hZzy commited on
Commit
93e6ca2
1 Parent(s): 0d6f856

Model save

Browse files
README.md ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: hZzy/qwen2.5-0.5b-sft-news-IFT
4
+ tags:
5
+ - trl
6
+ - expo
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: qwen2.5-0.5b-expo-DPO-EXPERIMENT-1W
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/zhiyuzha-university-of-florida/huggingface/runs/vdo32atw)
17
+ # qwen2.5-0.5b-expo-DPO-EXPERIMENT-1W
18
+
19
+ This model is a fine-tuned version of [hZzy/qwen2.5-0.5b-sft-news-IFT](https://huggingface.co/hZzy/qwen2.5-0.5b-sft-news-IFT) on an unknown dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 12664.2070
22
+ - Logps: -80.5751
23
+ - Logits: -1.3556
24
+ - Objective: 12926.7227
25
+ - Dpo Loss: 12926.7227
26
+ - Regularize: 12926.7227
27
+ - Ranking Simple: 0.5227
28
+ - Ranking Idealized: 0.5093
29
+ - Ranking Idealized Expo: 0.5093
30
+
31
+ ## Model description
32
+
33
+ More information needed
34
+
35
+ ## Intended uses & limitations
36
+
37
+ More information needed
38
+
39
+ ## Training and evaluation data
40
+
41
+ More information needed
42
+
43
+ ## Training procedure
44
+
45
+ ### Training hyperparameters
46
+
47
+ The following hyperparameters were used during training:
48
+ - learning_rate: 5e-06
49
+ - train_batch_size: 4
50
+ - eval_batch_size: 4
51
+ - seed: 42
52
+ - distributed_type: multi-GPU
53
+ - num_devices: 6
54
+ - gradient_accumulation_steps: 12
55
+ - total_train_batch_size: 288
56
+ - total_eval_batch_size: 24
57
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
58
+ - lr_scheduler_type: cosine
59
+ - lr_scheduler_warmup_ratio: 0.1
60
+ - num_epochs: 1
61
+
62
+ ### Training results
63
+
64
+ | Training Loss | Epoch | Step | Validation Loss | Logps | Logits | Objective | Dpo Loss | Regularize | Ranking Simple | Ranking Idealized | Ranking Idealized Expo |
65
+ |:-------------:|:------:|:----:|:---------------:|:--------:|:-------:|:----------:|:----------:|:----------:|:--------------:|:-----------------:|:----------------------:|
66
+ | 10470.1625 | 0.2834 | 50 | 11322.9775 | -84.9886 | -1.4586 | 11815.3389 | 11815.3389 | 11815.3389 | 0.5165 | 0.5093 | 0.5093 |
67
+ | 8558.2836 | 0.5668 | 100 | 12877.2754 | -82.3385 | -1.3720 | 13248.5166 | 13248.5166 | 13248.5166 | 0.5176 | 0.5093 | 0.5093 |
68
+ | 7521.95 | 0.8503 | 150 | 12664.2070 | -80.5751 | -1.3556 | 12926.7227 | 12926.7227 | 12926.7227 | 0.5227 | 0.5093 | 0.5093 |
69
+
70
+
71
+ ### Framework versions
72
+
73
+ - Transformers 4.42.0
74
+ - Pytorch 2.3.0+cu121
75
+ - Datasets 2.19.1
76
+ - Tokenizers 0.19.1
added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
all_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.997638167217761,
3
+ "total_flos": 0.0,
4
+ "train_loss": 8072.862625997175,
5
+ "train_runtime": 10876.2953,
6
+ "train_samples": 50802,
7
+ "train_samples_per_second": 4.671,
8
+ "train_steps_per_second": 0.016
9
+ }
config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "hZzy/qwen2.5-0.5b-sft-news-IFT",
3
+ "architectures": [
4
+ "Qwen2ForCausalLM"
5
+ ],
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 151644,
8
+ "eos_token_id": 151645,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 896,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 4864,
13
+ "max_position_embeddings": 32768,
14
+ "max_window_layers": 24,
15
+ "model_type": "qwen2",
16
+ "num_attention_heads": 14,
17
+ "num_hidden_layers": 24,
18
+ "num_key_value_heads": 2,
19
+ "pad_token_id": 151645,
20
+ "rms_norm_eps": 1e-06,
21
+ "rope_theta": 1000000.0,
22
+ "sliding_window": 32768,
23
+ "tie_word_embeddings": true,
24
+ "torch_dtype": "float32",
25
+ "transformers_version": "4.42.0",
26
+ "use_cache": false,
27
+ "use_mrope": false,
28
+ "use_sliding_window": false,
29
+ "vocab_size": 151665
30
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151644,
3
+ "eos_token_id": 151645,
4
+ "max_new_tokens": 2048,
5
+ "pad_token_id": 151645,
6
+ "transformers_version": "4.42.0"
7
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e25a672a0099634dac3b7237a79c2dd5bce57c5d814f490409b8430f859fe07e
3
+ size 1975192208
special_tokens_map.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>"
5
+ ],
6
+ "bos_token": {
7
+ "content": "<|im_start|>",
8
+ "lstrip": false,
9
+ "normalized": false,
10
+ "rstrip": false,
11
+ "single_word": false
12
+ },
13
+ "eos_token": {
14
+ "content": "<|im_end|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false
19
+ },
20
+ "pad_token": {
21
+ "content": "<|im_end|>",
22
+ "lstrip": false,
23
+ "normalized": false,
24
+ "rstrip": false,
25
+ "single_word": false
26
+ }
27
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>"
185
+ ],
186
+ "bos_token": "<|im_start|>",
187
+ "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n' + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}",
188
+ "clean_up_tokenization_spaces": false,
189
+ "eos_token": "<|im_end|>",
190
+ "errors": "replace",
191
+ "model_max_length": 2048,
192
+ "pad_token": "<|im_end|>",
193
+ "split_special_tokens": false,
194
+ "tokenizer_class": "Qwen2Tokenizer",
195
+ "unk_token": null
196
+ }
train_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 0.997638167217761,
3
+ "total_flos": 0.0,
4
+ "train_loss": 8072.862625997175,
5
+ "train_runtime": 10876.2953,
6
+ "train_samples": 50802,
7
+ "train_samples_per_second": 4.671,
8
+ "train_steps_per_second": 0.016
9
+ }
trainer_state.json ADDED
@@ -0,0 +1,630 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 0.997638167217761,
5
+ "eval_steps": 50,
6
+ "global_step": 176,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "dpo_loss": 0.6931471824645996,
13
+ "epoch": 0.005668398677373642,
14
+ "grad_norm": 1341367.1350394282,
15
+ "learning_rate": 2.7777777777777776e-07,
16
+ "logits": -1.3147305250167847,
17
+ "logps": -88.0877456665039,
18
+ "loss": 0.6931,
19
+ "objective": 0.6931471824645996,
20
+ "ranking_idealized": 0.5208333134651184,
21
+ "ranking_idealized_expo": 0.5208333134651184,
22
+ "ranking_simple": 0.5208333134651184,
23
+ "regularize": 0.6931471824645996,
24
+ "step": 1
25
+ },
26
+ {
27
+ "dpo_loss": 1280.619140625,
28
+ "epoch": 0.02834199338686821,
29
+ "grad_norm": 2623988.3634039685,
30
+ "learning_rate": 1.3888888888888892e-06,
31
+ "logits": -1.3680986166000366,
32
+ "logps": -84.14160919189453,
33
+ "loss": 1112.7681,
34
+ "objective": 1280.619140625,
35
+ "ranking_idealized": 0.546875,
36
+ "ranking_idealized_expo": 0.546875,
37
+ "ranking_simple": 0.5520833134651184,
38
+ "regularize": 1280.619140625,
39
+ "step": 5
40
+ },
41
+ {
42
+ "dpo_loss": 2070.714599609375,
43
+ "epoch": 0.05668398677373642,
44
+ "grad_norm": 2330391.207726475,
45
+ "learning_rate": 2.7777777777777783e-06,
46
+ "logits": -1.45966374874115,
47
+ "logps": -83.2505111694336,
48
+ "loss": 2407.7941,
49
+ "objective": 2070.714599609375,
50
+ "ranking_idealized": 0.5083333253860474,
51
+ "ranking_idealized_expo": 0.5083333253860474,
52
+ "ranking_simple": 0.5166666507720947,
53
+ "regularize": 2070.714599609375,
54
+ "step": 10
55
+ },
56
+ {
57
+ "dpo_loss": 5679.9853515625,
58
+ "epoch": 0.08502598016060463,
59
+ "grad_norm": 1644564.4852830889,
60
+ "learning_rate": 4.166666666666667e-06,
61
+ "logits": -1.4328060150146484,
62
+ "logps": -82.17012023925781,
63
+ "loss": 4264.0652,
64
+ "objective": 5679.9853515625,
65
+ "ranking_idealized": 0.5041666626930237,
66
+ "ranking_idealized_expo": 0.5041666626930237,
67
+ "ranking_simple": 0.5166666507720947,
68
+ "regularize": 5679.9853515625,
69
+ "step": 15
70
+ },
71
+ {
72
+ "dpo_loss": 7525.20068359375,
73
+ "epoch": 0.11336797354747284,
74
+ "grad_norm": 1515613.7017780922,
75
+ "learning_rate": 4.998023493068255e-06,
76
+ "logits": -1.4038361310958862,
77
+ "logps": -82.86460876464844,
78
+ "loss": 6967.5602,
79
+ "objective": 7525.20068359375,
80
+ "ranking_idealized": 0.5375000238418579,
81
+ "ranking_idealized_expo": 0.5375000238418579,
82
+ "ranking_simple": 0.5583333373069763,
83
+ "regularize": 7525.20068359375,
84
+ "step": 20
85
+ },
86
+ {
87
+ "dpo_loss": 8746.1123046875,
88
+ "epoch": 0.14170996693434104,
89
+ "grad_norm": 1687436.4259306653,
90
+ "learning_rate": 4.975823666181256e-06,
91
+ "logits": -1.371711254119873,
92
+ "logps": -82.0951156616211,
93
+ "loss": 9527.1836,
94
+ "objective": 8746.1123046875,
95
+ "ranking_idealized": 0.4958333373069763,
96
+ "ranking_idealized_expo": 0.4958333373069763,
97
+ "ranking_simple": 0.512499988079071,
98
+ "regularize": 8746.1123046875,
99
+ "step": 25
100
+ },
101
+ {
102
+ "dpo_loss": 11099.513671875,
103
+ "epoch": 0.17005196032120926,
104
+ "grad_norm": 1294226.520732075,
105
+ "learning_rate": 4.929173350101025e-06,
106
+ "logits": -1.3348666429519653,
107
+ "logps": -81.38406372070312,
108
+ "loss": 10099.4609,
109
+ "objective": 11099.513671875,
110
+ "ranking_idealized": 0.5416666865348816,
111
+ "ranking_idealized_expo": 0.5416666865348816,
112
+ "ranking_simple": 0.5458333492279053,
113
+ "regularize": 11099.513671875,
114
+ "step": 30
115
+ },
116
+ {
117
+ "dpo_loss": 12696.7705078125,
118
+ "epoch": 0.19839395370807747,
119
+ "grad_norm": 1391829.2894247118,
120
+ "learning_rate": 4.858533249305337e-06,
121
+ "logits": -1.411891222000122,
122
+ "logps": -77.1667251586914,
123
+ "loss": 9874.5609,
124
+ "objective": 12696.7705078125,
125
+ "ranking_idealized": 0.5041666626930237,
126
+ "ranking_idealized_expo": 0.5041666626930237,
127
+ "ranking_simple": 0.5375000238418579,
128
+ "regularize": 12696.7705078125,
129
+ "step": 35
130
+ },
131
+ {
132
+ "dpo_loss": 8431.9521484375,
133
+ "epoch": 0.22673594709494568,
134
+ "grad_norm": 1186854.0601841968,
135
+ "learning_rate": 4.764600984163809e-06,
136
+ "logits": -1.3430523872375488,
137
+ "logps": -77.41344451904297,
138
+ "loss": 10212.3687,
139
+ "objective": 8431.9521484375,
140
+ "ranking_idealized": 0.5208333134651184,
141
+ "ranking_idealized_expo": 0.5208333134651184,
142
+ "ranking_simple": 0.5874999761581421,
143
+ "regularize": 8431.9521484375,
144
+ "step": 40
145
+ },
146
+ {
147
+ "dpo_loss": 11319.8056640625,
148
+ "epoch": 0.25507794048181387,
149
+ "grad_norm": 1072714.7178744846,
150
+ "learning_rate": 4.648304201449153e-06,
151
+ "logits": -1.3935389518737793,
152
+ "logps": -79.65689086914062,
153
+ "loss": 10434.6648,
154
+ "objective": 11319.8056640625,
155
+ "ranking_idealized": 0.5375000238418579,
156
+ "ranking_idealized_expo": 0.5375000238418579,
157
+ "ranking_simple": 0.574999988079071,
158
+ "regularize": 11319.8056640625,
159
+ "step": 45
160
+ },
161
+ {
162
+ "dpo_loss": 9569.685546875,
163
+ "epoch": 0.2834199338686821,
164
+ "grad_norm": 950921.657886262,
165
+ "learning_rate": 4.510791413176912e-06,
166
+ "logits": -1.3682667016983032,
167
+ "logps": -77.96815490722656,
168
+ "loss": 10470.1625,
169
+ "objective": 9569.685546875,
170
+ "ranking_idealized": 0.42500001192092896,
171
+ "ranking_idealized_expo": 0.42500001192092896,
172
+ "ranking_simple": 0.5,
173
+ "regularize": 9569.685546875,
174
+ "step": 50
175
+ },
176
+ {
177
+ "epoch": 0.2834199338686821,
178
+ "eval_dpo_loss": 11815.3388671875,
179
+ "eval_logits": -1.4585951566696167,
180
+ "eval_logps": -84.9886474609375,
181
+ "eval_loss": 11322.9775390625,
182
+ "eval_objective": 11815.3388671875,
183
+ "eval_ranking_idealized": 0.5092975497245789,
184
+ "eval_ranking_idealized_expo": 0.5092975497245789,
185
+ "eval_ranking_simple": 0.5165289044380188,
186
+ "eval_regularize": 11815.3388671875,
187
+ "eval_runtime": 1211.6497,
188
+ "eval_samples_per_second": 4.779,
189
+ "eval_steps_per_second": 0.2,
190
+ "step": 50
191
+ },
192
+ {
193
+ "dpo_loss": 12371.0048828125,
194
+ "epoch": 0.3117619272555503,
195
+ "grad_norm": 1109923.0216191527,
196
+ "learning_rate": 4.353420654246546e-06,
197
+ "logits": -1.4771044254302979,
198
+ "logps": -79.130126953125,
199
+ "loss": 11113.0828,
200
+ "objective": 12371.0048828125,
201
+ "ranking_idealized": 0.5416666865348816,
202
+ "ranking_idealized_expo": 0.5416666865348816,
203
+ "ranking_simple": 0.5541666746139526,
204
+ "regularize": 12371.0048828125,
205
+ "step": 55
206
+ },
207
+ {
208
+ "dpo_loss": 8734.4072265625,
209
+ "epoch": 0.3401039206424185,
210
+ "grad_norm": 1211359.9780672283,
211
+ "learning_rate": 4.177746070897593e-06,
212
+ "logits": -1.4155789613723755,
213
+ "logps": -77.57923126220703,
214
+ "loss": 11080.9961,
215
+ "objective": 8734.4072265625,
216
+ "ranking_idealized": 0.512499988079071,
217
+ "ranking_idealized_expo": 0.512499988079071,
218
+ "ranking_simple": 0.5708333253860474,
219
+ "regularize": 8734.4072265625,
220
+ "step": 60
221
+ },
222
+ {
223
+ "dpo_loss": 9739.5908203125,
224
+ "epoch": 0.3684459140292867,
225
+ "grad_norm": 1057255.008796819,
226
+ "learning_rate": 3.985502572429277e-06,
227
+ "logits": -1.3343302011489868,
228
+ "logps": -77.00284576416016,
229
+ "loss": 11711.2734,
230
+ "objective": 9739.5908203125,
231
+ "ranking_idealized": 0.5083333253860474,
232
+ "ranking_idealized_expo": 0.5083333253860474,
233
+ "ranking_simple": 0.5541666746139526,
234
+ "regularize": 9739.5908203125,
235
+ "step": 65
236
+ },
237
+ {
238
+ "dpo_loss": 9313.849609375,
239
+ "epoch": 0.39678790741615494,
240
+ "grad_norm": 1088725.4948180364,
241
+ "learning_rate": 3.7785886977585562e-06,
242
+ "logits": -1.3323314189910889,
243
+ "logps": -76.60472106933594,
244
+ "loss": 11694.0484,
245
+ "objective": 9313.849609375,
246
+ "ranking_idealized": 0.5166666507720947,
247
+ "ranking_idealized_expo": 0.5166666507720947,
248
+ "ranking_simple": 0.5416666865348816,
249
+ "regularize": 9313.849609375,
250
+ "step": 70
251
+ },
252
+ {
253
+ "dpo_loss": 10158.4462890625,
254
+ "epoch": 0.42512990080302315,
255
+ "grad_norm": 999614.3040187941,
256
+ "learning_rate": 3.5590478660213214e-06,
257
+ "logits": -1.2671582698822021,
258
+ "logps": -77.10884094238281,
259
+ "loss": 11305.65,
260
+ "objective": 10158.4462890625,
261
+ "ranking_idealized": 0.5458333492279053,
262
+ "ranking_idealized_expo": 0.5458333492279053,
263
+ "ranking_simple": 0.5916666388511658,
264
+ "regularize": 10158.4462890625,
265
+ "step": 75
266
+ },
267
+ {
268
+ "dpo_loss": 13367.1591796875,
269
+ "epoch": 0.45347189418989137,
270
+ "grad_norm": 1036547.2837553191,
271
+ "learning_rate": 3.32904819638017e-06,
272
+ "logits": -1.2884188890457153,
273
+ "logps": -76.65129852294922,
274
+ "loss": 11108.9961,
275
+ "objective": 13367.1591796875,
276
+ "ranking_idealized": 0.5166666507720947,
277
+ "ranking_idealized_expo": 0.5166666507720947,
278
+ "ranking_simple": 0.5583333373069763,
279
+ "regularize": 13367.1591796875,
280
+ "step": 80
281
+ },
282
+ {
283
+ "dpo_loss": 9063.92578125,
284
+ "epoch": 0.4818138875767596,
285
+ "grad_norm": 878439.4100295339,
286
+ "learning_rate": 3.090861096332263e-06,
287
+ "logits": -1.2684974670410156,
288
+ "logps": -75.37580108642578,
289
+ "loss": 9492.8891,
290
+ "objective": 9063.92578125,
291
+ "ranking_idealized": 0.5249999761581421,
292
+ "ranking_idealized_expo": 0.5249999761581421,
293
+ "ranking_simple": 0.5916666388511658,
294
+ "regularize": 9063.92578125,
295
+ "step": 85
296
+ },
297
+ {
298
+ "dpo_loss": 12140.421875,
299
+ "epoch": 0.5101558809636277,
300
+ "grad_norm": 839896.1614839052,
301
+ "learning_rate": 2.8468388299726714e-06,
302
+ "logits": -1.1826907396316528,
303
+ "logps": -77.36532592773438,
304
+ "loss": 8699.0336,
305
+ "objective": 12140.421875,
306
+ "ranking_idealized": 0.4791666567325592,
307
+ "ranking_idealized_expo": 0.4791666567325592,
308
+ "ranking_simple": 0.5166666507720947,
309
+ "regularize": 12140.421875,
310
+ "step": 90
311
+ },
312
+ {
313
+ "dpo_loss": 7882.3046875,
314
+ "epoch": 0.538497874350496,
315
+ "grad_norm": 1169515.5696928622,
316
+ "learning_rate": 2.599391287742315e-06,
317
+ "logits": -1.2656235694885254,
318
+ "logps": -76.93645477294922,
319
+ "loss": 8873.2109,
320
+ "objective": 7882.3046875,
321
+ "ranking_idealized": 0.5416666865348816,
322
+ "ranking_idealized_expo": 0.5416666865348816,
323
+ "ranking_simple": 0.6166666746139526,
324
+ "regularize": 7882.3046875,
325
+ "step": 95
326
+ },
327
+ {
328
+ "dpo_loss": 7251.060546875,
329
+ "epoch": 0.5668398677373642,
330
+ "grad_norm": 1006084.9040205956,
331
+ "learning_rate": 2.3509621870754505e-06,
332
+ "logits": -1.342369794845581,
333
+ "logps": -78.5166244506836,
334
+ "loss": 8558.2836,
335
+ "objective": 7251.060546875,
336
+ "ranking_idealized": 0.5416666865348816,
337
+ "ranking_idealized_expo": 0.5416666865348816,
338
+ "ranking_simple": 0.6041666865348816,
339
+ "regularize": 7251.060546875,
340
+ "step": 100
341
+ },
342
+ {
343
+ "epoch": 0.5668398677373642,
344
+ "eval_dpo_loss": 13248.5166015625,
345
+ "eval_logits": -1.3720418214797974,
346
+ "eval_logps": -82.3385238647461,
347
+ "eval_loss": 12877.275390625,
348
+ "eval_objective": 13248.5166015625,
349
+ "eval_ranking_idealized": 0.5092975497245789,
350
+ "eval_ranking_idealized_expo": 0.5092975497245789,
351
+ "eval_ranking_simple": 0.5175619721412659,
352
+ "eval_regularize": 13248.5166015625,
353
+ "eval_runtime": 258.8572,
354
+ "eval_samples_per_second": 22.368,
355
+ "eval_steps_per_second": 0.935,
356
+ "step": 100
357
+ },
358
+ {
359
+ "dpo_loss": 9016.806640625,
360
+ "epoch": 0.5951818611242324,
361
+ "grad_norm": 861666.5853674684,
362
+ "learning_rate": 2.1040049389819628e-06,
363
+ "logits": -1.238232970237732,
364
+ "logps": -78.26189422607422,
365
+ "loss": 8298.018,
366
+ "objective": 9016.806640625,
367
+ "ranking_idealized": 0.5333333611488342,
368
+ "ranking_idealized_expo": 0.5333333611488342,
369
+ "ranking_simple": 0.612500011920929,
370
+ "regularize": 9016.806640625,
371
+ "step": 105
372
+ },
373
+ {
374
+ "dpo_loss": 7970.41748046875,
375
+ "epoch": 0.6235238545111006,
376
+ "grad_norm": 865124.6869036607,
377
+ "learning_rate": 1.8609584188988135e-06,
378
+ "logits": -1.330941081047058,
379
+ "logps": -76.76597595214844,
380
+ "loss": 7414.1281,
381
+ "objective": 7970.41748046875,
382
+ "ranking_idealized": 0.44999998807907104,
383
+ "ranking_idealized_expo": 0.44999998807907104,
384
+ "ranking_simple": 0.5416666865348816,
385
+ "regularize": 7970.41748046875,
386
+ "step": 110
387
+ },
388
+ {
389
+ "dpo_loss": 9534.12890625,
390
+ "epoch": 0.6518658478979689,
391
+ "grad_norm": 840576.3563797984,
392
+ "learning_rate": 1.6242228810904393e-06,
393
+ "logits": -1.3068194389343262,
394
+ "logps": -76.78305053710938,
395
+ "loss": 8769.2398,
396
+ "objective": 9534.12890625,
397
+ "ranking_idealized": 0.4833333194255829,
398
+ "ranking_idealized_expo": 0.4833333194255829,
399
+ "ranking_simple": 0.5458333492279053,
400
+ "regularize": 9534.12890625,
401
+ "step": 115
402
+ },
403
+ {
404
+ "dpo_loss": 8552.21484375,
405
+ "epoch": 0.680207841284837,
406
+ "grad_norm": 794346.4700274817,
407
+ "learning_rate": 1.3961362544602215e-06,
408
+ "logits": -1.2461904287338257,
409
+ "logps": -77.25305938720703,
410
+ "loss": 7377.8172,
411
+ "objective": 8552.21484375,
412
+ "ranking_idealized": 0.5333333611488342,
413
+ "ranking_idealized_expo": 0.5333333611488342,
414
+ "ranking_simple": 0.5874999761581421,
415
+ "regularize": 8552.21484375,
416
+ "step": 120
417
+ },
418
+ {
419
+ "dpo_loss": 9080.310546875,
420
+ "epoch": 0.7085498346717053,
421
+ "grad_norm": 866794.9526252571,
422
+ "learning_rate": 1.1789510538684524e-06,
423
+ "logits": -1.2470247745513916,
424
+ "logps": -75.66565704345703,
425
+ "loss": 7754.6125,
426
+ "objective": 9080.310546875,
427
+ "ranking_idealized": 0.550000011920929,
428
+ "ranking_idealized_expo": 0.550000011920929,
429
+ "ranking_simple": 0.6000000238418579,
430
+ "regularize": 9080.310546875,
431
+ "step": 125
432
+ },
433
+ {
434
+ "dpo_loss": 7816.40869140625,
435
+ "epoch": 0.7368918280585735,
436
+ "grad_norm": 914484.5347620883,
437
+ "learning_rate": 9.74812134973689e-07,
438
+ "logits": -1.362328290939331,
439
+ "logps": -74.44100952148438,
440
+ "loss": 7071.1984,
441
+ "objective": 7816.40869140625,
442
+ "ranking_idealized": 0.44999998807907104,
443
+ "ranking_idealized_expo": 0.44999998807907104,
444
+ "ranking_simple": 0.5041666626930237,
445
+ "regularize": 7816.40869140625,
446
+ "step": 130
447
+ },
448
+ {
449
+ "dpo_loss": 6297.37939453125,
450
+ "epoch": 0.7652338214454416,
451
+ "grad_norm": 908053.7687284376,
452
+ "learning_rate": 7.857355122839674e-07,
453
+ "logits": -1.3475323915481567,
454
+ "logps": -76.51807403564453,
455
+ "loss": 7334.7656,
456
+ "objective": 6297.37939453125,
457
+ "ranking_idealized": 0.5291666388511658,
458
+ "ranking_idealized_expo": 0.5291666388511658,
459
+ "ranking_simple": 0.5833333134651184,
460
+ "regularize": 6297.37939453125,
461
+ "step": 135
462
+ },
463
+ {
464
+ "dpo_loss": 5713.158203125,
465
+ "epoch": 0.7935758148323099,
466
+ "grad_norm": 765346.0794896972,
467
+ "learning_rate": 6.135884496044245e-07,
468
+ "logits": -1.2585749626159668,
469
+ "logps": -75.7166519165039,
470
+ "loss": 6126.0992,
471
+ "objective": 5713.158203125,
472
+ "ranking_idealized": 0.5041666626930237,
473
+ "ranking_idealized_expo": 0.5041666626930237,
474
+ "ranking_simple": 0.5708333253860474,
475
+ "regularize": 5713.158203125,
476
+ "step": 140
477
+ },
478
+ {
479
+ "dpo_loss": 6023.36669921875,
480
+ "epoch": 0.821917808219178,
481
+ "grad_norm": 864232.6098968883,
482
+ "learning_rate": 4.6007101950209827e-07,
483
+ "logits": -1.311624526977539,
484
+ "logps": -75.54283905029297,
485
+ "loss": 6440.6867,
486
+ "objective": 6023.36669921875,
487
+ "ranking_idealized": 0.4833333194255829,
488
+ "ranking_idealized_expo": 0.4833333194255829,
489
+ "ranking_simple": 0.5333333611488342,
490
+ "regularize": 6023.36669921875,
491
+ "step": 145
492
+ },
493
+ {
494
+ "dpo_loss": 9931.4228515625,
495
+ "epoch": 0.8502598016060463,
496
+ "grad_norm": 789706.2833060493,
497
+ "learning_rate": 3.266993139010438e-07,
498
+ "logits": -1.3248921632766724,
499
+ "logps": -73.5119857788086,
500
+ "loss": 7521.95,
501
+ "objective": 9931.4228515625,
502
+ "ranking_idealized": 0.5375000238418579,
503
+ "ranking_idealized_expo": 0.5375000238418579,
504
+ "ranking_simple": 0.5833333134651184,
505
+ "regularize": 9931.4228515625,
506
+ "step": 150
507
+ },
508
+ {
509
+ "epoch": 0.8502598016060463,
510
+ "eval_dpo_loss": 12926.72265625,
511
+ "eval_logits": -1.3555928468704224,
512
+ "eval_logps": -80.5750961303711,
513
+ "eval_loss": 12664.20703125,
514
+ "eval_objective": 12926.72265625,
515
+ "eval_ranking_idealized": 0.5092975497245789,
516
+ "eval_ranking_idealized_expo": 0.5092975497245789,
517
+ "eval_ranking_simple": 0.5227272510528564,
518
+ "eval_regularize": 12926.72265625,
519
+ "eval_runtime": 258.9258,
520
+ "eval_samples_per_second": 22.362,
521
+ "eval_steps_per_second": 0.935,
522
+ "step": 150
523
+ },
524
+ {
525
+ "dpo_loss": 5853.70361328125,
526
+ "epoch": 0.8786017949929145,
527
+ "grad_norm": 805936.4979433398,
528
+ "learning_rate": 2.1479047161491351e-07,
529
+ "logits": -1.3135905265808105,
530
+ "logps": -73.38835144042969,
531
+ "loss": 5683.3855,
532
+ "objective": 5853.70361328125,
533
+ "ranking_idealized": 0.512499988079071,
534
+ "ranking_idealized_expo": 0.512499988079071,
535
+ "ranking_simple": 0.612500011920929,
536
+ "regularize": 5853.70361328125,
537
+ "step": 155
538
+ },
539
+ {
540
+ "dpo_loss": 6223.60693359375,
541
+ "epoch": 0.9069437883797827,
542
+ "grad_norm": 771613.5624330686,
543
+ "learning_rate": 1.2544967068054332e-07,
544
+ "logits": -1.2275891304016113,
545
+ "logps": -74.46540069580078,
546
+ "loss": 6799.2719,
547
+ "objective": 6223.60693359375,
548
+ "ranking_idealized": 0.4958333373069763,
549
+ "ranking_idealized_expo": 0.4958333373069763,
550
+ "ranking_simple": 0.574999988079071,
551
+ "regularize": 6223.60693359375,
552
+ "step": 160
553
+ },
554
+ {
555
+ "dpo_loss": 4967.171875,
556
+ "epoch": 0.9352857817666509,
557
+ "grad_norm": 1004063.6300908878,
558
+ "learning_rate": 5.955921395237318e-08,
559
+ "logits": -1.2412575483322144,
560
+ "logps": -75.6294174194336,
561
+ "loss": 6197.7492,
562
+ "objective": 4967.171875,
563
+ "ranking_idealized": 0.5208333134651184,
564
+ "ranking_idealized_expo": 0.5208333134651184,
565
+ "ranking_simple": 0.5958333611488342,
566
+ "regularize": 4967.171875,
567
+ "step": 165
568
+ },
569
+ {
570
+ "dpo_loss": 7402.99267578125,
571
+ "epoch": 0.9636277751535192,
572
+ "grad_norm": 778643.4373397075,
573
+ "learning_rate": 1.7769815745066476e-08,
574
+ "logits": -1.2604658603668213,
575
+ "logps": -75.88407135009766,
576
+ "loss": 6623.1367,
577
+ "objective": 7402.99267578125,
578
+ "ranking_idealized": 0.5083333253860474,
579
+ "ranking_idealized_expo": 0.5083333253860474,
580
+ "ranking_simple": 0.5666666626930237,
581
+ "regularize": 7402.99267578125,
582
+ "step": 170
583
+ },
584
+ {
585
+ "dpo_loss": 5726.1806640625,
586
+ "epoch": 0.9919697685403873,
587
+ "grad_norm": 762884.4185589707,
588
+ "learning_rate": 4.941755748361088e-10,
589
+ "logits": -1.2217018604278564,
590
+ "logps": -76.01050567626953,
591
+ "loss": 5562.8855,
592
+ "objective": 5726.1806640625,
593
+ "ranking_idealized": 0.574999988079071,
594
+ "ranking_idealized_expo": 0.574999988079071,
595
+ "ranking_simple": 0.6333333253860474,
596
+ "regularize": 5726.1806640625,
597
+ "step": 175
598
+ },
599
+ {
600
+ "epoch": 0.997638167217761,
601
+ "step": 176,
602
+ "total_flos": 0.0,
603
+ "train_loss": 8072.862625997175,
604
+ "train_runtime": 10876.2953,
605
+ "train_samples_per_second": 4.671,
606
+ "train_steps_per_second": 0.016
607
+ }
608
+ ],
609
+ "logging_steps": 5,
610
+ "max_steps": 176,
611
+ "num_input_tokens_seen": 0,
612
+ "num_train_epochs": 1,
613
+ "save_steps": 500,
614
+ "stateful_callbacks": {
615
+ "TrainerControl": {
616
+ "args": {
617
+ "should_epoch_stop": false,
618
+ "should_evaluate": false,
619
+ "should_log": false,
620
+ "should_save": false,
621
+ "should_training_stop": false
622
+ },
623
+ "attributes": {}
624
+ }
625
+ },
626
+ "total_flos": 0.0,
627
+ "train_batch_size": 4,
628
+ "trial_name": null,
629
+ "trial_params": null
630
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b8198cad37243392c8c785ac399571ba23fcba30addbad1b0966add4517521a
3
+ size 8120
vocab.json ADDED
The diff for this file is too large to render. See raw diff