zfz1 commited on
Commit
9adce4c
1 Parent(s): b04eac5

Model save

Browse files
README.md ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ base_model: deepseek-ai/deepseek-math-7b-base
4
+ tags:
5
+ - trl
6
+ - orpo
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: deepseek-8b-orpo-full
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ [<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/thuzfz1/huggingface/runs/dtbyb3v7)
17
+ # deepseek-8b-orpo-full
18
+
19
+ This model is a fine-tuned version of [deepseek-ai/deepseek-math-7b-base](https://huggingface.co/deepseek-ai/deepseek-math-7b-base) on an unknown dataset.
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 2e-05
39
+ - train_batch_size: 4
40
+ - eval_batch_size: 4
41
+ - seed: 43
42
+ - distributed_type: multi-GPU
43
+ - num_devices: 4
44
+ - gradient_accumulation_steps: 2
45
+ - total_train_batch_size: 32
46
+ - total_eval_batch_size: 16
47
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
+ - lr_scheduler_type: cosine
49
+ - lr_scheduler_warmup_ratio: 0.1
50
+ - num_epochs: 2
51
+
52
+ ### Training results
53
+
54
+
55
+
56
+ ### Framework versions
57
+
58
+ - Transformers 4.42.3
59
+ - Pytorch 2.1.2
60
+ - Datasets 2.20.0
61
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.0,
3
+ "total_flos": 0.0,
4
+ "train_loss": 0.2130171504020691,
5
+ "train_runtime": 11716.0982,
6
+ "train_samples": 20000,
7
+ "train_samples_per_second": 3.414,
8
+ "train_steps_per_second": 0.107
9
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 100000,
4
+ "eos_token_id": 100001,
5
+ "transformers_version": "4.42.3"
6
+ }
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfa72038bb07d6e7b2db06365ddf749a1f0d2e7ead871483a2bcbdf7e6d2d45a
3
+ size 4987202208
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48aa886de79efc86115d136b985a023b9e80b8940d5aadb8da0faa2e767a26cc
3
+ size 4980945440
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f1204f6ccf8c8f580baba662b33eb215db8ea21273dab74aa457a0dd671d832
3
+ size 3852615520
model.safetensors.index.json ADDED
@@ -0,0 +1,280 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 13820731392
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "model-00003-of-00003.safetensors",
7
+ "model.embed_tokens.weight": "model-00001-of-00003.safetensors",
8
+ "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
9
+ "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
10
+ "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
11
+ "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
12
+ "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
13
+ "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
14
+ "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
15
+ "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
16
+ "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
17
+ "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
18
+ "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
19
+ "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
20
+ "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
21
+ "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
22
+ "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
23
+ "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
24
+ "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
25
+ "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
26
+ "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
27
+ "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
28
+ "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
29
+ "model.layers.10.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
30
+ "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
31
+ "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
32
+ "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
33
+ "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
34
+ "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
35
+ "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
36
+ "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
37
+ "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
38
+ "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
39
+ "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
40
+ "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
41
+ "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
42
+ "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
43
+ "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
44
+ "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
45
+ "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
46
+ "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
47
+ "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
48
+ "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
49
+ "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
50
+ "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
51
+ "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
52
+ "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
53
+ "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
54
+ "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
55
+ "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
56
+ "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
57
+ "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
58
+ "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
59
+ "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
60
+ "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
61
+ "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
62
+ "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
63
+ "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
64
+ "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
65
+ "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
66
+ "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
67
+ "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
68
+ "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
69
+ "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
70
+ "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
71
+ "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
72
+ "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
73
+ "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
74
+ "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
75
+ "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
76
+ "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
77
+ "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
78
+ "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
79
+ "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
80
+ "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
81
+ "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
82
+ "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
83
+ "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
84
+ "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
85
+ "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
86
+ "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
87
+ "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
88
+ "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
89
+ "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
90
+ "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
91
+ "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
92
+ "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
93
+ "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
94
+ "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
95
+ "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
96
+ "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
97
+ "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
98
+ "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
99
+ "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
100
+ "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
101
+ "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
102
+ "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
103
+ "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
104
+ "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
105
+ "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
106
+ "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
107
+ "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors",
108
+ "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
109
+ "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
110
+ "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
111
+ "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
112
+ "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
113
+ "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
114
+ "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
115
+ "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
116
+ "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
117
+ "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
118
+ "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
119
+ "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
120
+ "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
121
+ "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
122
+ "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
123
+ "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
124
+ "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
125
+ "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors",
126
+ "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
127
+ "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
128
+ "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
129
+ "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
130
+ "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
131
+ "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
132
+ "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
133
+ "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
134
+ "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors",
135
+ "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
136
+ "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
137
+ "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
138
+ "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
139
+ "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
140
+ "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
141
+ "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
142
+ "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
143
+ "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
144
+ "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
145
+ "model.layers.22.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
146
+ "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
147
+ "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
148
+ "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
149
+ "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
150
+ "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
151
+ "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
152
+ "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
153
+ "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
154
+ "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
155
+ "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
156
+ "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
157
+ "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
158
+ "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
159
+ "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
160
+ "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
161
+ "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
162
+ "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
163
+ "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
164
+ "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
165
+ "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
166
+ "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
167
+ "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
168
+ "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
169
+ "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
170
+ "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
171
+ "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
172
+ "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
173
+ "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
174
+ "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
175
+ "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
176
+ "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
177
+ "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
178
+ "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
179
+ "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
180
+ "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
181
+ "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
182
+ "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
183
+ "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
184
+ "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
185
+ "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
186
+ "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
187
+ "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
188
+ "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
189
+ "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
190
+ "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
191
+ "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
192
+ "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
193
+ "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
194
+ "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
195
+ "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
196
+ "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
197
+ "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
198
+ "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
199
+ "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
200
+ "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
201
+ "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
202
+ "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
203
+ "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
204
+ "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
205
+ "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
206
+ "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
207
+ "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
208
+ "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
209
+ "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
210
+ "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
211
+ "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
212
+ "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
213
+ "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
214
+ "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
215
+ "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
216
+ "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
217
+ "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
218
+ "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
219
+ "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
220
+ "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
221
+ "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
222
+ "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
223
+ "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
224
+ "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
225
+ "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
226
+ "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
227
+ "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
228
+ "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
229
+ "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
230
+ "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
231
+ "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
232
+ "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
233
+ "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
234
+ "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
235
+ "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
236
+ "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
237
+ "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
238
+ "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
239
+ "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
240
+ "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
241
+ "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
242
+ "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
243
+ "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
244
+ "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
245
+ "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
246
+ "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
247
+ "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
248
+ "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
249
+ "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
250
+ "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
251
+ "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
252
+ "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
253
+ "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
254
+ "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
255
+ "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
256
+ "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
257
+ "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
258
+ "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
259
+ "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
260
+ "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors",
261
+ "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
262
+ "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
263
+ "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
264
+ "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
265
+ "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
266
+ "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
267
+ "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
268
+ "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
269
+ "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors",
270
+ "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
271
+ "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
272
+ "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
273
+ "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
274
+ "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
275
+ "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
276
+ "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
277
+ "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
278
+ "model.norm.weight": "model-00003-of-00003.safetensors"
279
+ }
280
+ }
runs/Jul15_19-44-09_phe108-jieyuzhao-01/events.out.tfevents.1721098092.phe108-jieyuzhao-01.232196.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a87d5797b869c6dbb4faffaf0aff6b2c1ec244b01dec06128502e2c1a988f646
3
- size 112664
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c5c518b0a102820dbacdc15c5838d8eaf3571d7db3e200476d0e101c28377b3
3
+ size 113018
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 2.0,
3
+ "total_flos": 0.0,
4
+ "train_loss": 0.2130171504020691,
5
+ "train_runtime": 11716.0982,
6
+ "train_samples": 20000,
7
+ "train_samples_per_second": 3.414,
8
+ "train_steps_per_second": 0.107
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,2292 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.0,
5
+ "eval_steps": 10000,
6
+ "global_step": 1250,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.016,
13
+ "grad_norm": 7.157915223499978,
14
+ "learning_rate": 1.6000000000000001e-06,
15
+ "log_odds_chosen": 0.08687938749790192,
16
+ "log_odds_ratio": -0.6765035390853882,
17
+ "logits/chosen": 42.733116149902344,
18
+ "logits/rejected": 42.47905731201172,
19
+ "logps/chosen": -1.0179941654205322,
20
+ "logps/rejected": -1.0789813995361328,
21
+ "loss": 1.6603,
22
+ "nll_loss": 1.6577844619750977,
23
+ "rewards/accuracies": 0.574999988079071,
24
+ "rewards/chosen": -0.10179942846298218,
25
+ "rewards/margins": 0.0060987225733697414,
26
+ "rewards/rejected": -0.10789815336465836,
27
+ "step": 10
28
+ },
29
+ {
30
+ "epoch": 0.032,
31
+ "grad_norm": 4.787027456951044,
32
+ "learning_rate": 3.2000000000000003e-06,
33
+ "log_odds_chosen": 0.16144487261772156,
34
+ "log_odds_ratio": -0.6409393548965454,
35
+ "logits/chosen": 39.06427764892578,
36
+ "logits/rejected": 37.91705322265625,
37
+ "logps/chosen": -0.8086442947387695,
38
+ "logps/rejected": -0.9057012796401978,
39
+ "loss": 1.4104,
40
+ "nll_loss": 1.320190191268921,
41
+ "rewards/accuracies": 0.637499988079071,
42
+ "rewards/chosen": -0.08086444437503815,
43
+ "rewards/margins": 0.009705692529678345,
44
+ "rewards/rejected": -0.09057013690471649,
45
+ "step": 20
46
+ },
47
+ {
48
+ "epoch": 0.048,
49
+ "grad_norm": 3.619155990390428,
50
+ "learning_rate": 4.800000000000001e-06,
51
+ "log_odds_chosen": -0.09232249855995178,
52
+ "log_odds_ratio": -0.7732639312744141,
53
+ "logits/chosen": 33.01856231689453,
54
+ "logits/rejected": 34.8216552734375,
55
+ "logps/chosen": -0.5950765013694763,
56
+ "logps/rejected": -0.5527861714363098,
57
+ "loss": 0.9477,
58
+ "nll_loss": 0.9017837643623352,
59
+ "rewards/accuracies": 0.4375,
60
+ "rewards/chosen": -0.05950764939188957,
61
+ "rewards/margins": -0.004229038953781128,
62
+ "rewards/rejected": -0.05527861788868904,
63
+ "step": 30
64
+ },
65
+ {
66
+ "epoch": 0.064,
67
+ "grad_norm": 0.8193426378065174,
68
+ "learning_rate": 6.4000000000000006e-06,
69
+ "log_odds_chosen": 0.13158965110778809,
70
+ "log_odds_ratio": -0.6801391243934631,
71
+ "logits/chosen": 34.19690704345703,
72
+ "logits/rejected": 33.757015228271484,
73
+ "logps/chosen": -0.38923197984695435,
74
+ "logps/rejected": -0.42596691846847534,
75
+ "loss": 0.7871,
76
+ "nll_loss": 0.7480510473251343,
77
+ "rewards/accuracies": 0.48750001192092896,
78
+ "rewards/chosen": -0.038923196494579315,
79
+ "rewards/margins": 0.003673497587442398,
80
+ "rewards/rejected": -0.04259669408202171,
81
+ "step": 40
82
+ },
83
+ {
84
+ "epoch": 0.08,
85
+ "grad_norm": 0.837886863416422,
86
+ "learning_rate": 8.000000000000001e-06,
87
+ "log_odds_chosen": 0.010798773728311062,
88
+ "log_odds_ratio": -0.7395683526992798,
89
+ "logits/chosen": 36.046871185302734,
90
+ "logits/rejected": 35.39847946166992,
91
+ "logps/chosen": -0.3819189667701721,
92
+ "logps/rejected": -0.36964151263237,
93
+ "loss": 0.7614,
94
+ "nll_loss": 0.6979484558105469,
95
+ "rewards/accuracies": 0.5375000238418579,
96
+ "rewards/chosen": -0.038191892206668854,
97
+ "rewards/margins": -0.0012277422938495874,
98
+ "rewards/rejected": -0.03696415573358536,
99
+ "step": 50
100
+ },
101
+ {
102
+ "epoch": 0.096,
103
+ "grad_norm": 0.934326744395183,
104
+ "learning_rate": 9.600000000000001e-06,
105
+ "log_odds_chosen": 0.17773893475532532,
106
+ "log_odds_ratio": -0.6542357206344604,
107
+ "logits/chosen": 37.91405487060547,
108
+ "logits/rejected": 38.55655288696289,
109
+ "logps/chosen": -0.35373184084892273,
110
+ "logps/rejected": -0.40366658568382263,
111
+ "loss": 0.7369,
112
+ "nll_loss": 0.6853512525558472,
113
+ "rewards/accuracies": 0.699999988079071,
114
+ "rewards/chosen": -0.03537318482995033,
115
+ "rewards/margins": 0.0049934713169932365,
116
+ "rewards/rejected": -0.040366653352975845,
117
+ "step": 60
118
+ },
119
+ {
120
+ "epoch": 0.112,
121
+ "grad_norm": 0.8140121698145116,
122
+ "learning_rate": 1.1200000000000001e-05,
123
+ "log_odds_chosen": 0.27008742094039917,
124
+ "log_odds_ratio": -0.6275990009307861,
125
+ "logits/chosen": 37.89257049560547,
126
+ "logits/rejected": 36.801631927490234,
127
+ "logps/chosen": -0.3447234630584717,
128
+ "logps/rejected": -0.44380050897598267,
129
+ "loss": 0.7405,
130
+ "nll_loss": 0.6623181104660034,
131
+ "rewards/accuracies": 0.6499999761581421,
132
+ "rewards/chosen": -0.03447234630584717,
133
+ "rewards/margins": 0.009907705709338188,
134
+ "rewards/rejected": -0.04438004642724991,
135
+ "step": 70
136
+ },
137
+ {
138
+ "epoch": 0.128,
139
+ "grad_norm": 0.8866561757309486,
140
+ "learning_rate": 1.2800000000000001e-05,
141
+ "log_odds_chosen": 0.7000471949577332,
142
+ "log_odds_ratio": -0.46674174070358276,
143
+ "logits/chosen": 39.06570816040039,
144
+ "logits/rejected": 38.14867401123047,
145
+ "logps/chosen": -0.3046243488788605,
146
+ "logps/rejected": -0.541035532951355,
147
+ "loss": 0.6756,
148
+ "nll_loss": 0.5991424322128296,
149
+ "rewards/accuracies": 0.8374999761581421,
150
+ "rewards/chosen": -0.030462438240647316,
151
+ "rewards/margins": 0.023641113191843033,
152
+ "rewards/rejected": -0.0541035458445549,
153
+ "step": 80
154
+ },
155
+ {
156
+ "epoch": 0.144,
157
+ "grad_norm": 1.1050711696617728,
158
+ "learning_rate": 1.4400000000000001e-05,
159
+ "log_odds_chosen": 0.4890363812446594,
160
+ "log_odds_ratio": -0.5526093244552612,
161
+ "logits/chosen": 39.77109146118164,
162
+ "logits/rejected": 41.683753967285156,
163
+ "logps/chosen": -0.36524319648742676,
164
+ "logps/rejected": -0.5409058928489685,
165
+ "loss": 0.6761,
166
+ "nll_loss": 0.6159157156944275,
167
+ "rewards/accuracies": 0.7250000238418579,
168
+ "rewards/chosen": -0.036524318158626556,
169
+ "rewards/margins": 0.017566271126270294,
170
+ "rewards/rejected": -0.05409059673547745,
171
+ "step": 90
172
+ },
173
+ {
174
+ "epoch": 0.16,
175
+ "grad_norm": 1.1283739516296225,
176
+ "learning_rate": 1.6000000000000003e-05,
177
+ "log_odds_chosen": 0.5452580451965332,
178
+ "log_odds_ratio": -0.5334872603416443,
179
+ "logits/chosen": 41.09711837768555,
180
+ "logits/rejected": 39.49372100830078,
181
+ "logps/chosen": -0.3422545790672302,
182
+ "logps/rejected": -0.5715359449386597,
183
+ "loss": 0.6441,
184
+ "nll_loss": 0.5890800952911377,
185
+ "rewards/accuracies": 0.762499988079071,
186
+ "rewards/chosen": -0.0342254601418972,
187
+ "rewards/margins": 0.022928137332201004,
188
+ "rewards/rejected": -0.05715359374880791,
189
+ "step": 100
190
+ },
191
+ {
192
+ "epoch": 0.176,
193
+ "grad_norm": 1.044328943254365,
194
+ "learning_rate": 1.76e-05,
195
+ "log_odds_chosen": 0.6057835817337036,
196
+ "log_odds_ratio": -0.5027212500572205,
197
+ "logits/chosen": 41.39733123779297,
198
+ "logits/rejected": 39.85953140258789,
199
+ "logps/chosen": -0.35752102732658386,
200
+ "logps/rejected": -0.590412437915802,
201
+ "loss": 0.6068,
202
+ "nll_loss": 0.5553559064865112,
203
+ "rewards/accuracies": 0.800000011920929,
204
+ "rewards/chosen": -0.035752106457948685,
205
+ "rewards/margins": 0.023289144039154053,
206
+ "rewards/rejected": -0.059041254222393036,
207
+ "step": 110
208
+ },
209
+ {
210
+ "epoch": 0.192,
211
+ "grad_norm": 1.0225837980847625,
212
+ "learning_rate": 1.9200000000000003e-05,
213
+ "log_odds_chosen": 0.5523650050163269,
214
+ "log_odds_ratio": -0.5369630455970764,
215
+ "logits/chosen": 43.63589859008789,
216
+ "logits/rejected": 41.659324645996094,
217
+ "logps/chosen": -0.3336974084377289,
218
+ "logps/rejected": -0.5324398279190063,
219
+ "loss": 0.5734,
220
+ "nll_loss": 0.5098426938056946,
221
+ "rewards/accuracies": 0.75,
222
+ "rewards/chosen": -0.03336974233388901,
223
+ "rewards/margins": 0.01987423747777939,
224
+ "rewards/rejected": -0.05324398726224899,
225
+ "step": 120
226
+ },
227
+ {
228
+ "epoch": 0.208,
229
+ "grad_norm": 1.1101012093036131,
230
+ "learning_rate": 1.9999025240093045e-05,
231
+ "log_odds_chosen": 0.7986687421798706,
232
+ "log_odds_ratio": -0.471863329410553,
233
+ "logits/chosen": 43.426692962646484,
234
+ "logits/rejected": 40.31127166748047,
235
+ "logps/chosen": -0.35013002157211304,
236
+ "logps/rejected": -0.6971026659011841,
237
+ "loss": 0.5401,
238
+ "nll_loss": 0.508377730846405,
239
+ "rewards/accuracies": 0.824999988079071,
240
+ "rewards/chosen": -0.03501300513744354,
241
+ "rewards/margins": 0.03469725698232651,
242
+ "rewards/rejected": -0.06971026211977005,
243
+ "step": 130
244
+ },
245
+ {
246
+ "epoch": 0.224,
247
+ "grad_norm": 1.5810963333813628,
248
+ "learning_rate": 1.9991228300988586e-05,
249
+ "log_odds_chosen": 0.9270819425582886,
250
+ "log_odds_ratio": -0.4264460504055023,
251
+ "logits/chosen": 40.22356414794922,
252
+ "logits/rejected": 38.093299865722656,
253
+ "logps/chosen": -0.3407229781150818,
254
+ "logps/rejected": -0.7779335975646973,
255
+ "loss": 0.5243,
256
+ "nll_loss": 0.4749884605407715,
257
+ "rewards/accuracies": 0.8125,
258
+ "rewards/chosen": -0.03407229855656624,
259
+ "rewards/margins": 0.04372106492519379,
260
+ "rewards/rejected": -0.07779336720705032,
261
+ "step": 140
262
+ },
263
+ {
264
+ "epoch": 0.24,
265
+ "grad_norm": 0.9914026891453022,
266
+ "learning_rate": 1.9975640502598243e-05,
267
+ "log_odds_chosen": 0.7473500967025757,
268
+ "log_odds_ratio": -0.4457695484161377,
269
+ "logits/chosen": 38.593631744384766,
270
+ "logits/rejected": 40.06364059448242,
271
+ "logps/chosen": -0.35150039196014404,
272
+ "logps/rejected": -0.6378719210624695,
273
+ "loss": 0.507,
274
+ "nll_loss": 0.45233845710754395,
275
+ "rewards/accuracies": 0.875,
276
+ "rewards/chosen": -0.035150036215782166,
277
+ "rewards/margins": 0.02863715961575508,
278
+ "rewards/rejected": -0.06378719955682755,
279
+ "step": 150
280
+ },
281
+ {
282
+ "epoch": 0.256,
283
+ "grad_norm": 1.0114005741143277,
284
+ "learning_rate": 1.9952273999818312e-05,
285
+ "log_odds_chosen": 0.7789567708969116,
286
+ "log_odds_ratio": -0.47668343782424927,
287
+ "logits/chosen": 40.021446228027344,
288
+ "logits/rejected": 39.03896713256836,
289
+ "logps/chosen": -0.32147353887557983,
290
+ "logps/rejected": -0.6195149421691895,
291
+ "loss": 0.4762,
292
+ "nll_loss": 0.40015798807144165,
293
+ "rewards/accuracies": 0.7749999761581421,
294
+ "rewards/chosen": -0.0321473553776741,
295
+ "rewards/margins": 0.02980414591729641,
296
+ "rewards/rejected": -0.061951495707035065,
297
+ "step": 160
298
+ },
299
+ {
300
+ "epoch": 0.272,
301
+ "grad_norm": 1.1072013212789444,
302
+ "learning_rate": 1.9921147013144782e-05,
303
+ "log_odds_chosen": 0.9789760708808899,
304
+ "log_odds_ratio": -0.3996427059173584,
305
+ "logits/chosen": 38.469261169433594,
306
+ "logits/rejected": 39.36458206176758,
307
+ "logps/chosen": -0.31273895502090454,
308
+ "logps/rejected": -0.7282775640487671,
309
+ "loss": 0.4569,
310
+ "nll_loss": 0.4199884533882141,
311
+ "rewards/accuracies": 0.887499988079071,
312
+ "rewards/chosen": -0.031273894011974335,
313
+ "rewards/margins": 0.041553862392902374,
314
+ "rewards/rejected": -0.07282774895429611,
315
+ "step": 170
316
+ },
317
+ {
318
+ "epoch": 0.288,
319
+ "grad_norm": 1.0540475357463217,
320
+ "learning_rate": 1.988228381446553e-05,
321
+ "log_odds_chosen": 0.9965242147445679,
322
+ "log_odds_ratio": -0.4183652997016907,
323
+ "logits/chosen": 40.18573760986328,
324
+ "logits/rejected": 40.521148681640625,
325
+ "logps/chosen": -0.33785638213157654,
326
+ "logps/rejected": -0.7628054022789001,
327
+ "loss": 0.4311,
328
+ "nll_loss": 0.43037042021751404,
329
+ "rewards/accuracies": 0.824999988079071,
330
+ "rewards/chosen": -0.03378564119338989,
331
+ "rewards/margins": 0.04249490052461624,
332
+ "rewards/rejected": -0.07628054916858673,
333
+ "step": 180
334
+ },
335
+ {
336
+ "epoch": 0.304,
337
+ "grad_norm": 1.2807805662820098,
338
+ "learning_rate": 1.983571470813386e-05,
339
+ "log_odds_chosen": 0.9251772165298462,
340
+ "log_odds_ratio": -0.43279653787612915,
341
+ "logits/chosen": 39.66691970825195,
342
+ "logits/rejected": 39.22489929199219,
343
+ "logps/chosen": -0.3330400586128235,
344
+ "logps/rejected": -0.6816459894180298,
345
+ "loss": 0.4062,
346
+ "nll_loss": 0.36826083064079285,
347
+ "rewards/accuracies": 0.8125,
348
+ "rewards/chosen": -0.03330400586128235,
349
+ "rewards/margins": 0.03486059606075287,
350
+ "rewards/rejected": -0.06816460937261581,
351
+ "step": 190
352
+ },
353
+ {
354
+ "epoch": 0.32,
355
+ "grad_norm": 1.247095232534241,
356
+ "learning_rate": 1.9781476007338058e-05,
357
+ "log_odds_chosen": 1.1870229244232178,
358
+ "log_odds_ratio": -0.331691175699234,
359
+ "logits/chosen": 36.31359100341797,
360
+ "logits/rejected": 38.259849548339844,
361
+ "logps/chosen": -0.2683125436306,
362
+ "logps/rejected": -0.6734791994094849,
363
+ "loss": 0.3637,
364
+ "nll_loss": 0.29906129837036133,
365
+ "rewards/accuracies": 0.949999988079071,
366
+ "rewards/chosen": -0.026831254363059998,
367
+ "rewards/margins": 0.04051666334271431,
368
+ "rewards/rejected": -0.0673479214310646,
369
+ "step": 200
370
+ },
371
+ {
372
+ "epoch": 0.336,
373
+ "grad_norm": 1.023611287110811,
374
+ "learning_rate": 1.9719610005785466e-05,
375
+ "log_odds_chosen": 1.10099196434021,
376
+ "log_odds_ratio": -0.3911464810371399,
377
+ "logits/chosen": 36.06108856201172,
378
+ "logits/rejected": 36.09987258911133,
379
+ "logps/chosen": -0.3143238425254822,
380
+ "logps/rejected": -0.7127642035484314,
381
+ "loss": 0.3608,
382
+ "nll_loss": 0.33582496643066406,
383
+ "rewards/accuracies": 0.862500011920929,
384
+ "rewards/chosen": -0.0314323827624321,
385
+ "rewards/margins": 0.03984404355287552,
386
+ "rewards/rejected": -0.07127643376588821,
387
+ "step": 210
388
+ },
389
+ {
390
+ "epoch": 0.352,
391
+ "grad_norm": 1.0618187057425008,
392
+ "learning_rate": 1.9650164944723116e-05,
393
+ "log_odds_chosen": 1.4383821487426758,
394
+ "log_odds_ratio": -0.313861608505249,
395
+ "logits/chosen": 34.82101821899414,
396
+ "logits/rejected": 38.44565200805664,
397
+ "logps/chosen": -0.28008711338043213,
398
+ "logps/rejected": -0.8806719779968262,
399
+ "loss": 0.3331,
400
+ "nll_loss": 0.2972494959831238,
401
+ "rewards/accuracies": 0.9375,
402
+ "rewards/chosen": -0.028008710592985153,
403
+ "rewards/margins": 0.06005847454071045,
404
+ "rewards/rejected": -0.0880671888589859,
405
+ "step": 220
406
+ },
407
+ {
408
+ "epoch": 0.368,
409
+ "grad_norm": 0.9398000108663768,
410
+ "learning_rate": 1.9573194975320672e-05,
411
+ "log_odds_chosen": 1.4377901554107666,
412
+ "log_odds_ratio": -0.34182658791542053,
413
+ "logits/chosen": 33.84839630126953,
414
+ "logits/rejected": 35.625667572021484,
415
+ "logps/chosen": -0.30297523736953735,
416
+ "logps/rejected": -0.9298135042190552,
417
+ "loss": 0.3403,
418
+ "nll_loss": 0.2947668433189392,
419
+ "rewards/accuracies": 0.8999999761581421,
420
+ "rewards/chosen": -0.030297517776489258,
421
+ "rewards/margins": 0.0626838356256485,
422
+ "rewards/rejected": -0.09298135340213776,
423
+ "step": 230
424
+ },
425
+ {
426
+ "epoch": 0.384,
427
+ "grad_norm": 1.103321212338863,
428
+ "learning_rate": 1.9488760116444966e-05,
429
+ "log_odds_chosen": 1.4992961883544922,
430
+ "log_odds_ratio": -0.3107326030731201,
431
+ "logits/chosen": 35.9039421081543,
432
+ "logits/rejected": 36.633705139160156,
433
+ "logps/chosen": -0.2847585380077362,
434
+ "logps/rejected": -0.9570012092590332,
435
+ "loss": 0.3166,
436
+ "nll_loss": 0.30076318979263306,
437
+ "rewards/accuracies": 0.925000011920929,
438
+ "rewards/chosen": -0.02847585640847683,
439
+ "rewards/margins": 0.06722426414489746,
440
+ "rewards/rejected": -0.09570012241601944,
441
+ "step": 240
442
+ },
443
+ {
444
+ "epoch": 0.4,
445
+ "grad_norm": 0.9188155900142966,
446
+ "learning_rate": 1.9396926207859085e-05,
447
+ "log_odds_chosen": 1.232299566268921,
448
+ "log_odds_ratio": -0.36939191818237305,
449
+ "logits/chosen": 36.425289154052734,
450
+ "logits/rejected": 38.76382827758789,
451
+ "logps/chosen": -0.306942343711853,
452
+ "logps/rejected": -0.8188571929931641,
453
+ "loss": 0.3212,
454
+ "nll_loss": 0.2955878674983978,
455
+ "rewards/accuracies": 0.824999988079071,
456
+ "rewards/chosen": -0.030694235116243362,
457
+ "rewards/margins": 0.05119148641824722,
458
+ "rewards/rejected": -0.08188571780920029,
459
+ "step": 250
460
+ },
461
+ {
462
+ "epoch": 0.416,
463
+ "grad_norm": 1.1371525474896846,
464
+ "learning_rate": 1.9297764858882516e-05,
465
+ "log_odds_chosen": 1.5188441276550293,
466
+ "log_odds_ratio": -0.3032824695110321,
467
+ "logits/chosen": 33.87938690185547,
468
+ "logits/rejected": 34.950599670410156,
469
+ "logps/chosen": -0.2652571499347687,
470
+ "logps/rejected": -0.904030442237854,
471
+ "loss": 0.282,
472
+ "nll_loss": 0.2362031489610672,
473
+ "rewards/accuracies": 0.9125000238418579,
474
+ "rewards/chosen": -0.026525715366005898,
475
+ "rewards/margins": 0.06387733668088913,
476
+ "rewards/rejected": -0.09040305018424988,
477
+ "step": 260
478
+ },
479
+ {
480
+ "epoch": 0.432,
481
+ "grad_norm": 0.9722134512964291,
482
+ "learning_rate": 1.9191353392552346e-05,
483
+ "log_odds_chosen": 1.6056066751480103,
484
+ "log_odds_ratio": -0.273274302482605,
485
+ "logits/chosen": 31.2882080078125,
486
+ "logits/rejected": 32.222434997558594,
487
+ "logps/chosen": -0.2462386190891266,
488
+ "logps/rejected": -0.843704342842102,
489
+ "loss": 0.2621,
490
+ "nll_loss": 0.23743471503257751,
491
+ "rewards/accuracies": 0.949999988079071,
492
+ "rewards/chosen": -0.02462385967373848,
493
+ "rewards/margins": 0.059746574610471725,
494
+ "rewards/rejected": -0.0843704342842102,
495
+ "step": 270
496
+ },
497
+ {
498
+ "epoch": 0.448,
499
+ "grad_norm": 0.9536358910314567,
500
+ "learning_rate": 1.907777478532909e-05,
501
+ "log_odds_chosen": 1.3453700542449951,
502
+ "log_odds_ratio": -0.3348150849342346,
503
+ "logits/chosen": 34.44597625732422,
504
+ "logits/rejected": 35.61604690551758,
505
+ "logps/chosen": -0.25559094548225403,
506
+ "logps/rejected": -0.7815641164779663,
507
+ "loss": 0.2815,
508
+ "nll_loss": 0.25891661643981934,
509
+ "rewards/accuracies": 0.9125000238418579,
510
+ "rewards/chosen": -0.025559093803167343,
511
+ "rewards/margins": 0.052597321569919586,
512
+ "rewards/rejected": -0.07815642654895782,
513
+ "step": 280
514
+ },
515
+ {
516
+ "epoch": 0.464,
517
+ "grad_norm": 1.4793964733191685,
518
+ "learning_rate": 1.895711760239413e-05,
519
+ "log_odds_chosen": 1.8677561283111572,
520
+ "log_odds_ratio": -0.2598610520362854,
521
+ "logits/chosen": 30.939416885375977,
522
+ "logits/rejected": 32.69816589355469,
523
+ "logps/chosen": -0.25819864869117737,
524
+ "logps/rejected": -1.080700159072876,
525
+ "loss": 0.2526,
526
+ "nll_loss": 0.24270136654376984,
527
+ "rewards/accuracies": 0.9375,
528
+ "rewards/chosen": -0.025819867849349976,
529
+ "rewards/margins": 0.08225013315677643,
530
+ "rewards/rejected": -0.1080700010061264,
531
+ "step": 290
532
+ },
533
+ {
534
+ "epoch": 0.48,
535
+ "grad_norm": 1.2223140757678745,
536
+ "learning_rate": 1.8829475928589272e-05,
537
+ "log_odds_chosen": 1.873132348060608,
538
+ "log_odds_ratio": -0.21418969333171844,
539
+ "logits/chosen": 30.050670623779297,
540
+ "logits/rejected": 31.51558494567871,
541
+ "logps/chosen": -0.24227066338062286,
542
+ "logps/rejected": -1.0071208477020264,
543
+ "loss": 0.2446,
544
+ "nll_loss": 0.20659121870994568,
545
+ "rewards/accuracies": 0.9750000238418579,
546
+ "rewards/chosen": -0.024227067828178406,
547
+ "rewards/margins": 0.07648502290248871,
548
+ "rewards/rejected": -0.10071208328008652,
549
+ "step": 300
550
+ },
551
+ {
552
+ "epoch": 0.496,
553
+ "grad_norm": 0.8813575124105774,
554
+ "learning_rate": 1.869494929505219e-05,
555
+ "log_odds_chosen": 1.9893770217895508,
556
+ "log_odds_ratio": -0.22330129146575928,
557
+ "logits/chosen": 33.173133850097656,
558
+ "logits/rejected": 34.073055267333984,
559
+ "logps/chosen": -0.2551695704460144,
560
+ "logps/rejected": -1.1849584579467773,
561
+ "loss": 0.232,
562
+ "nll_loss": 0.2051832675933838,
563
+ "rewards/accuracies": 0.987500011920929,
564
+ "rewards/chosen": -0.025516962632536888,
565
+ "rewards/margins": 0.09297889471054077,
566
+ "rewards/rejected": -0.11849584430456161,
567
+ "step": 310
568
+ },
569
+ {
570
+ "epoch": 0.512,
571
+ "grad_norm": 0.9094503804430618,
572
+ "learning_rate": 1.855364260160507e-05,
573
+ "log_odds_chosen": 1.7108672857284546,
574
+ "log_odds_ratio": -0.2914069592952728,
575
+ "logits/chosen": 31.070571899414062,
576
+ "logits/rejected": 31.337520599365234,
577
+ "logps/chosen": -0.26724526286125183,
578
+ "logps/rejected": -0.9716242551803589,
579
+ "loss": 0.2355,
580
+ "nll_loss": 0.2270585000514984,
581
+ "rewards/accuracies": 0.8999999761581421,
582
+ "rewards/chosen": -0.026724526658654213,
583
+ "rewards/margins": 0.07043790072202682,
584
+ "rewards/rejected": -0.09716243296861649,
585
+ "step": 320
586
+ },
587
+ {
588
+ "epoch": 0.528,
589
+ "grad_norm": 0.9058310290840669,
590
+ "learning_rate": 1.8405666034956842e-05,
591
+ "log_odds_chosen": 2.1886684894561768,
592
+ "log_odds_ratio": -0.19184692203998566,
593
+ "logits/chosen": 30.936025619506836,
594
+ "logits/rejected": 31.9810848236084,
595
+ "logps/chosen": -0.2256220132112503,
596
+ "logps/rejected": -1.2340004444122314,
597
+ "loss": 0.2216,
598
+ "nll_loss": 0.18992918729782104,
599
+ "rewards/accuracies": 0.9750000238418579,
600
+ "rewards/chosen": -0.02256220206618309,
601
+ "rewards/margins": 0.100837841629982,
602
+ "rewards/rejected": -0.12340004742145538,
603
+ "step": 330
604
+ },
605
+ {
606
+ "epoch": 0.544,
607
+ "grad_norm": 0.7923767872883756,
608
+ "learning_rate": 1.8251134982782952e-05,
609
+ "log_odds_chosen": 1.8938252925872803,
610
+ "log_odds_ratio": -0.23439836502075195,
611
+ "logits/chosen": 30.1374454498291,
612
+ "logits/rejected": 31.97110939025879,
613
+ "logps/chosen": -0.25573670864105225,
614
+ "logps/rejected": -1.058814525604248,
615
+ "loss": 0.2256,
616
+ "nll_loss": 0.2036077082157135,
617
+ "rewards/accuracies": 0.9624999761581421,
618
+ "rewards/chosen": -0.025573670864105225,
619
+ "rewards/margins": 0.08030776679515839,
620
+ "rewards/rejected": -0.1058814525604248,
621
+ "step": 340
622
+ },
623
+ {
624
+ "epoch": 0.56,
625
+ "grad_norm": 0.941711390597777,
626
+ "learning_rate": 1.8090169943749477e-05,
627
+ "log_odds_chosen": 1.890398621559143,
628
+ "log_odds_ratio": -0.23317229747772217,
629
+ "logits/chosen": 29.39468002319336,
630
+ "logits/rejected": 31.831829071044922,
631
+ "logps/chosen": -0.26384299993515015,
632
+ "logps/rejected": -1.0387157201766968,
633
+ "loss": 0.2102,
634
+ "nll_loss": 0.19437000155448914,
635
+ "rewards/accuracies": 0.949999988079071,
636
+ "rewards/chosen": -0.026384299620985985,
637
+ "rewards/margins": 0.0774872750043869,
638
+ "rewards/rejected": -0.10387158393859863,
639
+ "step": 350
640
+ },
641
+ {
642
+ "epoch": 0.576,
643
+ "grad_norm": 0.9511427594232491,
644
+ "learning_rate": 1.792289643355191e-05,
645
+ "log_odds_chosen": 2.1849417686462402,
646
+ "log_odds_ratio": -0.20272669196128845,
647
+ "logits/chosen": 28.536510467529297,
648
+ "logits/rejected": 31.184783935546875,
649
+ "logps/chosen": -0.21713824570178986,
650
+ "logps/rejected": -1.1126606464385986,
651
+ "loss": 0.196,
652
+ "nll_loss": 0.1743771880865097,
653
+ "rewards/accuracies": 0.9750000238418579,
654
+ "rewards/chosen": -0.021713824942708015,
655
+ "rewards/margins": 0.08955223858356476,
656
+ "rewards/rejected": -0.11126606166362762,
657
+ "step": 360
658
+ },
659
+ {
660
+ "epoch": 0.592,
661
+ "grad_norm": 1.3234030569644144,
662
+ "learning_rate": 1.7749444887041797e-05,
663
+ "log_odds_chosen": 2.2127633094787598,
664
+ "log_odds_ratio": -0.1819957047700882,
665
+ "logits/chosen": 28.459209442138672,
666
+ "logits/rejected": 30.699054718017578,
667
+ "logps/chosen": -0.19446463882923126,
668
+ "logps/rejected": -1.064205527305603,
669
+ "loss": 0.2095,
670
+ "nll_loss": 0.157680481672287,
671
+ "rewards/accuracies": 1.0,
672
+ "rewards/chosen": -0.019446464255452156,
673
+ "rewards/margins": 0.08697408437728882,
674
+ "rewards/rejected": -0.10642055422067642,
675
+ "step": 370
676
+ },
677
+ {
678
+ "epoch": 0.608,
679
+ "grad_norm": 0.7887424541285082,
680
+ "learning_rate": 1.7569950556517566e-05,
681
+ "log_odds_chosen": 2.0950214862823486,
682
+ "log_odds_ratio": -0.2056867778301239,
683
+ "logits/chosen": 26.343616485595703,
684
+ "logits/rejected": 30.175817489624023,
685
+ "logps/chosen": -0.22634263336658478,
686
+ "logps/rejected": -1.0738613605499268,
687
+ "loss": 0.1995,
688
+ "nll_loss": 0.1639939546585083,
689
+ "rewards/accuracies": 0.9750000238418579,
690
+ "rewards/chosen": -0.022634264081716537,
691
+ "rewards/margins": 0.08475185930728912,
692
+ "rewards/rejected": -0.10738612711429596,
693
+ "step": 380
694
+ },
695
+ {
696
+ "epoch": 0.624,
697
+ "grad_norm": 0.8925541324230298,
698
+ "learning_rate": 1.7384553406258842e-05,
699
+ "log_odds_chosen": 2.156597852706909,
700
+ "log_odds_ratio": -0.1930965632200241,
701
+ "logits/chosen": 26.64847755432129,
702
+ "logits/rejected": 28.833852767944336,
703
+ "logps/chosen": -0.2270650863647461,
704
+ "logps/rejected": -1.0798944234848022,
705
+ "loss": 0.2017,
706
+ "nll_loss": 0.1784597635269165,
707
+ "rewards/accuracies": 0.987500011920929,
708
+ "rewards/chosen": -0.022706512361764908,
709
+ "rewards/margins": 0.08528293669223785,
710
+ "rewards/rejected": -0.10798945277929306,
711
+ "step": 390
712
+ },
713
+ {
714
+ "epoch": 0.64,
715
+ "grad_norm": 1.0272557886917288,
716
+ "learning_rate": 1.7193398003386514e-05,
717
+ "log_odds_chosen": 2.46358323097229,
718
+ "log_odds_ratio": -0.17916469275951385,
719
+ "logits/chosen": 28.102930068969727,
720
+ "logits/rejected": 29.97959327697754,
721
+ "logps/chosen": -0.21100203692913055,
722
+ "logps/rejected": -1.2793248891830444,
723
+ "loss": 0.1938,
724
+ "nll_loss": 0.1722893863916397,
725
+ "rewards/accuracies": 0.9750000238418579,
726
+ "rewards/chosen": -0.021100204437971115,
727
+ "rewards/margins": 0.10683228820562363,
728
+ "rewards/rejected": -0.12793250381946564,
729
+ "step": 400
730
+ },
731
+ {
732
+ "epoch": 0.656,
733
+ "grad_norm": 0.7328015065358303,
734
+ "learning_rate": 1.6996633405133656e-05,
735
+ "log_odds_chosen": 2.6361804008483887,
736
+ "log_odds_ratio": -0.14345327019691467,
737
+ "logits/chosen": 27.147369384765625,
738
+ "logits/rejected": 29.60892105102539,
739
+ "logps/chosen": -0.18612393736839294,
740
+ "logps/rejected": -1.346756935119629,
741
+ "loss": 0.1836,
742
+ "nll_loss": 0.14793823659420013,
743
+ "rewards/accuracies": 1.0,
744
+ "rewards/chosen": -0.018612395972013474,
745
+ "rewards/margins": 0.11606329679489136,
746
+ "rewards/rejected": -0.13467569649219513,
747
+ "step": 410
748
+ },
749
+ {
750
+ "epoch": 0.672,
751
+ "grad_norm": 0.7570321808098591,
752
+ "learning_rate": 1.6794413042615168e-05,
753
+ "log_odds_chosen": 2.324608325958252,
754
+ "log_odds_ratio": -0.16639328002929688,
755
+ "logits/chosen": 26.477365493774414,
756
+ "logits/rejected": 29.83447265625,
757
+ "logps/chosen": -0.19487398862838745,
758
+ "logps/rejected": -1.086843729019165,
759
+ "loss": 0.1708,
760
+ "nll_loss": 0.15096671879291534,
761
+ "rewards/accuracies": 1.0,
762
+ "rewards/chosen": -0.019487399607896805,
763
+ "rewards/margins": 0.08919696509838104,
764
+ "rewards/rejected": -0.10868436098098755,
765
+ "step": 420
766
+ },
767
+ {
768
+ "epoch": 0.688,
769
+ "grad_norm": 0.8589237619818589,
770
+ "learning_rate": 1.6586894601186804e-05,
771
+ "log_odds_chosen": 2.4892306327819824,
772
+ "log_odds_ratio": -0.1399163007736206,
773
+ "logits/chosen": 26.335521697998047,
774
+ "logits/rejected": 28.351062774658203,
775
+ "logps/chosen": -0.185403510928154,
776
+ "logps/rejected": -1.1550278663635254,
777
+ "loss": 0.1719,
778
+ "nll_loss": 0.15061026811599731,
779
+ "rewards/accuracies": 0.987500011920929,
780
+ "rewards/chosen": -0.01854034885764122,
781
+ "rewards/margins": 0.09696242958307266,
782
+ "rewards/rejected": -0.11550277471542358,
783
+ "step": 430
784
+ },
785
+ {
786
+ "epoch": 0.704,
787
+ "grad_norm": 0.9410946432421042,
788
+ "learning_rate": 1.63742398974869e-05,
789
+ "log_odds_chosen": 2.6492185592651367,
790
+ "log_odds_ratio": -0.11772862821817398,
791
+ "logits/chosen": 25.387847900390625,
792
+ "logits/rejected": 27.318252563476562,
793
+ "logps/chosen": -0.18541622161865234,
794
+ "logps/rejected": -1.3411946296691895,
795
+ "loss": 0.168,
796
+ "nll_loss": 0.1478230208158493,
797
+ "rewards/accuracies": 1.0,
798
+ "rewards/chosen": -0.018541622906923294,
799
+ "rewards/margins": 0.11557783931493759,
800
+ "rewards/rejected": -0.13411946594715118,
801
+ "step": 440
802
+ },
803
+ {
804
+ "epoch": 0.72,
805
+ "grad_norm": 0.913121815013067,
806
+ "learning_rate": 1.6156614753256583e-05,
807
+ "log_odds_chosen": 2.5952377319335938,
808
+ "log_odds_ratio": -0.13427497446537018,
809
+ "logits/chosen": 25.367713928222656,
810
+ "logits/rejected": 26.90452003479004,
811
+ "logps/chosen": -0.1814824938774109,
812
+ "logps/rejected": -1.2440539598464966,
813
+ "loss": 0.1621,
814
+ "nll_loss": 0.13832086324691772,
815
+ "rewards/accuracies": 0.987500011920929,
816
+ "rewards/chosen": -0.018148250877857208,
817
+ "rewards/margins": 0.10625714063644409,
818
+ "rewards/rejected": -0.1244053840637207,
819
+ "step": 450
820
+ },
821
+ {
822
+ "epoch": 0.736,
823
+ "grad_norm": 0.8697507877833889,
824
+ "learning_rate": 1.5934188866037017e-05,
825
+ "log_odds_chosen": 2.944627046585083,
826
+ "log_odds_ratio": -0.11624413728713989,
827
+ "logits/chosen": 25.267337799072266,
828
+ "logits/rejected": 29.403518676757812,
829
+ "logps/chosen": -0.17217525839805603,
830
+ "logps/rejected": -1.3934593200683594,
831
+ "loss": 0.1569,
832
+ "nll_loss": 0.14144101738929749,
833
+ "rewards/accuracies": 0.987500011920929,
834
+ "rewards/chosen": -0.017217526212334633,
835
+ "rewards/margins": 0.12212841212749481,
836
+ "rewards/rejected": -0.1393459290266037,
837
+ "step": 460
838
+ },
839
+ {
840
+ "epoch": 0.752,
841
+ "grad_norm": 0.6621793550764674,
842
+ "learning_rate": 1.570713567684432e-05,
843
+ "log_odds_chosen": 2.606139659881592,
844
+ "log_odds_ratio": -0.1426299810409546,
845
+ "logits/chosen": 29.288564682006836,
846
+ "logits/rejected": 29.074045181274414,
847
+ "logps/chosen": -0.19370333850383759,
848
+ "logps/rejected": -1.2478258609771729,
849
+ "loss": 0.1792,
850
+ "nll_loss": 0.16615137457847595,
851
+ "rewards/accuracies": 0.987500011920929,
852
+ "rewards/chosen": -0.019370336085557938,
853
+ "rewards/margins": 0.10541225969791412,
854
+ "rewards/rejected": -0.12478259950876236,
855
+ "step": 470
856
+ },
857
+ {
858
+ "epoch": 0.768,
859
+ "grad_norm": 0.6543673919224716,
860
+ "learning_rate": 1.5475632234925505e-05,
861
+ "log_odds_chosen": 2.9211063385009766,
862
+ "log_odds_ratio": -0.105687215924263,
863
+ "logits/chosen": 24.23305892944336,
864
+ "logits/rejected": 27.091516494750977,
865
+ "logps/chosen": -0.1637393683195114,
866
+ "logps/rejected": -1.3047716617584229,
867
+ "loss": 0.1509,
868
+ "nll_loss": 0.12331098318099976,
869
+ "rewards/accuracies": 1.0,
870
+ "rewards/chosen": -0.01637393794953823,
871
+ "rewards/margins": 0.11410322040319443,
872
+ "rewards/rejected": -0.1304771602153778,
873
+ "step": 480
874
+ },
875
+ {
876
+ "epoch": 0.784,
877
+ "grad_norm": 0.741926124626365,
878
+ "learning_rate": 1.5239859059700794e-05,
879
+ "log_odds_chosen": 2.723001003265381,
880
+ "log_odds_ratio": -0.13461025059223175,
881
+ "logits/chosen": 25.58415412902832,
882
+ "logits/rejected": 27.281457901000977,
883
+ "logps/chosen": -0.17167863249778748,
884
+ "logps/rejected": -1.1783361434936523,
885
+ "loss": 0.149,
886
+ "nll_loss": 0.1352442055940628,
887
+ "rewards/accuracies": 0.9750000238418579,
888
+ "rewards/chosen": -0.017167862504720688,
889
+ "rewards/margins": 0.10066574811935425,
890
+ "rewards/rejected": -0.11783361434936523,
891
+ "step": 490
892
+ },
893
+ {
894
+ "epoch": 0.8,
895
+ "grad_norm": 1.4262079801129468,
896
+ "learning_rate": 1.5000000000000002e-05,
897
+ "log_odds_chosen": 2.723966121673584,
898
+ "log_odds_ratio": -0.116343654692173,
899
+ "logits/chosen": 22.94664764404297,
900
+ "logits/rejected": 26.360061645507812,
901
+ "logps/chosen": -0.17480744421482086,
902
+ "logps/rejected": -1.2685492038726807,
903
+ "loss": 0.145,
904
+ "nll_loss": 0.13372007012367249,
905
+ "rewards/accuracies": 1.0,
906
+ "rewards/chosen": -0.017480744048953056,
907
+ "rewards/margins": 0.10937418043613434,
908
+ "rewards/rejected": -0.12685491144657135,
909
+ "step": 500
910
+ },
911
+ {
912
+ "epoch": 0.816,
913
+ "grad_norm": 1.961973010059407,
914
+ "learning_rate": 1.4756242090702756e-05,
915
+ "log_odds_chosen": 2.9430527687072754,
916
+ "log_odds_ratio": -0.1177729144692421,
917
+ "logits/chosen": 24.766284942626953,
918
+ "logits/rejected": 28.823827743530273,
919
+ "logps/chosen": -0.16511258482933044,
920
+ "logps/rejected": -1.238128423690796,
921
+ "loss": 0.1441,
922
+ "nll_loss": 0.12999919056892395,
923
+ "rewards/accuracies": 0.987500011920929,
924
+ "rewards/chosen": -0.016511257737874985,
925
+ "rewards/margins": 0.10730159282684326,
926
+ "rewards/rejected": -0.12381285429000854,
927
+ "step": 510
928
+ },
929
+ {
930
+ "epoch": 0.832,
931
+ "grad_norm": 0.8171513438848954,
932
+ "learning_rate": 1.4508775406894308e-05,
933
+ "log_odds_chosen": 2.880876064300537,
934
+ "log_odds_ratio": -0.10505475103855133,
935
+ "logits/chosen": 22.384557723999023,
936
+ "logits/rejected": 26.74051856994629,
937
+ "logps/chosen": -0.1716020405292511,
938
+ "logps/rejected": -1.2871472835540771,
939
+ "loss": 0.1375,
940
+ "nll_loss": 0.12949416041374207,
941
+ "rewards/accuracies": 1.0,
942
+ "rewards/chosen": -0.01716020330786705,
943
+ "rewards/margins": 0.11155451834201813,
944
+ "rewards/rejected": -0.12871472537517548,
945
+ "step": 520
946
+ },
947
+ {
948
+ "epoch": 0.848,
949
+ "grad_norm": 0.6656893252356583,
950
+ "learning_rate": 1.4257792915650728e-05,
951
+ "log_odds_chosen": 3.0356669425964355,
952
+ "log_odds_ratio": -0.10586954653263092,
953
+ "logits/chosen": 19.979793548583984,
954
+ "logits/rejected": 23.590118408203125,
955
+ "logps/chosen": -0.1698727309703827,
956
+ "logps/rejected": -1.3184887170791626,
957
+ "loss": 0.127,
958
+ "nll_loss": 0.12099827826023102,
959
+ "rewards/accuracies": 1.0,
960
+ "rewards/chosen": -0.0169872734695673,
961
+ "rewards/margins": 0.11486159265041351,
962
+ "rewards/rejected": -0.13184888660907745,
963
+ "step": 530
964
+ },
965
+ {
966
+ "epoch": 0.864,
967
+ "grad_norm": 0.5859755094542713,
968
+ "learning_rate": 1.4003490325568953e-05,
969
+ "log_odds_chosen": 3.2319984436035156,
970
+ "log_odds_ratio": -0.09660033136606216,
971
+ "logits/chosen": 21.624605178833008,
972
+ "logits/rejected": 24.835023880004883,
973
+ "logps/chosen": -0.1984686404466629,
974
+ "logps/rejected": -1.5682591199874878,
975
+ "loss": 0.1372,
976
+ "nll_loss": 0.12878955900669098,
977
+ "rewards/accuracies": 0.987500011920929,
978
+ "rewards/chosen": -0.01984686590731144,
979
+ "rewards/margins": 0.13697902858257294,
980
+ "rewards/rejected": -0.1568259298801422,
981
+ "step": 540
982
+ },
983
+ {
984
+ "epoch": 0.88,
985
+ "grad_norm": 0.779222909675585,
986
+ "learning_rate": 1.3746065934159123e-05,
987
+ "log_odds_chosen": 3.039996862411499,
988
+ "log_odds_ratio": -0.10127159208059311,
989
+ "logits/chosen": 25.058979034423828,
990
+ "logits/rejected": 27.096914291381836,
991
+ "logps/chosen": -0.18606023490428925,
992
+ "logps/rejected": -1.4767307043075562,
993
+ "loss": 0.137,
994
+ "nll_loss": 0.14917424321174622,
995
+ "rewards/accuracies": 1.0,
996
+ "rewards/chosen": -0.018606023862957954,
997
+ "rewards/margins": 0.12906703352928162,
998
+ "rewards/rejected": -0.1476730853319168,
999
+ "step": 550
1000
+ },
1001
+ {
1002
+ "epoch": 0.896,
1003
+ "grad_norm": 0.5131532913955095,
1004
+ "learning_rate": 1.3485720473218153e-05,
1005
+ "log_odds_chosen": 3.5183768272399902,
1006
+ "log_odds_ratio": -0.08511906862258911,
1007
+ "logits/chosen": 21.076614379882812,
1008
+ "logits/rejected": 24.55303955078125,
1009
+ "logps/chosen": -0.15598443150520325,
1010
+ "logps/rejected": -1.6431341171264648,
1011
+ "loss": 0.1361,
1012
+ "nll_loss": 0.12268916517496109,
1013
+ "rewards/accuracies": 1.0,
1014
+ "rewards/chosen": -0.015598444268107414,
1015
+ "rewards/margins": 0.14871497452259064,
1016
+ "rewards/rejected": -0.164313405752182,
1017
+ "step": 560
1018
+ },
1019
+ {
1020
+ "epoch": 0.912,
1021
+ "grad_norm": 0.7652230384669911,
1022
+ "learning_rate": 1.3222656952305113e-05,
1023
+ "log_odds_chosen": 3.6853690147399902,
1024
+ "log_odds_ratio": -0.08712838590145111,
1025
+ "logits/chosen": 20.190637588500977,
1026
+ "logits/rejected": 25.091556549072266,
1027
+ "logps/chosen": -0.13156656920909882,
1028
+ "logps/rejected": -1.6401069164276123,
1029
+ "loss": 0.121,
1030
+ "nll_loss": 0.10304377228021622,
1031
+ "rewards/accuracies": 1.0,
1032
+ "rewards/chosen": -0.01315665990114212,
1033
+ "rewards/margins": 0.15085402131080627,
1034
+ "rewards/rejected": -0.1640106737613678,
1035
+ "step": 570
1036
+ },
1037
+ {
1038
+ "epoch": 0.928,
1039
+ "grad_norm": 0.7300290151409546,
1040
+ "learning_rate": 1.2957080500440469e-05,
1041
+ "log_odds_chosen": 3.642732620239258,
1042
+ "log_odds_ratio": -0.07864442467689514,
1043
+ "logits/chosen": 19.066701889038086,
1044
+ "logits/rejected": 23.404861450195312,
1045
+ "logps/chosen": -0.14534644782543182,
1046
+ "logps/rejected": -1.6337239742279053,
1047
+ "loss": 0.1262,
1048
+ "nll_loss": 0.11143051087856293,
1049
+ "rewards/accuracies": 1.0,
1050
+ "rewards/chosen": -0.014534644782543182,
1051
+ "rewards/margins": 0.14883777499198914,
1052
+ "rewards/rejected": -0.16337242722511292,
1053
+ "step": 580
1054
+ },
1055
+ {
1056
+ "epoch": 0.944,
1057
+ "grad_norm": 0.581668451170283,
1058
+ "learning_rate": 1.2689198206152657e-05,
1059
+ "log_odds_chosen": 3.713766098022461,
1060
+ "log_odds_ratio": -0.0588177926838398,
1061
+ "logits/chosen": 22.885316848754883,
1062
+ "logits/rejected": 25.525924682617188,
1063
+ "logps/chosen": -0.14779958128929138,
1064
+ "logps/rejected": -1.8394057750701904,
1065
+ "loss": 0.1234,
1066
+ "nll_loss": 0.11549869924783707,
1067
+ "rewards/accuracies": 1.0,
1068
+ "rewards/chosen": -0.014779957942664623,
1069
+ "rewards/margins": 0.1691606342792511,
1070
+ "rewards/rejected": -0.1839405745267868,
1071
+ "step": 590
1072
+ },
1073
+ {
1074
+ "epoch": 0.96,
1075
+ "grad_norm": 0.566646644489029,
1076
+ "learning_rate": 1.2419218955996677e-05,
1077
+ "log_odds_chosen": 3.757345199584961,
1078
+ "log_odds_ratio": -0.06075813248753548,
1079
+ "logits/chosen": 17.20478057861328,
1080
+ "logits/rejected": 22.157577514648438,
1081
+ "logps/chosen": -0.11853750050067902,
1082
+ "logps/rejected": -1.5277284383773804,
1083
+ "loss": 0.1127,
1084
+ "nll_loss": 0.09783687442541122,
1085
+ "rewards/accuracies": 1.0,
1086
+ "rewards/chosen": -0.011853748932480812,
1087
+ "rewards/margins": 0.14091908931732178,
1088
+ "rewards/rejected": -0.15277281403541565,
1089
+ "step": 600
1090
+ },
1091
+ {
1092
+ "epoch": 0.976,
1093
+ "grad_norm": 0.6375581513651442,
1094
+ "learning_rate": 1.2147353271670634e-05,
1095
+ "log_odds_chosen": 3.8227837085723877,
1096
+ "log_odds_ratio": -0.09637900441884995,
1097
+ "logits/chosen": 16.64263916015625,
1098
+ "logits/rejected": 21.630008697509766,
1099
+ "logps/chosen": -0.1534954011440277,
1100
+ "logps/rejected": -1.8045285940170288,
1101
+ "loss": 0.1165,
1102
+ "nll_loss": 0.11545213311910629,
1103
+ "rewards/accuracies": 0.9750000238418579,
1104
+ "rewards/chosen": -0.015349539928138256,
1105
+ "rewards/margins": 0.16510331630706787,
1106
+ "rewards/rejected": -0.1804528534412384,
1107
+ "step": 610
1108
+ },
1109
+ {
1110
+ "epoch": 0.992,
1111
+ "grad_norm": 0.5653677883458513,
1112
+ "learning_rate": 1.187381314585725e-05,
1113
+ "log_odds_chosen": 3.73738431930542,
1114
+ "log_odds_ratio": -0.06889805942773819,
1115
+ "logits/chosen": 20.336971282958984,
1116
+ "logits/rejected": 23.924381256103516,
1117
+ "logps/chosen": -0.12372555583715439,
1118
+ "logps/rejected": -1.6133959293365479,
1119
+ "loss": 0.1249,
1120
+ "nll_loss": 0.1038857102394104,
1121
+ "rewards/accuracies": 1.0,
1122
+ "rewards/chosen": -0.012372555211186409,
1123
+ "rewards/margins": 0.14896704256534576,
1124
+ "rewards/rejected": -0.16133958101272583,
1125
+ "step": 620
1126
+ },
1127
+ {
1128
+ "epoch": 1.008,
1129
+ "grad_norm": 0.41947247295916856,
1130
+ "learning_rate": 1.159881187691835e-05,
1131
+ "log_odds_chosen": 3.937540054321289,
1132
+ "log_odds_ratio": -0.05745882913470268,
1133
+ "logits/chosen": 18.518375396728516,
1134
+ "logits/rejected": 22.057506561279297,
1135
+ "logps/chosen": -0.12291697412729263,
1136
+ "logps/rejected": -1.742640495300293,
1137
+ "loss": 0.1059,
1138
+ "nll_loss": 0.09989026188850403,
1139
+ "rewards/accuracies": 1.0,
1140
+ "rewards/chosen": -0.012291697785258293,
1141
+ "rewards/margins": 0.16197235882282257,
1142
+ "rewards/rejected": -0.17426404356956482,
1143
+ "step": 630
1144
+ },
1145
+ {
1146
+ "epoch": 1.024,
1147
+ "grad_norm": 0.6597379471767634,
1148
+ "learning_rate": 1.1322563902571227e-05,
1149
+ "log_odds_chosen": 4.9311842918396,
1150
+ "log_odds_ratio": -0.028991302475333214,
1151
+ "logits/chosen": 12.143542289733887,
1152
+ "logits/rejected": 17.73841094970703,
1153
+ "logps/chosen": -0.06773394346237183,
1154
+ "logps/rejected": -2.039699077606201,
1155
+ "loss": 0.0792,
1156
+ "nll_loss": 0.07453403621912003,
1157
+ "rewards/accuracies": 1.0,
1158
+ "rewards/chosen": -0.0067733945325016975,
1159
+ "rewards/margins": 0.19719651341438293,
1160
+ "rewards/rejected": -0.20396992564201355,
1161
+ "step": 640
1162
+ },
1163
+ {
1164
+ "epoch": 1.04,
1165
+ "grad_norm": 0.8720926641059171,
1166
+ "learning_rate": 1.1045284632676535e-05,
1167
+ "log_odds_chosen": 4.683058261871338,
1168
+ "log_odds_ratio": -0.025352224707603455,
1169
+ "logits/chosen": 13.8761625289917,
1170
+ "logits/rejected": 19.721921920776367,
1171
+ "logps/chosen": -0.08715043216943741,
1172
+ "logps/rejected": -1.9542983770370483,
1173
+ "loss": 0.0855,
1174
+ "nll_loss": 0.0856931135058403,
1175
+ "rewards/accuracies": 1.0,
1176
+ "rewards/chosen": -0.008715043775737286,
1177
+ "rewards/margins": 0.18671482801437378,
1178
+ "rewards/rejected": -0.19542984664440155,
1179
+ "step": 650
1180
+ },
1181
+ {
1182
+ "epoch": 1.056,
1183
+ "grad_norm": 0.5401242210481009,
1184
+ "learning_rate": 1.0767190281268187e-05,
1185
+ "log_odds_chosen": 4.751006603240967,
1186
+ "log_odds_ratio": -0.03235021233558655,
1187
+ "logits/chosen": 14.378767013549805,
1188
+ "logits/rejected": 19.829517364501953,
1189
+ "logps/chosen": -0.07587035000324249,
1190
+ "logps/rejected": -2.011232614517212,
1191
+ "loss": 0.0829,
1192
+ "nll_loss": 0.07639533281326294,
1193
+ "rewards/accuracies": 1.0,
1194
+ "rewards/chosen": -0.007587035186588764,
1195
+ "rewards/margins": 0.19353622198104858,
1196
+ "rewards/rejected": -0.20112323760986328,
1197
+ "step": 660
1198
+ },
1199
+ {
1200
+ "epoch": 1.072,
1201
+ "grad_norm": 0.4975498251334291,
1202
+ "learning_rate": 1.0488497697956134e-05,
1203
+ "log_odds_chosen": 4.578000068664551,
1204
+ "log_odds_ratio": -0.03245697170495987,
1205
+ "logits/chosen": 13.391039848327637,
1206
+ "logits/rejected": 18.039913177490234,
1207
+ "logps/chosen": -0.08072780072689056,
1208
+ "logps/rejected": -1.7768867015838623,
1209
+ "loss": 0.0835,
1210
+ "nll_loss": 0.07963033765554428,
1211
+ "rewards/accuracies": 1.0,
1212
+ "rewards/chosen": -0.008072780445218086,
1213
+ "rewards/margins": 0.16961589455604553,
1214
+ "rewards/rejected": -0.17768868803977966,
1215
+ "step": 670
1216
+ },
1217
+ {
1218
+ "epoch": 1.088,
1219
+ "grad_norm": 0.4579496342491343,
1220
+ "learning_rate": 1.0209424198833571e-05,
1221
+ "log_odds_chosen": 4.71366024017334,
1222
+ "log_odds_ratio": -0.029907289892435074,
1223
+ "logits/chosen": 11.174861907958984,
1224
+ "logits/rejected": 16.83877182006836,
1225
+ "logps/chosen": -0.08100271224975586,
1226
+ "logps/rejected": -1.965436339378357,
1227
+ "loss": 0.0782,
1228
+ "nll_loss": 0.07603878527879715,
1229
+ "rewards/accuracies": 1.0,
1230
+ "rewards/chosen": -0.008100271224975586,
1231
+ "rewards/margins": 0.1884433925151825,
1232
+ "rewards/rejected": -0.19654366374015808,
1233
+ "step": 680
1234
+ },
1235
+ {
1236
+ "epoch": 1.104,
1237
+ "grad_norm": 0.41635397248386163,
1238
+ "learning_rate": 9.930187397020385e-06,
1239
+ "log_odds_chosen": 4.987896919250488,
1240
+ "log_odds_ratio": -0.029992487281560898,
1241
+ "logits/chosen": 10.937190055847168,
1242
+ "logits/rejected": 16.221263885498047,
1243
+ "logps/chosen": -0.0843561664223671,
1244
+ "logps/rejected": -1.950840950012207,
1245
+ "loss": 0.0802,
1246
+ "nll_loss": 0.07738173007965088,
1247
+ "rewards/accuracies": 1.0,
1248
+ "rewards/chosen": -0.00843561626970768,
1249
+ "rewards/margins": 0.18664847314357758,
1250
+ "rewards/rejected": -0.1950840950012207,
1251
+ "step": 690
1252
+ },
1253
+ {
1254
+ "epoch": 1.12,
1255
+ "grad_norm": 0.42432934191231536,
1256
+ "learning_rate": 9.651005032974994e-06,
1257
+ "log_odds_chosen": 5.254538059234619,
1258
+ "log_odds_ratio": -0.019603287801146507,
1259
+ "logits/chosen": 12.815790176391602,
1260
+ "logits/rejected": 16.598247528076172,
1261
+ "logps/chosen": -0.07913420349359512,
1262
+ "logps/rejected": -2.306891918182373,
1263
+ "loss": 0.0795,
1264
+ "nll_loss": 0.07783981412649155,
1265
+ "rewards/accuracies": 1.0,
1266
+ "rewards/chosen": -0.007913419976830482,
1267
+ "rewards/margins": 0.22277574241161346,
1268
+ "rewards/rejected": -0.23068919777870178,
1269
+ "step": 700
1270
+ },
1271
+ {
1272
+ "epoch": 1.1360000000000001,
1273
+ "grad_norm": 0.5144220041987062,
1274
+ "learning_rate": 9.372094804706867e-06,
1275
+ "log_odds_chosen": 5.051201820373535,
1276
+ "log_odds_ratio": -0.022446967661380768,
1277
+ "logits/chosen": 8.465599060058594,
1278
+ "logits/rejected": 15.262280464172363,
1279
+ "logps/chosen": -0.04941285401582718,
1280
+ "logps/rejected": -1.7998939752578735,
1281
+ "loss": 0.0699,
1282
+ "nll_loss": 0.06099081039428711,
1283
+ "rewards/accuracies": 1.0,
1284
+ "rewards/chosen": -0.0049412851221859455,
1285
+ "rewards/margins": 0.1750481128692627,
1286
+ "rewards/rejected": -0.17998939752578735,
1287
+ "step": 710
1288
+ },
1289
+ {
1290
+ "epoch": 1.152,
1291
+ "grad_norm": 0.4355953213178483,
1292
+ "learning_rate": 9.093674198022201e-06,
1293
+ "log_odds_chosen": 4.678675651550293,
1294
+ "log_odds_ratio": -0.03710417076945305,
1295
+ "logits/chosen": 9.093629837036133,
1296
+ "logits/rejected": 15.341288566589355,
1297
+ "logps/chosen": -0.0850113034248352,
1298
+ "logps/rejected": -1.8611046075820923,
1299
+ "loss": 0.0802,
1300
+ "nll_loss": 0.08102207630872726,
1301
+ "rewards/accuracies": 0.987500011920929,
1302
+ "rewards/chosen": -0.008501130156219006,
1303
+ "rewards/margins": 0.17760933935642242,
1304
+ "rewards/rejected": -0.1861104667186737,
1305
+ "step": 720
1306
+ },
1307
+ {
1308
+ "epoch": 1.168,
1309
+ "grad_norm": 0.4183273277614356,
1310
+ "learning_rate": 8.815960316934991e-06,
1311
+ "log_odds_chosen": 4.754403591156006,
1312
+ "log_odds_ratio": -0.026562413200736046,
1313
+ "logits/chosen": 11.237001419067383,
1314
+ "logits/rejected": 16.33772087097168,
1315
+ "logps/chosen": -0.0855894535779953,
1316
+ "logps/rejected": -2.021867275238037,
1317
+ "loss": 0.074,
1318
+ "nll_loss": 0.07607786357402802,
1319
+ "rewards/accuracies": 1.0,
1320
+ "rewards/chosen": -0.00855894573032856,
1321
+ "rewards/margins": 0.1936277598142624,
1322
+ "rewards/rejected": -0.2021867036819458,
1323
+ "step": 730
1324
+ },
1325
+ {
1326
+ "epoch": 1.184,
1327
+ "grad_norm": 0.5708367601297668,
1328
+ "learning_rate": 8.539169714375885e-06,
1329
+ "log_odds_chosen": 5.499794960021973,
1330
+ "log_odds_ratio": -0.015643080696463585,
1331
+ "logits/chosen": 11.671289443969727,
1332
+ "logits/rejected": 15.871813774108887,
1333
+ "logps/chosen": -0.0698837861418724,
1334
+ "logps/rejected": -2.4058656692504883,
1335
+ "loss": 0.0758,
1336
+ "nll_loss": 0.06816570460796356,
1337
+ "rewards/accuracies": 1.0,
1338
+ "rewards/chosen": -0.006988379172980785,
1339
+ "rewards/margins": 0.23359815776348114,
1340
+ "rewards/rejected": -0.2405865490436554,
1341
+ "step": 740
1342
+ },
1343
+ {
1344
+ "epoch": 1.2,
1345
+ "grad_norm": 0.46974989560288,
1346
+ "learning_rate": 8.263518223330698e-06,
1347
+ "log_odds_chosen": 5.374692916870117,
1348
+ "log_odds_ratio": -0.01755271479487419,
1349
+ "logits/chosen": 10.17614459991455,
1350
+ "logits/rejected": 15.561047554016113,
1351
+ "logps/chosen": -0.06151513010263443,
1352
+ "logps/rejected": -2.1926753520965576,
1353
+ "loss": 0.0765,
1354
+ "nll_loss": 0.06682710349559784,
1355
+ "rewards/accuracies": 1.0,
1356
+ "rewards/chosen": -0.006151513196527958,
1357
+ "rewards/margins": 0.21311601996421814,
1358
+ "rewards/rejected": -0.21926753222942352,
1359
+ "step": 750
1360
+ },
1361
+ {
1362
+ "epoch": 1.216,
1363
+ "grad_norm": 0.536798190903821,
1364
+ "learning_rate": 7.989220788540356e-06,
1365
+ "log_odds_chosen": 5.090784549713135,
1366
+ "log_odds_ratio": -0.024079417809844017,
1367
+ "logits/chosen": 10.42772102355957,
1368
+ "logits/rejected": 16.83907127380371,
1369
+ "logps/chosen": -0.0742868185043335,
1370
+ "logps/rejected": -1.9404573440551758,
1371
+ "loss": 0.0767,
1372
+ "nll_loss": 0.07289760559797287,
1373
+ "rewards/accuracies": 1.0,
1374
+ "rewards/chosen": -0.007428681943565607,
1375
+ "rewards/margins": 0.18661704659461975,
1376
+ "rewards/rejected": -0.19404573738574982,
1377
+ "step": 760
1378
+ },
1379
+ {
1380
+ "epoch": 1.232,
1381
+ "grad_norm": 0.7535095663117326,
1382
+ "learning_rate": 7.716491298893443e-06,
1383
+ "log_odds_chosen": 5.1247100830078125,
1384
+ "log_odds_ratio": -0.027670254930853844,
1385
+ "logits/chosen": 13.019775390625,
1386
+ "logits/rejected": 18.80875015258789,
1387
+ "logps/chosen": -0.08515826612710953,
1388
+ "logps/rejected": -2.227242946624756,
1389
+ "loss": 0.0799,
1390
+ "nll_loss": 0.08172322809696198,
1391
+ "rewards/accuracies": 1.0,
1392
+ "rewards/chosen": -0.008515825495123863,
1393
+ "rewards/margins": 0.2142084836959839,
1394
+ "rewards/rejected": -0.2227243185043335,
1395
+ "step": 770
1396
+ },
1397
+ {
1398
+ "epoch": 1.248,
1399
+ "grad_norm": 0.43802198454777963,
1400
+ "learning_rate": 7.445542420642097e-06,
1401
+ "log_odds_chosen": 4.944367408752441,
1402
+ "log_odds_ratio": -0.019510487094521523,
1403
+ "logits/chosen": 12.673688888549805,
1404
+ "logits/rejected": 17.701494216918945,
1405
+ "logps/chosen": -0.06077063828706741,
1406
+ "logps/rejected": -1.8956024646759033,
1407
+ "loss": 0.0722,
1408
+ "nll_loss": 0.06931096315383911,
1409
+ "rewards/accuracies": 1.0,
1410
+ "rewards/chosen": -0.0060770632699131966,
1411
+ "rewards/margins": 0.18348316848278046,
1412
+ "rewards/rejected": -0.18956024944782257,
1413
+ "step": 780
1414
+ },
1415
+ {
1416
+ "epoch": 1.264,
1417
+ "grad_norm": 0.5918184881161849,
1418
+ "learning_rate": 7.176585431571235e-06,
1419
+ "log_odds_chosen": 4.934049606323242,
1420
+ "log_odds_ratio": -0.022184232249855995,
1421
+ "logits/chosen": 10.117755889892578,
1422
+ "logits/rejected": 16.32207489013672,
1423
+ "logps/chosen": -0.06841601431369781,
1424
+ "logps/rejected": -1.8420988321304321,
1425
+ "loss": 0.076,
1426
+ "nll_loss": 0.07008203864097595,
1427
+ "rewards/accuracies": 1.0,
1428
+ "rewards/chosen": -0.006841603666543961,
1429
+ "rewards/margins": 0.17736826837062836,
1430
+ "rewards/rejected": -0.1842098832130432,
1431
+ "step": 790
1432
+ },
1433
+ {
1434
+ "epoch": 1.28,
1435
+ "grad_norm": 0.4109794535608094,
1436
+ "learning_rate": 6.909830056250527e-06,
1437
+ "log_odds_chosen": 5.385331153869629,
1438
+ "log_odds_ratio": -0.01698954775929451,
1439
+ "logits/chosen": 9.808797836303711,
1440
+ "logits/rejected": 15.439534187316895,
1441
+ "logps/chosen": -0.06334672123193741,
1442
+ "logps/rejected": -2.0740480422973633,
1443
+ "loss": 0.0735,
1444
+ "nll_loss": 0.06666533648967743,
1445
+ "rewards/accuracies": 1.0,
1446
+ "rewards/chosen": -0.006334672216325998,
1447
+ "rewards/margins": 0.2010701447725296,
1448
+ "rewards/rejected": -0.20740482211112976,
1449
+ "step": 800
1450
+ },
1451
+ {
1452
+ "epoch": 1.296,
1453
+ "grad_norm": 0.5786524901744241,
1454
+ "learning_rate": 6.645484302497452e-06,
1455
+ "log_odds_chosen": 5.305611610412598,
1456
+ "log_odds_ratio": -0.015638932585716248,
1457
+ "logits/chosen": 10.026269912719727,
1458
+ "logits/rejected": 16.038047790527344,
1459
+ "logps/chosen": -0.058640025556087494,
1460
+ "logps/rejected": -2.095346212387085,
1461
+ "loss": 0.0687,
1462
+ "nll_loss": 0.0625784695148468,
1463
+ "rewards/accuracies": 1.0,
1464
+ "rewards/chosen": -0.005864003207534552,
1465
+ "rewards/margins": 0.20367059111595154,
1466
+ "rewards/rejected": -0.20953461527824402,
1467
+ "step": 810
1468
+ },
1469
+ {
1470
+ "epoch": 1.312,
1471
+ "grad_norm": 0.413245918888085,
1472
+ "learning_rate": 6.383754299179079e-06,
1473
+ "log_odds_chosen": 5.338342189788818,
1474
+ "log_odds_ratio": -0.016953352838754654,
1475
+ "logits/chosen": 10.80897331237793,
1476
+ "logits/rejected": 16.324369430541992,
1477
+ "logps/chosen": -0.07062120735645294,
1478
+ "logps/rejected": -2.1152849197387695,
1479
+ "loss": 0.0753,
1480
+ "nll_loss": 0.07253970950841904,
1481
+ "rewards/accuracies": 1.0,
1482
+ "rewards/chosen": -0.007062119897454977,
1483
+ "rewards/margins": 0.20446637272834778,
1484
+ "rewards/rejected": -0.2115284949541092,
1485
+ "step": 820
1486
+ },
1487
+ {
1488
+ "epoch": 1.328,
1489
+ "grad_norm": 0.45473358832367533,
1490
+ "learning_rate": 6.124844135478971e-06,
1491
+ "log_odds_chosen": 4.9816741943359375,
1492
+ "log_odds_ratio": -0.016600508242845535,
1493
+ "logits/chosen": 12.60502815246582,
1494
+ "logits/rejected": 17.482601165771484,
1495
+ "logps/chosen": -0.07652293890714645,
1496
+ "logps/rejected": -1.9486392736434937,
1497
+ "loss": 0.0753,
1498
+ "nll_loss": 0.0775555893778801,
1499
+ "rewards/accuracies": 1.0,
1500
+ "rewards/chosen": -0.007652295287698507,
1501
+ "rewards/margins": 0.18721160292625427,
1502
+ "rewards/rejected": -0.1948639154434204,
1503
+ "step": 830
1504
+ },
1505
+ {
1506
+ "epoch": 1.3439999999999999,
1507
+ "grad_norm": 0.43221707053355657,
1508
+ "learning_rate": 5.868955701754584e-06,
1509
+ "log_odds_chosen": 5.058526039123535,
1510
+ "log_odds_ratio": -0.02234785258769989,
1511
+ "logits/chosen": 10.739975929260254,
1512
+ "logits/rejected": 14.592979431152344,
1513
+ "logps/chosen": -0.08031310141086578,
1514
+ "logps/rejected": -2.022430896759033,
1515
+ "loss": 0.071,
1516
+ "nll_loss": 0.07450385391712189,
1517
+ "rewards/accuracies": 1.0,
1518
+ "rewards/chosen": -0.008031310513615608,
1519
+ "rewards/margins": 0.19421175122261047,
1520
+ "rewards/rejected": -0.20224305987358093,
1521
+ "step": 840
1522
+ },
1523
+ {
1524
+ "epoch": 1.3599999999999999,
1525
+ "grad_norm": 0.36475019984330276,
1526
+ "learning_rate": 5.616288532109225e-06,
1527
+ "log_odds_chosen": 5.216296672821045,
1528
+ "log_odds_ratio": -0.02249590866267681,
1529
+ "logits/chosen": 10.627111434936523,
1530
+ "logits/rejected": 16.160968780517578,
1531
+ "logps/chosen": -0.08645172417163849,
1532
+ "logps/rejected": -2.255333662033081,
1533
+ "loss": 0.0741,
1534
+ "nll_loss": 0.07864928245544434,
1535
+ "rewards/accuracies": 1.0,
1536
+ "rewards/chosen": -0.008645173162221909,
1537
+ "rewards/margins": 0.2168881893157959,
1538
+ "rewards/rejected": -0.2255333662033081,
1539
+ "step": 850
1540
+ },
1541
+ {
1542
+ "epoch": 1.376,
1543
+ "grad_norm": 0.5386028507331131,
1544
+ "learning_rate": 5.367039648801386e-06,
1545
+ "log_odds_chosen": 5.349644184112549,
1546
+ "log_odds_ratio": -0.019762977957725525,
1547
+ "logits/chosen": 11.73155689239502,
1548
+ "logits/rejected": 16.905214309692383,
1549
+ "logps/chosen": -0.07144279778003693,
1550
+ "logps/rejected": -2.1353299617767334,
1551
+ "loss": 0.073,
1552
+ "nll_loss": 0.07140424847602844,
1553
+ "rewards/accuracies": 1.0,
1554
+ "rewards/chosen": -0.007144278846681118,
1555
+ "rewards/margins": 0.20638874173164368,
1556
+ "rewards/rejected": -0.21353299915790558,
1557
+ "step": 860
1558
+ },
1559
+ {
1560
+ "epoch": 1.392,
1561
+ "grad_norm": 0.4861620483022773,
1562
+ "learning_rate": 5.121403408612672e-06,
1563
+ "log_odds_chosen": 5.4971466064453125,
1564
+ "log_odds_ratio": -0.013068562373518944,
1565
+ "logits/chosen": 10.185466766357422,
1566
+ "logits/rejected": 15.52686595916748,
1567
+ "logps/chosen": -0.05803658813238144,
1568
+ "logps/rejected": -2.2335634231567383,
1569
+ "loss": 0.071,
1570
+ "nll_loss": 0.06299170106649399,
1571
+ "rewards/accuracies": 1.0,
1572
+ "rewards/chosen": -0.005803658626973629,
1573
+ "rewards/margins": 0.21755269169807434,
1574
+ "rewards/rejected": -0.22335636615753174,
1575
+ "step": 870
1576
+ },
1577
+ {
1578
+ "epoch": 1.408,
1579
+ "grad_norm": 0.7221352144813555,
1580
+ "learning_rate": 4.879571351294287e-06,
1581
+ "log_odds_chosen": 5.298310279846191,
1582
+ "log_odds_ratio": -0.015498277731239796,
1583
+ "logits/chosen": 8.6812744140625,
1584
+ "logits/rejected": 14.908788681030273,
1585
+ "logps/chosen": -0.04598275199532509,
1586
+ "logps/rejected": -1.8670889139175415,
1587
+ "loss": 0.0695,
1588
+ "nll_loss": 0.05738530308008194,
1589
+ "rewards/accuracies": 1.0,
1590
+ "rewards/chosen": -0.004598274827003479,
1591
+ "rewards/margins": 0.18211060762405396,
1592
+ "rewards/rejected": -0.18670889735221863,
1593
+ "step": 880
1594
+ },
1595
+ {
1596
+ "epoch": 1.424,
1597
+ "grad_norm": 0.4395795057245911,
1598
+ "learning_rate": 4.641732050210032e-06,
1599
+ "log_odds_chosen": 5.754244327545166,
1600
+ "log_odds_ratio": -0.013112393207848072,
1601
+ "logits/chosen": 10.254205703735352,
1602
+ "logits/rejected": 16.640913009643555,
1603
+ "logps/chosen": -0.07648586481809616,
1604
+ "logps/rejected": -2.492318630218506,
1605
+ "loss": 0.0712,
1606
+ "nll_loss": 0.07544175535440445,
1607
+ "rewards/accuracies": 1.0,
1608
+ "rewards/chosen": -0.007648587226867676,
1609
+ "rewards/margins": 0.24158325791358948,
1610
+ "rewards/rejected": -0.24923184514045715,
1611
+ "step": 890
1612
+ },
1613
+ {
1614
+ "epoch": 1.44,
1615
+ "grad_norm": 0.6571181936107011,
1616
+ "learning_rate": 4.408070965292534e-06,
1617
+ "log_odds_chosen": 5.259117126464844,
1618
+ "log_odds_ratio": -0.016550693660974503,
1619
+ "logits/chosen": 10.593389511108398,
1620
+ "logits/rejected": 15.269793510437012,
1621
+ "logps/chosen": -0.08531410992145538,
1622
+ "logps/rejected": -2.2608280181884766,
1623
+ "loss": 0.0758,
1624
+ "nll_loss": 0.07715515047311783,
1625
+ "rewards/accuracies": 1.0,
1626
+ "rewards/chosen": -0.008531411178410053,
1627
+ "rewards/margins": 0.21755139529705048,
1628
+ "rewards/rejected": -0.22608280181884766,
1629
+ "step": 900
1630
+ },
1631
+ {
1632
+ "epoch": 1.456,
1633
+ "grad_norm": 0.45529898147163606,
1634
+ "learning_rate": 4.178770298427107e-06,
1635
+ "log_odds_chosen": 5.506943702697754,
1636
+ "log_odds_ratio": -0.017402973026037216,
1637
+ "logits/chosen": 10.588479995727539,
1638
+ "logits/rejected": 16.51026153564453,
1639
+ "logps/chosen": -0.06945344805717468,
1640
+ "logps/rejected": -2.281757354736328,
1641
+ "loss": 0.0701,
1642
+ "nll_loss": 0.07317786663770676,
1643
+ "rewards/accuracies": 1.0,
1644
+ "rewards/chosen": -0.006945345550775528,
1645
+ "rewards/margins": 0.2212304174900055,
1646
+ "rewards/rejected": -0.22817575931549072,
1647
+ "step": 910
1648
+ },
1649
+ {
1650
+ "epoch": 1.472,
1651
+ "grad_norm": 0.4799597661016462,
1652
+ "learning_rate": 3.954008851376252e-06,
1653
+ "log_odds_chosen": 5.671328544616699,
1654
+ "log_odds_ratio": -0.01591522805392742,
1655
+ "logits/chosen": 7.850234031677246,
1656
+ "logits/rejected": 15.029336929321289,
1657
+ "logps/chosen": -0.04982997104525566,
1658
+ "logps/rejected": -2.066762685775757,
1659
+ "loss": 0.0637,
1660
+ "nll_loss": 0.059512246400117874,
1661
+ "rewards/accuracies": 1.0,
1662
+ "rewards/chosen": -0.0049829971976578236,
1663
+ "rewards/margins": 0.20169326663017273,
1664
+ "rewards/rejected": -0.20667627453804016,
1665
+ "step": 920
1666
+ },
1667
+ {
1668
+ "epoch": 1.488,
1669
+ "grad_norm": 0.42918774258741327,
1670
+ "learning_rate": 3.7339618863553983e-06,
1671
+ "log_odds_chosen": 5.558573246002197,
1672
+ "log_odds_ratio": -0.014333250001072884,
1673
+ "logits/chosen": 8.254855155944824,
1674
+ "logits/rejected": 14.497201919555664,
1675
+ "logps/chosen": -0.06610764563083649,
1676
+ "logps/rejected": -2.231790781021118,
1677
+ "loss": 0.0638,
1678
+ "nll_loss": 0.06769216060638428,
1679
+ "rewards/accuracies": 1.0,
1680
+ "rewards/chosen": -0.0066107651218771935,
1681
+ "rewards/margins": 0.21656832098960876,
1682
+ "rewards/rejected": -0.22317907214164734,
1683
+ "step": 930
1684
+ },
1685
+ {
1686
+ "epoch": 1.504,
1687
+ "grad_norm": 0.39881406218102516,
1688
+ "learning_rate": 3.5188009893686916e-06,
1689
+ "log_odds_chosen": 5.319655418395996,
1690
+ "log_odds_ratio": -0.0184173621237278,
1691
+ "logits/chosen": 10.073064804077148,
1692
+ "logits/rejected": 15.57422924041748,
1693
+ "logps/chosen": -0.07947548478841782,
1694
+ "logps/rejected": -2.1314804553985596,
1695
+ "loss": 0.0699,
1696
+ "nll_loss": 0.07717884331941605,
1697
+ "rewards/accuracies": 1.0,
1698
+ "rewards/chosen": -0.007947548292577267,
1699
+ "rewards/margins": 0.20520050823688507,
1700
+ "rewards/rejected": -0.21314804255962372,
1701
+ "step": 940
1702
+ },
1703
+ {
1704
+ "epoch": 1.52,
1705
+ "grad_norm": 0.3003778882690616,
1706
+ "learning_rate": 3.308693936411421e-06,
1707
+ "log_odds_chosen": 5.652432441711426,
1708
+ "log_odds_ratio": -0.018848979845643044,
1709
+ "logits/chosen": 7.4497809410095215,
1710
+ "logits/rejected": 14.933306694030762,
1711
+ "logps/chosen": -0.0582701750099659,
1712
+ "logps/rejected": -2.193925142288208,
1713
+ "loss": 0.0677,
1714
+ "nll_loss": 0.06635381281375885,
1715
+ "rewards/accuracies": 1.0,
1716
+ "rewards/chosen": -0.005827018059790134,
1717
+ "rewards/margins": 0.21356551349163055,
1718
+ "rewards/rejected": -0.21939250826835632,
1719
+ "step": 950
1720
+ },
1721
+ {
1722
+ "epoch": 1.536,
1723
+ "grad_norm": 0.3009824619938579,
1724
+ "learning_rate": 3.103804562643302e-06,
1725
+ "log_odds_chosen": 5.427401065826416,
1726
+ "log_odds_ratio": -0.014070016331970692,
1727
+ "logits/chosen": 8.830900192260742,
1728
+ "logits/rejected": 14.178878784179688,
1729
+ "logps/chosen": -0.06140875071287155,
1730
+ "logps/rejected": -2.2172341346740723,
1731
+ "loss": 0.0645,
1732
+ "nll_loss": 0.06464984267950058,
1733
+ "rewards/accuracies": 1.0,
1734
+ "rewards/chosen": -0.0061408756300807,
1735
+ "rewards/margins": 0.21558256447315216,
1736
+ "rewards/rejected": -0.22172346711158752,
1737
+ "step": 960
1738
+ },
1739
+ {
1740
+ "epoch": 1.552,
1741
+ "grad_norm": 0.4056333088268534,
1742
+ "learning_rate": 2.9042926346347932e-06,
1743
+ "log_odds_chosen": 5.441522598266602,
1744
+ "log_odds_ratio": -0.014954231679439545,
1745
+ "logits/chosen": 9.046590805053711,
1746
+ "logits/rejected": 15.852540969848633,
1747
+ "logps/chosen": -0.06886409223079681,
1748
+ "logps/rejected": -2.157127618789673,
1749
+ "loss": 0.0658,
1750
+ "nll_loss": 0.06967596709728241,
1751
+ "rewards/accuracies": 1.0,
1752
+ "rewards/chosen": -0.006886409129947424,
1753
+ "rewards/margins": 0.20882634818553925,
1754
+ "rewards/rejected": -0.2157127559185028,
1755
+ "step": 970
1756
+ },
1757
+ {
1758
+ "epoch": 1.568,
1759
+ "grad_norm": 0.3196857185408296,
1760
+ "learning_rate": 2.7103137257858867e-06,
1761
+ "log_odds_chosen": 5.551312446594238,
1762
+ "log_odds_ratio": -0.011502685025334358,
1763
+ "logits/chosen": 7.783066749572754,
1764
+ "logits/rejected": 14.421549797058105,
1765
+ "logps/chosen": -0.05041471868753433,
1766
+ "logps/rejected": -2.1501784324645996,
1767
+ "loss": 0.0606,
1768
+ "nll_loss": 0.05976518243551254,
1769
+ "rewards/accuracies": 1.0,
1770
+ "rewards/chosen": -0.005041471216827631,
1771
+ "rewards/margins": 0.20997636020183563,
1772
+ "rewards/rejected": -0.21501782536506653,
1773
+ "step": 980
1774
+ },
1775
+ {
1776
+ "epoch": 1.584,
1777
+ "grad_norm": 0.5263216429559819,
1778
+ "learning_rate": 2.522019095014683e-06,
1779
+ "log_odds_chosen": 5.743101119995117,
1780
+ "log_odds_ratio": -0.010009285993874073,
1781
+ "logits/chosen": 7.407317161560059,
1782
+ "logits/rejected": 14.168241500854492,
1783
+ "logps/chosen": -0.048953376710414886,
1784
+ "logps/rejected": -2.123490810394287,
1785
+ "loss": 0.0686,
1786
+ "nll_loss": 0.058644842356443405,
1787
+ "rewards/accuracies": 1.0,
1788
+ "rewards/chosen": -0.004895338322967291,
1789
+ "rewards/margins": 0.20745377242565155,
1790
+ "rewards/rejected": -0.2123490869998932,
1791
+ "step": 990
1792
+ },
1793
+ {
1794
+ "epoch": 1.6,
1795
+ "grad_norm": 0.36561256966761685,
1796
+ "learning_rate": 2.339555568810221e-06,
1797
+ "log_odds_chosen": 5.752957344055176,
1798
+ "log_odds_ratio": -0.008349338546395302,
1799
+ "logits/chosen": 7.5777788162231445,
1800
+ "logits/rejected": 14.192586898803711,
1801
+ "logps/chosen": -0.043931931257247925,
1802
+ "logps/rejected": -2.1989071369171143,
1803
+ "loss": 0.0669,
1804
+ "nll_loss": 0.053385306149721146,
1805
+ "rewards/accuracies": 1.0,
1806
+ "rewards/chosen": -0.004393193405121565,
1807
+ "rewards/margins": 0.21549753844738007,
1808
+ "rewards/rejected": -0.21989071369171143,
1809
+ "step": 1000
1810
+ },
1811
+ {
1812
+ "epoch": 1.616,
1813
+ "grad_norm": 0.5870370963384035,
1814
+ "learning_rate": 2.163065426741603e-06,
1815
+ "log_odds_chosen": 5.90677547454834,
1816
+ "log_odds_ratio": -0.012085262686014175,
1817
+ "logits/chosen": 6.974325656890869,
1818
+ "logits/rejected": 14.128486633300781,
1819
+ "logps/chosen": -0.0529697947204113,
1820
+ "logps/rejected": -2.190331220626831,
1821
+ "loss": 0.0645,
1822
+ "nll_loss": 0.05988318473100662,
1823
+ "rewards/accuracies": 1.0,
1824
+ "rewards/chosen": -0.0052969790995121,
1825
+ "rewards/margins": 0.2137361317873001,
1826
+ "rewards/rejected": -0.2190331220626831,
1827
+ "step": 1010
1828
+ },
1829
+ {
1830
+ "epoch": 1.6320000000000001,
1831
+ "grad_norm": 0.6243099880494768,
1832
+ "learning_rate": 1.9926862905126663e-06,
1833
+ "log_odds_chosen": 5.65887975692749,
1834
+ "log_odds_ratio": -0.014939057640731335,
1835
+ "logits/chosen": 7.151690483093262,
1836
+ "logits/rejected": 13.71330451965332,
1837
+ "logps/chosen": -0.055140770971775055,
1838
+ "logps/rejected": -2.053590774536133,
1839
+ "loss": 0.0673,
1840
+ "nll_loss": 0.06196935102343559,
1841
+ "rewards/accuracies": 1.0,
1842
+ "rewards/chosen": -0.005514076910912991,
1843
+ "rewards/margins": 0.19984501600265503,
1844
+ "rewards/rejected": -0.2053591012954712,
1845
+ "step": 1020
1846
+ },
1847
+ {
1848
+ "epoch": 1.6480000000000001,
1849
+ "grad_norm": 0.405678875969962,
1850
+ "learning_rate": 1.8285510166487154e-06,
1851
+ "log_odds_chosen": 5.765079021453857,
1852
+ "log_odds_ratio": -0.014157906174659729,
1853
+ "logits/chosen": 9.832456588745117,
1854
+ "logits/rejected": 15.122549057006836,
1855
+ "logps/chosen": -0.06559108942747116,
1856
+ "logps/rejected": -2.4253296852111816,
1857
+ "loss": 0.0694,
1858
+ "nll_loss": 0.07027839124202728,
1859
+ "rewards/accuracies": 1.0,
1860
+ "rewards/chosen": -0.006559108849614859,
1861
+ "rewards/margins": 0.23597386479377747,
1862
+ "rewards/rejected": -0.24253296852111816,
1863
+ "step": 1030
1864
+ },
1865
+ {
1866
+ "epoch": 1.6640000000000001,
1867
+ "grad_norm": 0.2511510020504382,
1868
+ "learning_rate": 1.6707875928990059e-06,
1869
+ "log_odds_chosen": 5.9670586585998535,
1870
+ "log_odds_ratio": -0.007793957833200693,
1871
+ "logits/chosen": 6.463404178619385,
1872
+ "logits/rejected": 13.94349479675293,
1873
+ "logps/chosen": -0.036587394773960114,
1874
+ "logps/rejected": -2.103496789932251,
1875
+ "loss": 0.0584,
1876
+ "nll_loss": 0.05108030512928963,
1877
+ "rewards/accuracies": 1.0,
1878
+ "rewards/chosen": -0.0036587391514331102,
1879
+ "rewards/margins": 0.2066909372806549,
1880
+ "rewards/rejected": -0.2103496789932251,
1881
+ "step": 1040
1882
+ },
1883
+ {
1884
+ "epoch": 1.6800000000000002,
1885
+ "grad_norm": 0.3812260869637052,
1886
+ "learning_rate": 1.5195190384357405e-06,
1887
+ "log_odds_chosen": 5.7870683670043945,
1888
+ "log_odds_ratio": -0.01724204048514366,
1889
+ "logits/chosen": 7.535168647766113,
1890
+ "logits/rejected": 14.668806076049805,
1891
+ "logps/chosen": -0.06202416867017746,
1892
+ "logps/rejected": -2.3643112182617188,
1893
+ "loss": 0.0648,
1894
+ "nll_loss": 0.067660391330719,
1895
+ "rewards/accuracies": 1.0,
1896
+ "rewards/chosen": -0.006202416960150003,
1897
+ "rewards/margins": 0.2302287369966507,
1898
+ "rewards/rejected": -0.23643112182617188,
1899
+ "step": 1050
1900
+ },
1901
+ {
1902
+ "epoch": 1.696,
1903
+ "grad_norm": 0.3058840630848455,
1904
+ "learning_rate": 1.3748633079274254e-06,
1905
+ "log_odds_chosen": 5.669142723083496,
1906
+ "log_odds_ratio": -0.010866614989936352,
1907
+ "logits/chosen": 8.615381240844727,
1908
+ "logits/rejected": 14.441973686218262,
1909
+ "logps/chosen": -0.05342826992273331,
1910
+ "logps/rejected": -2.128041982650757,
1911
+ "loss": 0.063,
1912
+ "nll_loss": 0.06444182246923447,
1913
+ "rewards/accuracies": 1.0,
1914
+ "rewards/chosen": -0.005342827644199133,
1915
+ "rewards/margins": 0.2074614018201828,
1916
+ "rewards/rejected": -0.21280424296855927,
1917
+ "step": 1060
1918
+ },
1919
+ {
1920
+ "epoch": 1.712,
1921
+ "grad_norm": 0.3235904810469797,
1922
+ "learning_rate": 1.2369331995613664e-06,
1923
+ "log_odds_chosen": 5.627545356750488,
1924
+ "log_odds_ratio": -0.008036193437874317,
1925
+ "logits/chosen": 6.999047756195068,
1926
+ "logits/rejected": 13.646265983581543,
1927
+ "logps/chosen": -0.048848383128643036,
1928
+ "logps/rejected": -2.056873321533203,
1929
+ "loss": 0.0634,
1930
+ "nll_loss": 0.05773681402206421,
1931
+ "rewards/accuracies": 1.0,
1932
+ "rewards/chosen": -0.004884839057922363,
1933
+ "rewards/margins": 0.2008024901151657,
1934
+ "rewards/rejected": -0.20568732917308807,
1935
+ "step": 1070
1936
+ },
1937
+ {
1938
+ "epoch": 1.728,
1939
+ "grad_norm": 0.410833487441517,
1940
+ "learning_rate": 1.1058362670870248e-06,
1941
+ "log_odds_chosen": 5.63392448425293,
1942
+ "log_odds_ratio": -0.012561176903545856,
1943
+ "logits/chosen": 7.00423526763916,
1944
+ "logits/rejected": 12.398062705993652,
1945
+ "logps/chosen": -0.06395191699266434,
1946
+ "logps/rejected": -2.2964274883270264,
1947
+ "loss": 0.0622,
1948
+ "nll_loss": 0.06452804803848267,
1949
+ "rewards/accuracies": 1.0,
1950
+ "rewards/chosen": -0.006395191885530949,
1951
+ "rewards/margins": 0.22324757277965546,
1952
+ "rewards/rejected": -0.22964277863502502,
1953
+ "step": 1080
1954
+ },
1955
+ {
1956
+ "epoch": 1.744,
1957
+ "grad_norm": 0.34256892405513334,
1958
+ "learning_rate": 9.816747359488632e-07,
1959
+ "log_odds_chosen": 6.084750652313232,
1960
+ "log_odds_ratio": -0.01084104459732771,
1961
+ "logits/chosen": 7.226653099060059,
1962
+ "logits/rejected": 13.381048202514648,
1963
+ "logps/chosen": -0.05891553312540054,
1964
+ "logps/rejected": -2.476435661315918,
1965
+ "loss": 0.0609,
1966
+ "nll_loss": 0.06247776746749878,
1967
+ "rewards/accuracies": 1.0,
1968
+ "rewards/chosen": -0.005891552660614252,
1969
+ "rewards/margins": 0.241752028465271,
1970
+ "rewards/rejected": -0.24764356017112732,
1971
+ "step": 1090
1972
+ },
1973
+ {
1974
+ "epoch": 1.76,
1975
+ "grad_norm": 0.34625995571735285,
1976
+ "learning_rate": 8.645454235739903e-07,
1977
+ "log_odds_chosen": 5.164887428283691,
1978
+ "log_odds_ratio": -0.01927667111158371,
1979
+ "logits/chosen": 7.622632026672363,
1980
+ "logits/rejected": 15.065897941589355,
1981
+ "logps/chosen": -0.05562591552734375,
1982
+ "logps/rejected": -1.8294832706451416,
1983
+ "loss": 0.0671,
1984
+ "nll_loss": 0.06367551535367966,
1985
+ "rewards/accuracies": 1.0,
1986
+ "rewards/chosen": -0.00556259136646986,
1987
+ "rewards/margins": 0.17738574743270874,
1988
+ "rewards/rejected": -0.18294833600521088,
1989
+ "step": 1100
1990
+ },
1991
+ {
1992
+ "epoch": 1.776,
1993
+ "grad_norm": 0.3129225650767717,
1994
+ "learning_rate": 7.545396638768698e-07,
1995
+ "log_odds_chosen": 5.7577362060546875,
1996
+ "log_odds_ratio": -0.009081291034817696,
1997
+ "logits/chosen": 7.138357639312744,
1998
+ "logits/rejected": 13.764391899108887,
1999
+ "logps/chosen": -0.048210419714450836,
2000
+ "logps/rejected": -2.1392695903778076,
2001
+ "loss": 0.0617,
2002
+ "nll_loss": 0.0597044937312603,
2003
+ "rewards/accuracies": 1.0,
2004
+ "rewards/chosen": -0.004821042064577341,
2005
+ "rewards/margins": 0.20910592377185822,
2006
+ "rewards/rejected": -0.2139269858598709,
2007
+ "step": 1110
2008
+ },
2009
+ {
2010
+ "epoch": 1.792,
2011
+ "grad_norm": 0.3636030975331139,
2012
+ "learning_rate": 6.517432360398556e-07,
2013
+ "log_odds_chosen": 5.995306015014648,
2014
+ "log_odds_ratio": -0.013278511352837086,
2015
+ "logits/chosen": 7.143318176269531,
2016
+ "logits/rejected": 14.408166885375977,
2017
+ "logps/chosen": -0.048548031598329544,
2018
+ "logps/rejected": -2.2793731689453125,
2019
+ "loss": 0.0626,
2020
+ "nll_loss": 0.06035660579800606,
2021
+ "rewards/accuracies": 1.0,
2022
+ "rewards/chosen": -0.0048548029735684395,
2023
+ "rewards/margins": 0.2230825424194336,
2024
+ "rewards/rejected": -0.22793734073638916,
2025
+ "step": 1120
2026
+ },
2027
+ {
2028
+ "epoch": 1.808,
2029
+ "grad_norm": 0.3502514097075585,
2030
+ "learning_rate": 5.562362976251901e-07,
2031
+ "log_odds_chosen": 5.784189224243164,
2032
+ "log_odds_ratio": -0.019179679453372955,
2033
+ "logits/chosen": 7.51906681060791,
2034
+ "logits/rejected": 13.482861518859863,
2035
+ "logps/chosen": -0.050465572625398636,
2036
+ "logps/rejected": -2.069855213165283,
2037
+ "loss": 0.0673,
2038
+ "nll_loss": 0.06206601858139038,
2039
+ "rewards/accuracies": 1.0,
2040
+ "rewards/chosen": -0.005046557169407606,
2041
+ "rewards/margins": 0.2019389569759369,
2042
+ "rewards/rejected": -0.20698551833629608,
2043
+ "step": 1130
2044
+ },
2045
+ {
2046
+ "epoch": 1.8239999999999998,
2047
+ "grad_norm": 0.2385312527933045,
2048
+ "learning_rate": 4.6809332207053083e-07,
2049
+ "log_odds_chosen": 6.114804267883301,
2050
+ "log_odds_ratio": -0.01152479462325573,
2051
+ "logits/chosen": 8.802262306213379,
2052
+ "logits/rejected": 14.353776931762695,
2053
+ "logps/chosen": -0.05430366471409798,
2054
+ "logps/rejected": -2.6986145973205566,
2055
+ "loss": 0.0617,
2056
+ "nll_loss": 0.062255967408418655,
2057
+ "rewards/accuracies": 1.0,
2058
+ "rewards/chosen": -0.00543036637827754,
2059
+ "rewards/margins": 0.26443108916282654,
2060
+ "rewards/rejected": -0.26986145973205566,
2061
+ "step": 1140
2062
+ },
2063
+ {
2064
+ "epoch": 1.8399999999999999,
2065
+ "grad_norm": 0.5023792704599986,
2066
+ "learning_rate": 3.8738304061681107e-07,
2067
+ "log_odds_chosen": 5.99333381652832,
2068
+ "log_odds_ratio": -0.010073414072394371,
2069
+ "logits/chosen": 6.594714164733887,
2070
+ "logits/rejected": 14.024187088012695,
2071
+ "logps/chosen": -0.04959163814783096,
2072
+ "logps/rejected": -2.3722965717315674,
2073
+ "loss": 0.0664,
2074
+ "nll_loss": 0.05878063291311264,
2075
+ "rewards/accuracies": 1.0,
2076
+ "rewards/chosen": -0.004959163721650839,
2077
+ "rewards/margins": 0.2322704792022705,
2078
+ "rewards/rejected": -0.23722966015338898,
2079
+ "step": 1150
2080
+ },
2081
+ {
2082
+ "epoch": 1.8559999999999999,
2083
+ "grad_norm": 0.18925156808932705,
2084
+ "learning_rate": 3.1416838871368925e-07,
2085
+ "log_odds_chosen": 5.59714412689209,
2086
+ "log_odds_ratio": -0.016117483377456665,
2087
+ "logits/chosen": 6.784099578857422,
2088
+ "logits/rejected": 14.297111511230469,
2089
+ "logps/chosen": -0.04300536960363388,
2090
+ "logps/rejected": -1.9598991870880127,
2091
+ "loss": 0.0594,
2092
+ "nll_loss": 0.06088510900735855,
2093
+ "rewards/accuracies": 1.0,
2094
+ "rewards/chosen": -0.0043005370534956455,
2095
+ "rewards/margins": 0.1916893571615219,
2096
+ "rewards/rejected": -0.19598989188671112,
2097
+ "step": 1160
2098
+ },
2099
+ {
2100
+ "epoch": 1.8719999999999999,
2101
+ "grad_norm": 0.4752392208294527,
2102
+ "learning_rate": 2.4850645694436736e-07,
2103
+ "log_odds_chosen": 5.687636375427246,
2104
+ "log_odds_ratio": -0.009939742274582386,
2105
+ "logits/chosen": 6.543469429016113,
2106
+ "logits/rejected": 13.751274108886719,
2107
+ "logps/chosen": -0.0570371076464653,
2108
+ "logps/rejected": -2.2096800804138184,
2109
+ "loss": 0.0612,
2110
+ "nll_loss": 0.06289631128311157,
2111
+ "rewards/accuracies": 1.0,
2112
+ "rewards/chosen": -0.005703710950911045,
2113
+ "rewards/margins": 0.21526429057121277,
2114
+ "rewards/rejected": -0.22096800804138184,
2115
+ "step": 1170
2116
+ },
2117
+ {
2118
+ "epoch": 1.888,
2119
+ "grad_norm": 0.40463907978157104,
2120
+ "learning_rate": 1.9044844650808468e-07,
2121
+ "log_odds_chosen": 5.943331718444824,
2122
+ "log_odds_ratio": -0.008882846683263779,
2123
+ "logits/chosen": 7.400280952453613,
2124
+ "logits/rejected": 14.994569778442383,
2125
+ "logps/chosen": -0.045132219791412354,
2126
+ "logps/rejected": -2.3415608406066895,
2127
+ "loss": 0.0612,
2128
+ "nll_loss": 0.05737587809562683,
2129
+ "rewards/accuracies": 1.0,
2130
+ "rewards/chosen": -0.00451322179287672,
2131
+ "rewards/margins": 0.22964289784431458,
2132
+ "rewards/rejected": -0.23415610194206238,
2133
+ "step": 1180
2134
+ },
2135
+ {
2136
+ "epoch": 1.904,
2137
+ "grad_norm": 0.35720023172691506,
2138
+ "learning_rate": 1.400396292949513e-07,
2139
+ "log_odds_chosen": 5.809309959411621,
2140
+ "log_odds_ratio": -0.008964666165411472,
2141
+ "logits/chosen": 8.318174362182617,
2142
+ "logits/rejected": 13.486407279968262,
2143
+ "logps/chosen": -0.053054939955472946,
2144
+ "logps/rejected": -2.4225754737854004,
2145
+ "loss": 0.0626,
2146
+ "nll_loss": 0.06207244470715523,
2147
+ "rewards/accuracies": 1.0,
2148
+ "rewards/chosen": -0.005305493716150522,
2149
+ "rewards/margins": 0.2369520664215088,
2150
+ "rewards/rejected": -0.24225755035877228,
2151
+ "step": 1190
2152
+ },
2153
+ {
2154
+ "epoch": 1.92,
2155
+ "grad_norm": 0.45375001244411867,
2156
+ "learning_rate": 9.731931258429638e-08,
2157
+ "log_odds_chosen": 6.011991024017334,
2158
+ "log_odds_ratio": -0.011115940287709236,
2159
+ "logits/chosen": 7.217115879058838,
2160
+ "logits/rejected": 13.6605806350708,
2161
+ "logps/chosen": -0.05640987679362297,
2162
+ "logps/rejected": -2.306795835494995,
2163
+ "loss": 0.0667,
2164
+ "nll_loss": 0.064894899725914,
2165
+ "rewards/accuracies": 1.0,
2166
+ "rewards/chosen": -0.005640988238155842,
2167
+ "rewards/margins": 0.22503860294818878,
2168
+ "rewards/rejected": -0.23067960143089294,
2169
+ "step": 1200
2170
+ },
2171
+ {
2172
+ "epoch": 1.936,
2173
+ "grad_norm": 0.3851516384032617,
2174
+ "learning_rate": 6.232080839403631e-08,
2175
+ "log_odds_chosen": 6.033963203430176,
2176
+ "log_odds_ratio": -0.008175736293196678,
2177
+ "logits/chosen": 5.674324035644531,
2178
+ "logits/rejected": 12.042490005493164,
2179
+ "logps/chosen": -0.041214533150196075,
2180
+ "logps/rejected": -2.353848934173584,
2181
+ "loss": 0.0594,
2182
+ "nll_loss": 0.05154932290315628,
2183
+ "rewards/accuracies": 1.0,
2184
+ "rewards/chosen": -0.0041214535012841225,
2185
+ "rewards/margins": 0.2312634438276291,
2186
+ "rewards/rejected": -0.23538486659526825,
2187
+ "step": 1210
2188
+ },
2189
+ {
2190
+ "epoch": 1.952,
2191
+ "grad_norm": 0.2301620888800148,
2192
+ "learning_rate": 3.50714075049563e-08,
2193
+ "log_odds_chosen": 5.746167182922363,
2194
+ "log_odds_ratio": -0.01174002606421709,
2195
+ "logits/chosen": 7.780473232269287,
2196
+ "logits/rejected": 14.472818374633789,
2197
+ "logps/chosen": -0.052154283970594406,
2198
+ "logps/rejected": -2.3260200023651123,
2199
+ "loss": 0.0631,
2200
+ "nll_loss": 0.06034677103161812,
2201
+ "rewards/accuracies": 1.0,
2202
+ "rewards/chosen": -0.0052154287695884705,
2203
+ "rewards/margins": 0.22738659381866455,
2204
+ "rewards/rejected": -0.23260203003883362,
2205
+ "step": 1220
2206
+ },
2207
+ {
2208
+ "epoch": 1.968,
2209
+ "grad_norm": 0.29743582988851025,
2210
+ "learning_rate": 1.5592358180189782e-08,
2211
+ "log_odds_chosen": 5.381113529205322,
2212
+ "log_odds_ratio": -0.010704567655920982,
2213
+ "logits/chosen": 6.542645454406738,
2214
+ "logits/rejected": 14.09373664855957,
2215
+ "logps/chosen": -0.045278120785951614,
2216
+ "logps/rejected": -1.7775919437408447,
2217
+ "loss": 0.0605,
2218
+ "nll_loss": 0.057972025126218796,
2219
+ "rewards/accuracies": 1.0,
2220
+ "rewards/chosen": -0.004527812357991934,
2221
+ "rewards/margins": 0.1732313632965088,
2222
+ "rewards/rejected": -0.17775918543338776,
2223
+ "step": 1230
2224
+ },
2225
+ {
2226
+ "epoch": 1.984,
2227
+ "grad_norm": 0.3648997358256397,
2228
+ "learning_rate": 3.898849596456477e-09,
2229
+ "log_odds_chosen": 5.863680839538574,
2230
+ "log_odds_ratio": -0.009549752809107304,
2231
+ "logits/chosen": 6.157291889190674,
2232
+ "logits/rejected": 12.384984016418457,
2233
+ "logps/chosen": -0.05373894423246384,
2234
+ "logps/rejected": -2.254366397857666,
2235
+ "loss": 0.0613,
2236
+ "nll_loss": 0.06132974475622177,
2237
+ "rewards/accuracies": 1.0,
2238
+ "rewards/chosen": -0.005373895168304443,
2239
+ "rewards/margins": 0.2200627624988556,
2240
+ "rewards/rejected": -0.22543664276599884,
2241
+ "step": 1240
2242
+ },
2243
+ {
2244
+ "epoch": 2.0,
2245
+ "grad_norm": 0.6054356785823999,
2246
+ "learning_rate": 0.0,
2247
+ "log_odds_chosen": 5.457306861877441,
2248
+ "log_odds_ratio": -0.019121190533041954,
2249
+ "logits/chosen": 7.081140041351318,
2250
+ "logits/rejected": 13.820508003234863,
2251
+ "logps/chosen": -0.0638023242354393,
2252
+ "logps/rejected": -1.9536654949188232,
2253
+ "loss": 0.0679,
2254
+ "nll_loss": 0.067476287484169,
2255
+ "rewards/accuracies": 1.0,
2256
+ "rewards/chosen": -0.006380232982337475,
2257
+ "rewards/margins": 0.18898631632328033,
2258
+ "rewards/rejected": -0.19536656141281128,
2259
+ "step": 1250
2260
+ },
2261
+ {
2262
+ "epoch": 2.0,
2263
+ "step": 1250,
2264
+ "total_flos": 0.0,
2265
+ "train_loss": 0.2130171504020691,
2266
+ "train_runtime": 11716.0982,
2267
+ "train_samples_per_second": 3.414,
2268
+ "train_steps_per_second": 0.107
2269
+ }
2270
+ ],
2271
+ "logging_steps": 10,
2272
+ "max_steps": 1250,
2273
+ "num_input_tokens_seen": 0,
2274
+ "num_train_epochs": 2,
2275
+ "save_steps": 500,
2276
+ "stateful_callbacks": {
2277
+ "TrainerControl": {
2278
+ "args": {
2279
+ "should_epoch_stop": false,
2280
+ "should_evaluate": false,
2281
+ "should_log": false,
2282
+ "should_save": true,
2283
+ "should_training_stop": true
2284
+ },
2285
+ "attributes": {}
2286
+ }
2287
+ },
2288
+ "total_flos": 0.0,
2289
+ "train_batch_size": 4,
2290
+ "trial_name": null,
2291
+ "trial_params": null
2292
+ }