532658100128174,
113
+ "rewards/real": 1.5284594297409058,
114
  "step": 60
115
  },
116
  {
117
  "epoch": 0.992,
118
  "step": 62,
119
  "total_flos": 0.0,
120
+ "train_loss": 0.7473330401605175,
121
+ "train_runtime": 796.1787,
122
+ "train_samples_per_second": 2.511,
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 sha256:159fef2ce51b22c7f5d1faead1105882d5e9288a7d66381f5668c6e5ecd11177
3
  size 6392
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a195599b62489affb65b07d3476c850902409ec25a72a42aac6a5edc1c41eb2b
3
  size 6392