Hugging Face's logo

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

Safetensors
mistral
Generated from Trainer
Model card Files Community
AmberYifan commited on
Commit
d60c9b6
•
1 Parent(s): 4cbee17

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": 0.7342928696063257,
5
- "train_runtime": 746.6525,
6
  "train_samples": 1999,
7
- "train_samples_per_second": 2.677,
8
- "train_steps_per_second": 0.083
9
  }
 
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
+ "train_loss": 0.648834865900778,
5
+ "train_runtime": 793.7409,
6
  "train_samples": 1999,
7
+ "train_samples_per_second": 2.518,
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:ffb4a031e075be9569c3eee1e468a2abf7638cadaf48a0a1fd1bda488608f91a
3
  size 4949453792
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:118d64adaabcd562feba4892db17c8144cc53a4f944b473dfef44f29d5e2d183
3
  size 4949453792
model-00002-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:69eddf625b7f8dbae204ab4a66f676b6e799a5385778ebe796d486e613570ecd
3
  size 4999819336
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f6396df81e7edc14a4163cd2fa3cdbc13a645e55f045c14b3d18164e6e8a280
3
  size 4999819336
model-00003-of-00003.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3b725b9bf68a01c5a65b51e9f01ab33c1e94bed2c58a7d93e8e2f47682e35fb3
3
  size 4546807800
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f368260fe3431ff517de51d9330d0b0655b7c18451791898b3da28620d7aa6b7
3
  size 4546807800
train_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
- "train_loss": 0.7342928696063257,
5
- "train_runtime": 746.6525,
6
  "train_samples": 1999,
7
- "train_samples_per_second": 2.677,
8
- "train_steps_per_second": 0.083
9
  }
 
1
  {
2
  "epoch": 0.992,
3
  "total_flos": 0.0,
4
+ "train_loss": 0.648834865900778,
5
+ "train_runtime": 793.7409,
6
  "train_samples": 1999,
7
+ "train_samples_per_second": 2.518,
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": 198.7469829190425,
14
  "learning_rate": 7.142857142857142e-08,
15
- "logits/generated": -2.8518388271331787,
16
- "logits/real": -2.9862747192382812,
17
- "logps/generated": -241.12493896484375,
18
- "logps/real": -248.8436279296875,
19
- "loss": 0.9204,
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": 191.54263414525988,
29
  "learning_rate": 4.727272727272727e-07,
30
- "logits/generated": -3.05552339553833,
31
- "logits/real": -3.0378544330596924,
32
- "logps/generated": -243.76849365234375,
33
- "logps/real": -230.84719848632812,
34
- "loss": 0.8961,
35
- "rewards/accuracies": 0.5833333134651184,
36
- "rewards/generated": 0.18684834241867065,
37
- "rewards/margins": 0.20998667180538177,
38
- "rewards/real": 0.396835058927536,
39
  "step": 10
40
  },
41
  {
42
  "epoch": 0.32,
43
- "grad_norm": 139.26955693169643,
44
  "learning_rate": 3.818181818181818e-07,
45
- "logits/generated": -2.9985575675964355,
46
- "logits/real": -2.984205484390259,
47
- "logps/generated": -250.49853515625,
48
- "logps/real": -217.1739044189453,
49
- "loss": 0.8668,
50
- "rewards/accuracies": 0.675000011920929,
51
- "rewards/generated": -1.4985097646713257,
52
- "rewards/margins": 0.8482303619384766,
53
- "rewards/real": -0.6502794623374939,
54
  "step": 20
55
  },
56
  {
57
  "epoch": 0.48,
58
- "grad_norm": 117.51802509366077,
59
  "learning_rate": 2.909090909090909e-07,
60
- "logits/generated": -2.9175362586975098,
61
- "logits/real": -2.964449644088745,
62
- "logps/generated": -262.0832824707031,
63
- "logps/real": -228.2645721435547,
64
- "loss": 0.7592,
65
- "rewards/accuracies": 0.8125,
66
- "rewards/generated": -1.5061546564102173,
67
- "rewards/margins": 1.8053114414215088,
68
- "rewards/real": 0.29915687441825867,
69
  "step": 30
70
  },
71
  {
72
  "epoch": 0.64,
73
- "grad_norm": 176.64698900823225,
74
  "learning_rate": 2e-07,
75
- "logits/generated": -2.9176225662231445,
76
- "logits/real": -2.9303154945373535,
77
- "logps/generated": -260.50018310546875,
78
- "logps/real": -230.49288940429688,
79
- "loss": 0.6242,
80
- "rewards/accuracies": 0.824999988079071,
81
- "rewards/generated": -2.2268033027648926,
82
- "rewards/margins": 2.359907627105713,
83
- "rewards/real": 0.13310480117797852,
84
  "step": 40
85
  },
86
  {
87
  "epoch": 0.8,
88
- "grad_norm": 183.16237322543932,
89
  "learning_rate": 1.0909090909090908e-07,
90
- "logits/generated": -2.88456654548645,
91
- "logits/real": -2.944023609161377,
92
- "logps/generated": -260.68035888671875,
93
- "logps/real": -228.1358184814453,
94
- "loss": 0.604,
95
- "rewards/accuracies": 0.800000011920929,
96
- "rewards/generated": -1.7078344821929932,
97
- "rewards/margins": 1.941307783126831,
98
- "rewards/real": 0.23347334563732147,
99
  "step": 50
100
  },
101
  {
102
  "epoch": 0.96,
103
- "grad_norm": 95.0210404383474,
104
  "learning_rate": 1.818181818181818e-08,
105
- "logits/generated": -2.917457342147827,
106
- "logits/real": -2.9479122161865234,
107
- "logps/generated": -249.3968048095703,
108
- "logps/real": -225.28262329101562,
109
- "loss": 0.6543,
110
- "rewards/accuracies": 0.824999988079071,
111
- "rewards/generated": -1.3457591533660889,
112
- "rewards/margins": 1.6494863033294678,
113
- "rewards/real": 0.30372729897499084,
114
  "step": 60
115
  },
116
  {
117
  "epoch": 0.992,
118
  "step": 62,
119
  "total_flos": 0.0,
120
- "train_loss": 0.7342928696063257,
121
- "train_runtime": 746.6525,
122
- "train_samples_per_second": 2.677,
123
- "train_steps_per_second": 0.083
124
  }
125
  ],
126
  "logging_steps": 10,
 
10
  "log_history": [
11
  {
12
  "epoch": 0.016,
13
+ "grad_norm": 228.67711231286552,
14
  "learning_rate": 7.142857142857142e-08,
15
+ "logits/generated": -3.182648181915283,
16
+ "logits/real": -3.236078977584839,
17
+ "logps/generated": -231.1735382080078,
18
+ "logps/real": -216.56446838378906,
19
+ "loss": 0.914,
20
  "rewards/accuracies": 0.0,
21
  "rewards/generated": 0.0,
22
  "rewards/margins": 0.0,
 
25
  },
26
  {
27
  "epoch": 0.16,
28
+ "grad_norm": 104.82372723323877,
29
  "learning_rate": 4.727272727272727e-07,
30
+ "logits/generated": -3.0453360080718994,
31
+ "logits/real": -3.0438425540924072,
32
+ "logps/generated": -241.46511840820312,
33
+ "logps/real": -229.99560546875,
34
+ "loss": 0.842,
35
+ "rewards/accuracies": 0.5694444179534912,
36
+ "rewards/generated": -0.5092797875404358,
37
+ "rewards/margins": 0.3939414620399475,
38
+ "rewards/real": -0.11533831059932709,
39
  "step": 10
40
  },
41
  {
42
  "epoch": 0.32,
43
+ "grad_norm": 160.6471422793089,
44
  "learning_rate": 3.818181818181818e-07,
45
+ "logits/generated": -3.0548062324523926,
46
+ "logits/real": -2.970714569091797,
47
+ "logps/generated": -253.4170379638672,
48
+ "logps/real": -231.45254516601562,
49
+ "loss": 0.7195,
50
+ "rewards/accuracies": 0.800000011920929,
51
+ "rewards/generated": -2.3348288536071777,
52
+ "rewards/margins": 1.3843599557876587,
53
+ "rewards/real": -0.9504690170288086,
54
  "step": 20
55
  },
56
  {
57
  "epoch": 0.48,
58
+ "grad_norm": 146.55739161612283,
59
  "learning_rate": 2.909090909090909e-07,
60
+ "logits/generated": -3.0757832527160645,
61
+ "logits/real": -3.142075300216675,
62
+ "logps/generated": -258.6484680175781,
63
+ "logps/real": -223.58065795898438,
64
+ "loss": 0.5971,
65
+ "rewards/accuracies": 0.8500000238418579,
66
+ "rewards/generated": -1.7125927209854126,
67
+ "rewards/margins": 1.9861854314804077,
68
+ "rewards/real": 0.2735927999019623,
69
  "step": 30
70
  },
71
  {
72
  "epoch": 0.64,
73
+ "grad_norm": 102.71447615342244,
74
  "learning_rate": 2e-07,
75
+ "logits/generated": -2.938872814178467,
76
+ "logits/real": -2.982527017593384,
77
+ "logps/generated": -247.6828155517578,
78
+ "logps/real": -228.6986541748047,
79
+ "loss": 0.5605,
80
+ "rewards/accuracies": 0.8374999761581421,
81
+ "rewards/generated": -2.5453975200653076,
82
+ "rewards/margins": 1.9871200323104858,
83
+ "rewards/real": -0.558277428150177,
84
  "step": 40
85
  },
86
  {
87
  "epoch": 0.8,
88
+ "grad_norm": 128.82542081084267,
89
  "learning_rate": 1.0909090909090908e-07,
90
+ "logits/generated": -2.9596309661865234,
91
+ "logits/real": -3.0371811389923096,
92
+ "logps/generated": -258.96600341796875,
93
+ "logps/real": -224.0094757080078,
94
+ "loss": 0.6154,
95
+ "rewards/accuracies": 0.8999999761581421,
96
+ "rewards/generated": -2.521024465560913,
97
+ "rewards/margins": 2.7490367889404297,
98
+ "rewards/real": 0.22801223397254944,
99
  "step": 50
100
  },
101
  {
102
  "epoch": 0.96,
103
+ "grad_norm": 140.9790674306219,
104
  "learning_rate": 1.818181818181818e-08,
105
+ "logits/generated": -2.8980393409729004,
106
+ "logits/real": -3.0217809677124023,
107
+ "logps/generated": -256.95941162109375,
108
+ "logps/real": -225.6009063720703,
109
+ "loss": 0.491,
110
+ "rewards/accuracies": 0.7875000238418579,
111
+ "rewards/generated": -2.6129565238952637,
112
+ "rewards/margins": 2.7330925464630127,
113
+ "rewards/real": 0.12013578414916992,
114
  "step": 60
115
  },
116
  {
117
  "epoch": 0.992,
118
  "step": 62,
119
  "total_flos": 0.0,
120
+ "train_loss": 0.648834865900778,
121
+ "train_runtime": 793.7409,
122
+ "train_samples_per_second": 2.518,
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:e380477d354c1079bad53f011ae55543312cf994f5c48739473bec0dd2d24758
3
  size 6456
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11abd944a7b0c9886b30dcf46efc0f6e3d502124d5969c6240c2e6b8edf142da
3
  size 6456